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 * Add "abspath" and "includedir" pragma's to config files, to prevent,
27 or modify relative pathname inclusion.
31 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
32 validated. The module is implemented as an OpenSSL provider, the so-called
33 FIPS provider. A list of all changes related to the FIPS provider would go
34 beyond the scope of this CHANGES file, please consult the README-FIPS and
35 README-PROVIDERS files, as well as the migration guide.
37 The FIPS provider is disabled by default and needs to be enabled explicitly
38 at configuration time using the `enable-fips` option. If it is enabled,
39 the FIPS provider gets built and installed in addition to the default and
40 the legacy provider. No separate installation procedure is necessary.
41 There is however a dedicated `install_fips` make target, which serves the
42 special purpose of installing only the FIPS provider into an existing
45 *OpenSSL team members and many third party contributors*
47 * For the key types DH and DHX the allowed settable parameters are now different.
48 Previously (in 1.1.1) these conflicting parameters were allowed, but will now
49 result in errors. See EVP_PKEY-DH(7) for further details. This affects the
50 behaviour of openssl-genpkey(1) for DH parameter generation.
54 * Added enhanced PKCS#12 APIs which accept a library context `OSSL_LIB_CTX`
55 and (where relevant) a property query. Other APIs which handle PKCS#7 and
56 PKCS#8 objects have also been enhanced where required. This includes:
58 PKCS12_add_key_ex(), PKCS12_add_safe_ex(), PKCS12_add_safes_ex(),
59 PKCS12_create_ex(), PKCS12_decrypt_skey_ex(), PKCS12_init_ex(),
60 PKCS12_item_decrypt_d2i_ex(), PKCS12_item_i2d_encrypt_ex(),
61 PKCS12_key_gen_asc_ex(), PKCS12_key_gen_uni_ex(), PKCS12_key_gen_utf8_ex(),
62 PKCS12_pack_p7encdata_ex(), PKCS12_pbe_crypt_ex(), PKCS12_PBE_keyivgen_ex(),
63 PKCS12_SAFEBAG_create_pkcs8_encrypt_ex(), PKCS5_pbe2_set_iv_ex(),
64 PKCS5_pbe_set0_algor_ex(), PKCS5_pbe_set_ex(), PKCS5_pbkdf2_set_ex(),
65 PKCS5_v2_PBE_keyivgen_ex(), PKCS5_v2_scrypt_keyivgen_ex(),
66 PKCS8_decrypt_ex(), PKCS8_encrypt_ex(), PKCS8_set0_pbe_ex().
68 As part of this change the EVP_PBE_xxx APIs can also accept a library
69 context and property query and will call an extended version of the key/IV
70 derivation function which supports these parameters. This includes
71 EVP_PBE_CipherInit_ex(), EVP_PBE_find_ex() and EVP_PBE_scrypt_ex().
75 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
79 * Added support for Kernel TLS (KTLS). In order to use KTLS, support for it
80 must be compiled in using the "enable-ktls" compile time option. It must
81 also be enabled at run time using the SSL_OP_ENABLE_KTLS option.
83 *Boris Pismenny, John Baldwin and Andrew Gallatin*
85 * Support for RFC 5746 secure renegotiation is now required by default for
86 SSL or TLS connections to succeed. Applications that require the ability
87 to connect to legacy peers will need to explicitly set
88 SSL_OP_LEGACY_SERVER_CONNECT. Accordingly, SSL_OP_LEGACY_SERVER_CONNECT
89 is no longer set as part of SSL_OP_ALL.
93 * The signature of the `copy` functional parameter of the
94 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
95 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
96 the signature of the `pub_decode` functional parameter of the
97 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
98 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
102 * The error return values from some control calls (ctrl) have changed.
103 One significant change is that controls which used to return -2 for
104 invalid inputs, now return -1 indicating a generic error condition instead.
108 * A public key check is now performed during EVP_PKEY_derive_set_peer().
109 Previously DH was internally doing this during EVP_PKEY_derive().
110 To disable this check use EVP_PKEY_derive_set_peer_ex(dh, peer, 0). This
111 may mean that an error can occur in EVP_PKEY_derive_set_peer() rather than
112 during EVP_PKEY_derive().
116 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
117 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
118 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
119 are deprecated. They are not invoked by the OpenSSL library anymore and
120 are replaced by direct checks of the key operation against the key type
121 when the operation is initialized.
125 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
126 more key types including RSA, DSA, ED25519, X25519, ED448 and X448.
127 Previously (in 1.1.1) they would return -2. For key types that do not have
128 parameters then EVP_PKEY_param_check() will always return 1.
130 * The output from numerous "printing" functions such as X509_signature_print(),
131 X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been
132 amended such that there may be cosmetic differences between the output
133 observed in 1.1.1 and 3.0. This also applies to the "-text" output from the
134 x509 and crl applications.
138 * Windows thread synchronization uses read/write primitives (SRWLock) when
139 supported by the OS, otherwise CriticalSection continues to be used.
143 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
144 work on read only BIO source/sinks that do not support these functions.
145 This allows piping or redirection of a file BIO using stdin to be buffered
146 into memory. This is used internally in OSSL_DECODER_from_bio().
150 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
151 this function would return one of the values OSSL_STORE_INFO_NAME,
152 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
153 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
154 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
155 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
156 using this function should be amended to handle the changed return value.
160 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
161 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
162 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
163 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
164 contains more than one certificate identifier: This means that all
165 certificates referenced there MUST be part of the validation chain.
169 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
170 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
171 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
172 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
173 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
174 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
175 using the EVP APIs to access these ciphers should instead use more modern
176 ciphers. If that is not possible then these applications should ensure that
177 the legacy provider has been loaded. This can be achieved either
178 programmatically or via configuration. See the provider(7) man page for
183 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
184 RIPEMD-160 have been moved to the legacy provider. Applications using the
185 EVP APIs to access these digests should instead use more modern digests. If
186 that is not possible then these applications should ensure that the legacy
187 provider has been loaded. This can be achieved either programmatically or via
188 configuration. See the provider(7) man page for further details.
192 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
197 * The deprecated functions EVP_PKEY_get0_RSA(),
198 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
199 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
200 well as the similarly named "get1" functions behave slightly differently in
201 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
202 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
203 provider. Calling these functions will only return a handle on the internal
204 key where the EVP_PKEY was constructed using this key in the first place, for
205 example using a function or macro such as EVP_PKEY_assign_RSA(),
206 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
207 then these functions now return a cached copy of the key. Changes to
208 the internal provider key that take place after the first time the cached key
209 is accessed will not be reflected back in the cached copy. Similarly any
210 changes made to the cached copy by application code will not be reflected
211 back in the internal provider key.
213 For the above reasons the keys returned from these functions should typically
214 be treated as read-only. To emphasise this the value returned from
215 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
216 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
217 break some existing code. Applications broken by this change should be
218 modified. The preferred solution is to refactor the code to avoid the use of
219 these deprecated functions. Failing this the code should be modified to use a
220 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
221 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
222 non-const pointer to enable them to be "freed". However they should also be
223 treated as read-only.
227 * A number of functions handling low level keys or engines were deprecated
228 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
229 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
230 EVP_PKEY_get0_siphash(). Applications using engines should instead use
231 providers. Applications getting or setting low level keys in an EVP_PKEY
232 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
233 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
237 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
238 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
239 and require returning octet ptr parameters from providers that
240 would like to support them which complicates provider implementations.
244 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
245 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
246 RAND_set_rand_method(). Provider based random number generators should
247 be used instead via EVP_RAND(3).
251 * The SRP APIs have been deprecated. The old APIs do not work via providers,
252 and there is no EVP interface to them. Unfortunately there is no replacement
253 for these APIs at this time.
257 * Add a compile time option to prevent the caching of provider fetched
258 algorithms. This is enabled by including the no-cached-fetch option
259 at configuration time.
263 * The default algorithms for pkcs12 creation with the PKCS12_create() function
264 were changed to more modern PBKDF2 and AES based algorithms. The default
265 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
266 with the password-based encryption iteration count. The default digest
267 algorithm for the MAC computation was changed to SHA-256. The pkcs12
268 application now supports -legacy option that restores the previous
269 default algorithms to support interoperability with legacy systems.
271 *Tomáš Mráz and Sahana Prasad*
273 * The openssl speed command does not use low-level API calls anymore. This
274 implies some of the performance numbers might not be fully comparable
275 with the previous releases due to higher overhead. This applies
276 particularly to measuring performance on smaller data chunks.
280 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
283 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
285 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
286 Typically if OpenSSL has no EC or DH algorithms then it cannot support
287 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
288 through providers. Therefore third party providers may supply group
289 implementations even where there are no built-in ones. Attempting to create
290 TLS connections in such a build without also disabling TLSv1.3 at run time or
291 using third party provider groups may result in handshake failures. TLSv1.3
292 can be disabled at compile time using the "no-tls1_3" Configure option.
296 * The undocumented function X509_certificate_type() has been deprecated;
297 applications can use X509_get0_pubkey() and X509_get0_signature() to
298 get the same information.
302 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
303 functions. They are identical to BN_rand() and BN_rand_range()
308 * Removed RSA padding mode for SSLv23 (which was only used for
309 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
310 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
315 * Deprecated the obsolete X9.31 RSA key generation related functions
316 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
317 BN_X931_generate_prime_ex().
321 * The default key generation method for the regular 2-prime RSA keys was
322 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
323 Conditions Based on Auxiliary Probable Primes). This method is slower
324 than the original method.
328 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
329 They are replaced with the BN_check_prime() function that avoids possible
330 misuse and always uses at least 64 rounds of the Miller-Rabin
331 primality test. At least 64 rounds of the Miller-Rabin test are now also
332 used for all prime generation, including RSA key generation.
333 This increases key generation time, especially for larger keys.
337 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
338 as they are not useful with non-deprecated functions.
342 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
343 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
344 OCSP_REQ_CTX_i2d() and its special form OCSP_REQ_CTX_set1_req(),
345 OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
346 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
347 were used to collect all necessary data to form a HTTP request, and to
348 perform the HTTP transfer with that request. With OpenSSL 3.0, the
349 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
350 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
351 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
352 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
353 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
354 OSSL_HTTP_REQ_CTX_set_max_response_length().
356 *Rich Salz and Richard Levitte*
358 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
362 * Validation of SM2 keys has been separated from the validation of regular EC
363 keys, allowing to improve the SM2 validation process to reject loaded private
364 keys that are not conforming to the SM2 ISO standard.
365 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
370 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
371 switches: a validation failure triggers an early exit, returning a failure
372 exit status to the parent process.
376 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
377 to ignore unknown ciphers.
381 * The `-cipher-commands` and `-digest-commands` options
382 of the command line utility `list` have been deprecated.
383 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
387 * All of the low level EC_KEY functions have been deprecated including:
389 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
390 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
391 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
392 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
393 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
394 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
395 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
396 EC_KEY_METHOD_get_verify,
397 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
398 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
399 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
400 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
401 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
402 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
403 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
404 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
405 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
406 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
407 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
408 Applications that need to implement an EC_KEY_METHOD need to consider
409 implementation of the functionality in a special provider.
410 For replacement of the functions manipulating the EC_KEY objects
411 see the EVP_PKEY-EC(7) manual page.
413 Additionally functions that read and write EC_KEY objects such as
414 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
415 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
416 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
417 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
418 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
419 have also been deprecated. Applications should instead use the
420 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
422 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
423 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
424 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
425 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
426 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
428 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
430 * Deprecated all the libcrypto and libssl error string loading
431 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
432 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
433 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
434 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
435 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
436 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
437 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
438 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
439 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
440 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
442 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
443 now loads error strings automatically.
447 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
448 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
449 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
450 are to be used by servers requiring ephemeral DH keys. Instead applications
451 should consider using the built-in DH parameters that are available by
452 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
453 necessary then applications can use the alternative functions
454 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
455 replacement for the "callback" functions. The callback was originally useful
456 in order to have different parameters for export and non-export ciphersuites.
457 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
458 functions should be replaced by one of the other methods described above.
462 * The `-crypt` option to the `passwd` command line tool has been removed.
466 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
471 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
473 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
474 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
475 The inverse ciphers use AES decryption for wrapping, and
476 AES encryption for unwrapping.
480 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
481 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
482 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
483 OpenSSL 3.0 these are replaced by the more generic functions
484 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
485 The old versions have been converted to deprecated macros that just call the
490 * The security callback, which can be customised by application code, supports
491 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
492 in the "other" parameter. In most places this is what is passed. All these
493 places occur server side. However there was one client side call of this
494 security operation and it passed a DH object instead. This is incorrect
495 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
496 of the other locations. Therefore this client side call has been changed to
497 pass an EVP_PKEY instead.
501 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
502 interface. Their functionality remains unchanged.
506 * Added new option for 'openssl list', '-providers', which will display the
507 list of loaded providers, their names, version and status. It optionally
508 displays their gettable parameters.
512 * Removed EVP_PKEY_set_alias_type(). This function was previously
513 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
514 type is internally recognised so the workaround is no longer needed.
516 This is a breaking change from previous OpenSSL versions.
520 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
521 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
525 * Changed all "STACK" functions to be macros instead of inline functions. Macro
526 parameters are still checked for type safety at compile time via helper
531 * Remove the RAND_DRBG API
533 The RAND_DRBG API did not fit well into the new provider concept as
534 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
535 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
536 and some of its API calls are rather low-level. This holds in particular
537 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
539 Adding a compatibility layer to continue supporting the RAND_DRBG API as
540 a legacy API for a regular deprecation period turned out to come at the
541 price of complicating the new provider API unnecessarily. Since the
542 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
545 *Paul Dale and Matthias St. Pierre*
547 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
548 as well as actual hostnames.
552 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
553 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
554 conversely, silently ignore DTLS protocol version bounds when configuring
555 TLS-based contexts. The commands can be repeated to set bounds of both
556 types. The same applies with the corresponding "min_protocol" and
557 "max_protocol" command-line switches, in case some application uses both TLS
560 SSL_CTX instances that are created for a fixed protocol version (e.g.
561 `TLSv1_server_method()`) also silently ignore version bounds. Previously
562 attempts to apply bounds to these protocol versions would result in an
563 error. Now only the "version-flexible" SSL_CTX instances are subject to
564 limits in configuration files in command-line options.
568 * Deprecated the `ENGINE` API. Engines should be replaced with providers
573 * Reworked the recorded ERR codes to make better space for system errors.
574 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
575 given code is a system error (true) or an OpenSSL error (false).
579 * Reworked the test perl framework to better allow parallel testing.
581 *Nicola Tuveri and David von Oheimb*
583 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
584 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
588 * 'Configure' has been changed to figure out the configuration target if
589 none is given on the command line. Consequently, the 'config' script is
590 now only a mere wrapper. All documentation is changed to only mention
593 *Rich Salz and Richard Levitte*
595 * Added a library context `OSSL_LIB_CTX` that applications as well as
596 other libraries can use to form a separate context within which
597 libcrypto operations are performed.
599 There are two ways this can be used:
601 - Directly, by passing a library context to functions that take
602 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
604 - Indirectly, by creating a new library context and then assigning
605 it as the new default, with `OSSL_LIB_CTX_set0_default`.
607 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
608 apart from the functions directly related to `OSSL_LIB_CTX`, accept
609 NULL to indicate that the default library context should be used.
611 Library code that changes the default library context using
612 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
613 second call before returning to the caller.
615 _(Note: the library context was initially called `OPENSSL_CTX` and
616 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
620 * Handshake now fails if Extended Master Secret extension is dropped
625 * Dropped interactive mode from the `openssl` program. From now on,
626 running it without arguments is equivalent to `openssl help`.
630 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()` since their
631 return values were confusing: Unlike other `_cmp()` functions
632 they do not return 0 when their arguments are equal.
633 The new replacement functions `EVP_PKEY_eq()` and `EVP_PKEY_parameters_eq()`
636 *David von Oheimb and Shane Lontis*
638 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
639 `EC_GROUP_get_field_type()`.
643 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
644 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
645 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
646 Applications should rely on the library automatically assigning a suitable
647 EC_METHOD internally upon EC_GROUP construction.
651 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
652 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
653 assigned internally without application intervention.
654 Users of EC_GROUP_new() should switch to a different suitable constructor.
658 * Add CAdES-BES signature verification support, mostly derived
659 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
661 *Filipe Raimundo da Silva*
663 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
667 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
668 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
669 is to support encryption and decryption of a digital envelope that is both
670 authenticated and encrypted using AES GCM mode.
674 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
675 functions are not widely used and now OpenSSL automatically perform this
676 conversion when needed.
680 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
681 EC_KEY_precompute_mult(). These functions are not widely used and
682 applications should instead switch to named curves which OpenSSL has
683 hardcoded lookup tables for.
687 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
688 should instead use the L<EC_POINT_mul(3)> function.
692 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
693 that are not applicable to the new provider model. Applications should
694 instead use EVP_default_properties_is_fips_enabled() and
695 EVP_default_properties_enable_fips().
699 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
700 is set, an unexpected EOF is ignored, it pretends a close notify was received
701 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
705 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
706 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
707 used and applications should instead use the
708 L<EC_POINT_set_affine_coordinates(3)> and
709 L<EC_POINT_get_affine_coordinates(3)> functions.
713 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
714 arrays to be more easily constructed via a series of utility functions.
715 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
716 the various push functions and finally convert to a passable OSSL_PARAM
717 array using OSSL_PARAM_BLD_to_param().
721 * The security strength of SHA1 and MD5 based signatures in TLS has been
722 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
723 working at the default security level of 1 and instead requires security
724 level 0. The security level can be changed either using the cipher string
725 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
726 that where the signature algorithms extension is missing from a ClientHello
727 then the handshake will fail in TLS 1.2 at security level 1. This is because,
728 although this extension is optional, failing to provide one means that
729 OpenSSL will fallback to a default set of signature algorithms. This default
730 set requires the availability of SHA1.
734 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
735 contain a provider side internal key.
739 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
740 They are old functions that we don't use, and that you could disable with
741 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
745 * Project text documents not yet having a proper file name extension
746 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
747 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
751 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
752 have been converted to Markdown with the goal to produce documents
753 which not only look pretty when viewed online in the browser, but
754 remain well readable inside a plain text editor.
756 To achieve this goal, a 'minimalistic' Markdown style has been applied
757 which avoids formatting elements that interfere too much with the
758 reading flow in the text file. For example, it
760 * avoids [ATX headings][] and uses [setext headings][] instead
761 (which works for `<h1>` and `<h2>` headings only).
762 * avoids [inline links][] and uses [reference links][] instead.
763 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
765 [ATX headings]: https://github.github.com/gfm/#atx-headings
766 [setext headings]: https://github.github.com/gfm/#setext-headings
767 [inline links]: https://github.github.com/gfm/#inline-link
768 [reference links]: https://github.github.com/gfm/#reference-link
769 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
770 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
772 *Matthias St. Pierre*
774 * The test suite is changed to preserve results of each test recipe.
775 A new directory test-runs/ with subdirectories named like the
776 test recipes are created in the build tree for this purpose.
780 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
781 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
782 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
784 *David von Oheimb, Martin Peylo*
786 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
787 The legacy OCSP-focused and only partly documented API is retained for
788 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
792 * Added `util/check-format.pl`, a tool for checking adherence to the
793 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
794 The checks performed are incomplete and yield some false positives.
795 Still the tool should be useful for detecting most typical glitches.
799 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
800 If domain name resolution yields multiple IP addresses all of them are tried
801 after `connect()` failures.
805 * All of the low level RSA functions have been deprecated including:
807 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
808 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
809 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
810 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
811 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
812 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
813 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
814 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
815 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
816 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
817 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
818 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
819 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
820 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
821 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
822 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
823 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
824 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
825 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
826 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
827 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
828 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
829 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
830 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
831 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
832 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
833 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
834 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
836 Use of these low level functions has been informally discouraged for a long
837 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
838 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
839 L<EVP_PKEY_decrypt(3)>.
841 All of these low level RSA functions have been deprecated without
844 RSA_blinding_off, RSA_blinding_on, RSA_clear_flags, RSA_get_version,
845 RSAPrivateKey_dup, RSAPublicKey_dup, RSA_set_flags, RSA_setup_blinding and
848 All of these RSA flags have been deprecated without replacement:
850 RSA_FLAG_BLINDING, RSA_FLAG_CACHE_PRIVATE, RSA_FLAG_CACHE_PUBLIC,
851 RSA_FLAG_EXT_PKEY, RSA_FLAG_NO_BLINDING, RSA_FLAG_THREAD_SAFE and
852 RSA_METHOD_FLAG_NO_CHECK.
856 * X509 certificates signed using SHA1 are no longer allowed at security
858 In TLS/SSL the default security level is 1. It can be set either
859 using the cipher string with `@SECLEVEL`, or calling
860 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
861 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
863 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
864 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
865 options of the commands.
869 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
870 modified to use PKEY APIs. These commands are now in maintenance mode
871 and no new features will be added to them.
875 * The command line utility rsautl has been deprecated.
876 Instead use the pkeyutl program.
880 * The command line utilities genrsa and rsa have been modified to use PKEY
881 APIs. They now write PKCS#8 keys by default. These commands are now in
882 maintenance mode and no new features will be added to them.
886 * All of the low level DH functions have been deprecated including:
888 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
889 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
890 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
891 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
892 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
893 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
894 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
895 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
896 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
897 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
898 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
899 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
900 DH_meth_get_generate_params and DH_meth_set_generate_params.
902 Use of these low level functions has been informally discouraged for a long
903 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
904 and L<EVP_PKEY_derive(3)>.
906 These low level DH functions have been deprecated without replacement:
908 DH_clear_flags, DH_get_1024_160, DH_get_2048_224, DH_get_2048_256,
909 DH_set_flags and DH_test_flags.
911 The DH_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
912 The DH_FLAG_TYPE_DH and DH_FLAG_TYPE_DHX have been deprecated. Use
913 EVP_PKEY_is_a() to determine the type of a key. There is no replacement for
916 Additionally functions that read and write DH objects such as d2i_DHparams,
917 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
918 functions have also been deprecated. Applications should instead use the
919 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
921 Finally functions that assign or obtain DH objects from an EVP_PKEY such as
922 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
923 `EVP_PKEY_set1_DH()` are also deprecated.
924 Applications should instead either read or write an
925 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
926 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
928 *Paul Dale and Matt Caswell*
930 * All of the low level DSA functions have been deprecated including:
932 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
933 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
934 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
935 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
936 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
937 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
938 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
939 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
940 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
941 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
942 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
943 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
944 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
945 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
946 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
947 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
948 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
950 Use of these low level functions has been informally discouraged for a long
951 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
952 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
954 These low level DSA functions have been deprecated without replacement:
956 DSA_clear_flags, DSA_dup_DH, DSAparams_dup, DSA_set_flags and
959 The DSA_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
961 Finally functions that assign or obtain DSA objects from an EVP_PKEY such as
962 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
963 `EVP_PKEY_set1_DSA()` are also deprecated.
964 Applications should instead either read or write an
965 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
966 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
970 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
971 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. This is a breaking
972 change from previous OpenSSL versions.
974 Unlike in previous OpenSSL versions, this means that applications must not
975 call `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
976 The `EVP_PKEY_set_alias_type` function has now been removed.
978 Parameter and key generation is also reworked to make it possible
979 to generate EVP_PKEY_SM2 parameters and keys. Applications must now generate
980 SM2 keys directly and must not create an EVP_PKEY_EC key first.
984 * Deprecated low level ECDH and ECDSA functions. These include:
986 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
987 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
990 Use of these low level functions has been informally discouraged for a long
991 time. Instead applications should use the EVP_PKEY_derive(3),
992 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
996 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
997 and EVP_PKEY_decrypt() instead.
998 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
999 and EVP_PKEY_encrypt() instead.
1003 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
1004 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
1005 a new formulation to include all the things it can be used for,
1006 as well as words of caution.
1010 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1011 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
1015 * All of the low level HMAC functions have been deprecated including:
1017 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
1018 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
1019 and HMAC_CTX_get_md.
1021 Use of these low level functions has been informally discouraged for a long
1022 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1023 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1024 and L<EVP_MAC_final(3)>.
1028 * Over two thousand fixes were made to the documentation, including:
1029 - Common options (such as -rand/-writerand, TLS version control, etc)
1030 were refactored and point to newly-enhanced descriptions in openssl.pod.
1031 - Added style conformance for all options (with help from Richard Levitte),
1032 documented all reported missing options, added a CI build to check
1033 that all options are documented and that no unimplemented options
1035 - Documented some internals, such as all use of environment variables.
1036 - Addressed all internal broken L<> references.
1040 * All of the low level CMAC functions have been deprecated including:
1042 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
1043 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
1045 Use of these low level functions has been informally discouraged for a long
1046 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1047 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1048 and L<EVP_MAC_final(3)>.
1052 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
1053 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
1056 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
1057 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
1058 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
1059 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
1060 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
1061 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
1062 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
1063 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
1064 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
1065 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
1067 Use of these low level functions has been informally discouraged
1068 for a long time. Applications should use the EVP_DigestInit_ex(3),
1069 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
1073 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1074 set of functions. The documentation mentioned negative values for some
1075 errors, but this was never the case, so the mention of negative values
1078 Code that followed the documentation and thereby check with something
1079 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1083 * All of the low level cipher functions have been deprecated including:
1085 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
1086 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
1087 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
1088 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
1089 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
1090 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
1091 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
1092 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
1093 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
1094 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
1095 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
1096 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
1097 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
1098 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
1099 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
1100 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
1101 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
1102 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
1103 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
1104 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
1105 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
1106 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
1107 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
1108 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
1109 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
1110 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
1111 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
1112 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
1113 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
1115 Use of these low level functions has been informally discouraged for
1116 a long time. Applications should use the high level EVP APIs, e.g.
1117 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
1118 equivalently named decrypt functions instead.
1120 *Matt Caswell and Paul Dale*
1122 * Removed include/openssl/opensslconf.h.in and replaced it with
1123 include/openssl/configuration.h.in, which differs in not including
1124 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1125 was added to include both.
1127 This allows internal hacks where one might need to modify the set
1128 of configured macros, for example this if deprecated symbols are
1129 still supposed to be available internally:
1131 #include <openssl/configuration.h>
1133 #undef OPENSSL_NO_DEPRECATED
1134 #define OPENSSL_SUPPRESS_DEPRECATED
1136 #include <openssl/macros.h>
1138 This should not be used by applications that use the exported
1139 symbols, as that will lead to linking errors.
1143 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1144 used in exponentiation with 512-bit moduli. No EC algorithms are
1145 affected. Analysis suggests that attacks against 2-prime RSA1024,
1146 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1147 difficult to perform and are not believed likely. Attacks against DH512
1148 are considered just feasible. However, for an attack the target would
1149 have to re-use the DH512 private key, which is not recommended anyway.
1150 Also applications directly using the low level API BN_mod_exp may be
1151 affected if they use BN_FLG_CONSTTIME.
1156 * Most memory-debug features have been deprecated, and the functionality
1157 replaced with no-ops.
1161 * Added documentation for the STACK API.
1165 * Introduced a new method type and API, OSSL_ENCODER, to
1166 represent generic encoders. An implementation is expected to
1167 be able to encode an object associated with a given name (such
1168 as an algorithm name for an asymmetric key) into forms given by
1169 implementation properties.
1171 Encoders are primarily used from inside libcrypto, through
1172 calls to functions like EVP_PKEY_print_private(),
1173 PEM_write_bio_PrivateKey() and similar.
1175 Encoders are specified in such a way that they can be made to
1176 directly handle the provider side portion of an object, if this
1177 provider side part comes from the same provider as the encoder
1178 itself, but can also be made to handle objects in parametrized
1179 form (as an OSSL_PARAM array of data). This allows a provider to
1180 offer generic encoders as a service for any other provider.
1184 * Added a .pragma directive to the syntax of configuration files, to
1185 allow varying behavior in a supported and predictable manner.
1186 Currently added pragma:
1190 This allows dollar signs to be a keyword character unless it's
1191 followed by a opening brace or parenthesis. This is useful for
1192 platforms where dollar signs are commonly used in names, such as
1193 volume names and system directory names on VMS.
1197 * Added functionality to create an EVP_PKEY from user data. This
1198 is effectively the same as creating a RSA, DH or DSA object and
1199 then assigning them to an EVP_PKEY, but directly using algorithm
1200 agnostic EVP functions. A benefit is that this should be future
1201 proof for public key algorithms to come.
1205 * Change the interpretation of the '--api' configuration option to
1206 mean that this is a desired API compatibility level with no
1207 further meaning. The previous interpretation, that this would
1208 also mean to remove all deprecated symbols up to and including
1209 the given version, no requires that 'no-deprecated' is also used
1210 in the configuration.
1212 When building applications, the desired API compatibility level
1213 can be set with the OPENSSL_API_COMPAT macro like before. For
1214 API compatibility version below 3.0, the old style numerical
1215 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1216 For version 3.0 and on, the value is expected to be the decimal
1217 value calculated from the major and minor version like this:
1219 MAJOR * 10000 + MINOR * 100
1223 -DOPENSSL_API_COMPAT=30000 For 3.0
1224 -DOPENSSL_API_COMPAT=30200 For 3.2
1226 To hide declarations that are deprecated up to and including the
1227 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1228 given when building the application as well.
1232 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1233 access to certificate and CRL stores via URIs and OSSL_STORE
1236 This adds the following functions:
1238 - X509_LOOKUP_store()
1239 - X509_STORE_load_file()
1240 - X509_STORE_load_path()
1241 - X509_STORE_load_store()
1242 - SSL_add_store_cert_subjects_to_stack()
1243 - SSL_CTX_set_default_verify_store()
1244 - SSL_CTX_load_verify_file()
1245 - SSL_CTX_load_verify_dir()
1246 - SSL_CTX_load_verify_store()
1250 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1251 The presence of this system service is determined at run-time.
1255 * Added functionality to create an EVP_PKEY context based on data
1256 for methods from providers. This takes an algorithm name and a
1257 property query string and simply stores them, with the intent
1258 that any operation that uses this context will use those strings
1259 to fetch the needed methods implicitly, thereby making the port
1260 of application written for pre-3.0 OpenSSL easier.
1264 * The undocumented function NCONF_WIN32() has been deprecated; for
1265 conversion details see the HISTORY section of doc/man5/config.pod
1269 * Introduced the new functions EVP_DigestSignInit_ex() and
1270 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1271 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1272 pages for further details.
1276 * Over two thousand fixes were made to the documentation, including:
1277 adding missing command flags, better style conformance, documentation
1280 *Rich Salz, Richard Levitte*
1282 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1283 X25519, X448, Ed25519 and Ed448.
1287 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1292 * Deprecated the public definition of `ERR_STATE` as well as the function
1293 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1298 * Added ERR functionality to give callers access to the stored function
1299 names that have replaced the older function code based functions.
1301 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1302 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1303 ERR_peek_error_all() and ERR_peek_last_error_all().
1305 These functions have become deprecated: ERR_get_error_line(),
1306 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1307 ERR_peek_last_error_line_data() and ERR_func_error_string().
1309 Users are recommended to use ERR_get_error_all(), or to pick information
1310 with ERR_peek functions and finish off with getting the error code by using
1315 * Extended testing to be verbose for failing tests only. The make variables
1316 VERBOSE_FAILURE or VF can be used to enable this:
1318 $ make VF=1 test # Unix
1319 $ mms /macro=(VF=1) test ! OpenVMS
1320 $ nmake VF=1 test # Windows
1324 * Added the `-copy_extensions` option to the `x509` command for use with
1325 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1326 all extensions in the request are copied to the certificate or vice versa.
1328 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1330 * Added the `-copy_extensions` option to the `req` command for use with
1331 `-x509`. When given with the `copy` or `copyall` argument,
1332 all extensions in the certification request are copied to the certificate.
1336 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1337 they generate are by default RFC 5280 compliant in the following sense:
1338 There is a subjectKeyIdentifier extension with a hash value of the public key
1339 and for not self-signed certs there is an authorityKeyIdentifier extension
1340 with a keyIdentifier field or issuer information identifying the signing key.
1341 This is done unless some configuration overrides the new default behavior,
1342 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1346 * Added several checks to `X509_verify_cert()` according to requirements in
1347 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1348 (which may be done by using the CLI option `-x509_strict`):
1349 * The basicConstraints of CA certificates must be marked critical.
1350 * CA certificates must explicitly include the keyUsage extension.
1351 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1352 * The issuer name of any certificate must not be empty.
1353 * The subject name of CA certs, certs with keyUsage crlSign,
1354 and certs without subjectAlternativeName must not be empty.
1355 * If a subjectAlternativeName extension is given it must not be empty.
1356 * The signatureAlgorithm field and the cert signature must be consistent.
1357 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1358 must not be marked critical.
1359 * The authorityKeyIdentifier must be given for X.509v3 certs
1360 unless they are self-signed.
1361 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1365 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1366 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1370 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1371 used even when parsing explicit parameters, when loading a encoded key
1372 or calling `EC_GROUP_new_from_ecpkparameters()`/
1373 `EC_GROUP_new_from_ecparameters()`.
1374 This prevents bypass of security hardening and performance gains,
1375 especially for curves with specialized EC_METHODs.
1376 By default, if a key encoded with explicit parameters is loaded and later
1377 encoded, the output is still encoded with explicit parameters, even if
1378 internally a "named" EC_GROUP is used for computation.
1382 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1383 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1384 NULL. After this change, only the cofactor parameter can be NULL. It also
1385 does some minimal sanity checks on the passed order.
1390 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1391 An attack is simple, if the first CMS_recipientInfo is valid but the
1392 second CMS_recipientInfo is chosen ciphertext. If the second
1393 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1394 encryption key will be replaced by garbage, and the message cannot be
1395 decoded, but if the RSA decryption fails, the correct encryption key is
1396 used and the recipient will not notice the attack.
1397 As a work around for this potential attack the length of the decrypted
1398 key must be equal to the cipher default key length, in case the
1399 certifiate is not given and all recipientInfo are tried out.
1400 The old behaviour can be re-enabled in the CMS code by setting the
1401 CMS_DEBUG_DECRYPT flag.
1405 * Early start up entropy quality from the DEVRANDOM seed source has been
1406 improved for older Linux systems. The RAND subsystem will wait for
1407 /dev/random to be producing output before seeding from /dev/urandom.
1408 The seeded state is stored for future library initialisations using
1409 a system global shared memory segment. The shared memory identifier
1410 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1411 the desired value. The default identifier is 114.
1415 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1416 when primes for RSA keys are computed.
1417 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1418 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1419 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1420 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1421 This avoids possible fingerprinting of newly generated RSA modules.
1425 * Correct the extended master secret constant on EBCDIC systems. Without this
1426 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1427 negotiate EMS will fail. Unfortunately this also means that TLS connections
1428 between EBCDIC systems with this fix, and EBCDIC systems without this
1429 fix will fail if they negotiate EMS.
1433 * Changed the library initialisation so that the config file is now loaded
1434 by default. This was already the case for libssl. It now occurs for both
1435 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1436 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1440 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1441 where the former acts as a replacement for `ERR_put_error()`, and the
1442 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1443 `ERR_raise_data()` adds more flexibility by taking a format string and
1444 an arbitrary number of arguments following it, to be processed with
1449 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1450 to check if a named provider is loaded and available. When called, it
1451 will also activate all fallback providers if such are still present.
1455 * Enforce a minimum DH modulus size of 512 bits.
1459 * Changed DH parameters to generate the order q subgroup instead of 2q.
1460 Previously generated DH parameters are still accepted by DH_check
1461 but DH_generate_key works around that by clearing bit 0 of the
1462 private key for those. This avoids leaking bit 0 of the private key.
1466 * Significantly reduce secure memory usage by the randomness pools.
1470 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1475 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1476 algorithms. An implementation of a key exchange algorithm can be obtained
1477 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1478 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1479 the older EVP_PKEY_derive_init() function. See the man pages for the new
1480 functions for further details.
1484 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1488 * Removed the function names from error messages and deprecated the
1491 * Removed NextStep support and the macro OPENSSL_UNISTD
1495 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1496 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1497 Also removed "export var as function" capability; we do not export
1498 variables, only functions.
1502 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1503 an error and 1 indicating success. In previous versions of OpenSSL this
1504 was a void type. If a key was set longer than the maximum possible this
1509 * Support SM2 signing and verification schemes with X509 certificate.
1513 * Use SHA256 as the default digest for TS query in the `ts` app.
1517 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1518 This checks that the salt length is at least 128 bits, the derived key
1519 length is at least 112 bits, and that the iteration count is at least 1000.
1520 For backwards compatibility these checks are disabled by default in the
1521 default provider, but are enabled by default in the fips provider.
1522 To enable or disable these checks use the control
1523 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1527 * Default cipher lists/suites are now available via a function, the
1528 #defines are deprecated.
1532 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1533 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1534 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1538 * Join the directories crypto/x509 and crypto/x509v3
1542 * Added command 'openssl kdf' that uses the EVP_KDF API.
1546 * Added command 'openssl mac' that uses the EVP_MAC API.
1550 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1551 as default directories. Also added the command 'openssl info'
1552 for scripting purposes.
1556 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1557 deprecated. These undocumented functions were never integrated into the EVP
1558 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1559 Bi-directional IGE mode. These modes were never formally standardised and
1560 usage of these functions is believed to be very small. In particular
1561 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1562 is ever used. The security implications are believed to be minimal, but
1563 this issue was never fixed for backwards compatibility reasons. New code
1564 should not use these modes.
1568 * Add prediction resistance to the DRBG reseeding process.
1572 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1573 mandated by IEEE Std 1619-2018.
1577 * Added newline escaping functionality to a filename when using openssl dgst.
1578 This output format is to replicate the output format found in the `*sum`
1579 checksum programs. This aims to preserve backward compatibility.
1581 *Matt Eaton, Richard Levitte, and Paul Dale*
1583 * Removed the heartbeat message in DTLS feature, as it has very
1584 little usage and doesn't seem to fulfill a valuable purpose.
1585 The configuration option is now deprecated.
1589 * Changed the output of 'openssl {digestname} < file' to display the
1590 digest name in its output.
1594 * Added a new generic trace API which provides support for enabling
1595 instrumentation through trace output. This feature is mainly intended
1596 as an aid for developers and is disabled by default. To utilize it,
1597 OpenSSL needs to be configured with the `enable-trace` option.
1599 If the tracing API is enabled, the application can activate trace output
1600 by registering BIOs as trace channels for a number of tracing and debugging
1603 The `openssl` program has been expanded to enable any of the types
1604 available via environment variables defined by the user, and serves as
1605 one possible example on how to use this functionality.
1607 *Richard Levitte & Matthias St. Pierre*
1609 * Added build tests for C++. These are generated files that only do one
1610 thing, to include one public OpenSSL head file each. This tests that
1611 the public header files can be usefully included in a C++ application.
1613 This test isn't enabled by default. It can be enabled with the option
1614 'enable-buildtest-c++'.
1618 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1622 * Add KMAC to EVP_MAC.
1626 * Added property based algorithm implementation selection framework to
1631 * Added SCA hardening for modular field inversion in EC_GROUP through
1632 a new dedicated field_inv() pointer in EC_METHOD.
1633 This also addresses a leakage affecting conversions from projective
1634 to affine coordinates.
1636 *Billy Bob Brumley, Nicola Tuveri*
1638 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1639 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1640 those algorithms that were already supported through the EVP_PKEY API
1641 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1642 and scrypt are now wrappers that call EVP_KDF.
1646 * Build devcrypto engine as a dynamic engine.
1648 *Eneas U de Queiroz*
1650 * Add keyed BLAKE2 to EVP_MAC.
1654 * Fix a bug in the computation of the endpoint-pair shared secret used
1655 by DTLS over SCTP. This breaks interoperability with older versions
1656 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1657 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1658 interoperability with such broken implementations. However, enabling
1659 this switch breaks interoperability with correct implementations.
1661 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1662 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1666 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1670 * Changed the license to the Apache License v2.0.
1674 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1676 - Major releases (indicated by incrementing the MAJOR release number)
1677 may introduce incompatible API/ABI changes.
1678 - Minor releases (indicated by incrementing the MINOR release number)
1679 may introduce new features but retain API/ABI compatibility.
1680 - Patch releases (indicated by incrementing the PATCH number)
1681 are intended for bug fixes and other improvements of existing
1682 features only (like improving performance or adding documentation)
1683 and retain API/ABI compatibility.
1687 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1691 * Remove the 'dist' target and add a tarball building script. The
1692 'dist' target has fallen out of use, and it shouldn't be
1693 necessary to configure just to create a source distribution.
1697 * Recreate the OS390-Unix config target. It no longer relies on a
1698 special script like it did for OpenSSL pre-1.1.0.
1702 * Instead of having the source directories listed in Configure, add
1703 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1708 * Add GMAC to EVP_MAC.
1712 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1716 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1717 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1718 to facilitate the continued use of MACs through raw private keys in
1719 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1723 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1724 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1728 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1729 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1730 are retained for backwards compatibility.
1734 * AES-XTS mode now enforces that its two keys are different to mitigate
1735 the attacked described in "Efficient Instantiations of Tweakable
1736 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1737 Details of this attack can be obtained from:
1738 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1742 * Rename the object files, i.e. give them other names than in previous
1743 versions. Their names now include the name of the final product, as
1744 well as its type mnemonic (bin, lib, shlib).
1748 * Added new option for 'openssl list', '-objects', which will display the
1749 list of built in objects, i.e. OIDs with names.
1753 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1754 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1759 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1760 improves application performance by removing data copies and providing
1761 applications with zero-copy system calls such as sendfile and splice.
1765 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1766 option is set, openssl cleanses (zeroize) plaintext bytes from
1767 internal buffers after delivering them to the application. Note,
1768 the application is still responsible for cleansing other copies
1769 (e.g.: data received by SSL_read(3)).
1773 * `PKCS12_parse` now maintains the order of the parsed certificates
1774 when outputting them via `*ca` (rather than reversing it).
1778 * Deprecated pthread fork support methods. These were unused so no
1779 replacement is required.
1781 - OPENSSL_fork_prepare()
1782 - OPENSSL_fork_parent()
1783 - OPENSSL_fork_child()
1787 * Added support for FFDHE key exchange in TLS 1.3.
1794 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1796 * Fixed a problem with verifying a certificate chain when using the
1797 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1798 the certificates present in a certificate chain. It is not set by default.
1800 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1801 the chain that have explicitly encoded elliptic curve parameters was added
1802 as an additional strict check.
1804 An error in the implementation of this check meant that the result of a
1805 previous check to confirm that certificates in the chain are valid CA
1806 certificates was overwritten. This effectively bypasses the check
1807 that non-CA certificates must not be able to issue other certificates.
1809 If a "purpose" has been configured then there is a subsequent opportunity
1810 for checks that the certificate is a valid CA. All of the named "purpose"
1811 values implemented in libcrypto perform this check. Therefore, where
1812 a purpose is set the certificate chain will still be rejected even when the
1813 strict flag has been used. A purpose is set by default in libssl client and
1814 server certificate verification routines, but it can be overridden or
1815 removed by an application.
1817 In order to be affected, an application must explicitly set the
1818 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1819 for the certificate verification or, in the case of TLS client or server
1820 applications, override the default purpose.
1825 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1826 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1827 renegotiation ClientHello omits the signature_algorithms extension (where it
1828 was present in the initial ClientHello), but includes a
1829 signature_algorithms_cert extension then a NULL pointer dereference will
1830 result, leading to a crash and a denial of service attack.
1832 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1833 (which is the default configuration). OpenSSL TLS clients are not impacted by
1837 *Peter Kästle and Samuel Sapalski*
1839 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1841 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1842 create a unique hash value based on the issuer and serial number data
1843 contained within an X509 certificate. However it was failing to correctly
1844 handle any errors that may occur while parsing the issuer field (which might
1845 occur if the issuer field is maliciously constructed). This may subsequently
1846 result in a NULL pointer deref and a crash leading to a potential denial of
1852 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1853 padding mode to correctly check for rollback attacks. This is considered a
1854 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1859 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1860 functions. Previously they could overflow the output length argument in some
1861 cases where the input length is close to the maximum permissable length for
1862 an integer on the platform. In such cases the return value from the function
1863 call would be 1 (indicating success), but the output length value would be
1864 negative. This could cause applications to behave incorrectly or crash.
1869 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1870 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1871 could be exploited in a side channel attack to recover the password. Since
1872 the attack is local host only this is outside of the current OpenSSL
1873 threat model and therefore no CVE is assigned.
1875 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1880 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1882 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1883 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1884 If an attacker can control both items being compared then this could lead
1885 to a possible denial of service attack. OpenSSL itself uses the
1886 GENERAL_NAME_cmp function for two purposes:
1887 1) Comparing CRL distribution point names between an available CRL and a
1888 CRL distribution point embedded in an X509 certificate
1889 2) When verifying that a timestamp response token signer matches the
1890 timestamp authority name (exposed via the API functions
1891 TS_RESP_verify_response and TS_RESP_verify_token)
1896 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1898 * Certificates with explicit curve parameters are now disallowed in
1899 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1903 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1904 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1905 conversely, silently ignore DTLS protocol version bounds when configuring
1906 TLS-based contexts. The commands can be repeated to set bounds of both
1907 types. The same applies with the corresponding "min_protocol" and
1908 "max_protocol" command-line switches, in case some application uses both TLS
1911 SSL_CTX instances that are created for a fixed protocol version (e.g.
1912 TLSv1_server_method()) also silently ignore version bounds. Previously
1913 attempts to apply bounds to these protocol versions would result in an
1914 error. Now only the "version-flexible" SSL_CTX instances are subject to
1915 limits in configuration files in command-line options.
1919 * Handshake now fails if Extended Master Secret extension is dropped
1924 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1926 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1928 * Fixed segmentation fault in SSL_check_chain()
1929 Server or client applications that call the SSL_check_chain() function
1930 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1931 dereference as a result of incorrect handling of the
1932 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1933 or unrecognised signature algorithm is received from the peer. This could
1934 be exploited by a malicious peer in a Denial of Service attack.
1939 * Added AES consttime code for no-asm configurations
1940 an optional constant time support for AES was added
1941 when building openssl for no-asm.
1942 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1943 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1944 At this time this feature is by default disabled.
1945 It will be enabled by default in 3.0.
1949 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1951 * Revert the change of EOF detection while reading in libssl to avoid
1952 regressions in applications depending on the current way of reporting
1953 the EOF. As the existing method is not fully accurate the change to
1954 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1955 branch and will be present in the 3.0 release.
1959 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1960 when primes for RSA keys are computed.
1961 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1962 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1963 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1964 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1965 This avoids possible fingerprinting of newly generated RSA modules.
1969 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1971 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1972 while reading in libssl then we would report an error back to the
1973 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1974 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1975 therefore give a hint as to what went wrong.
1979 * Check that ed25519 and ed448 are allowed by the security level. Previously
1980 signature algorithms not using an MD were not being checked that they were
1981 allowed by the security level.
1985 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1986 was not quite right. The behaviour was not consistent between resumption
1987 and normal handshakes, and also not quite consistent with historical
1988 behaviour. The behaviour in various scenarios has been clarified and
1989 it has been updated to make it match historical behaviour as closely as
1994 * *[VMS only]* The header files that the VMS compilers include automatically,
1995 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1996 that the C++ compiler doesn't understand. This is a shortcoming in the
1997 compiler, but can be worked around with `__cplusplus` guards.
1999 C++ applications that use OpenSSL libraries must be compiled using the
2000 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2001 functions. Otherwise, only functions with symbols of less than 31
2002 characters can be used, as the linker will not be able to successfully
2003 resolve symbols with longer names.
2007 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2008 The presence of this system service is determined at run-time.
2012 * Added newline escaping functionality to a filename when using openssl dgst.
2013 This output format is to replicate the output format found in the `*sum`
2014 checksum programs. This aims to preserve backward compatibility.
2016 *Matt Eaton, Richard Levitte, and Paul Dale*
2018 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2023 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2025 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2026 number generator (RNG). This was intended to include protection in the
2027 event of a fork() system call in order to ensure that the parent and child
2028 processes did not share the same RNG state. However this protection was not
2029 being used in the default case.
2031 A partial mitigation for this issue is that the output from a high
2032 precision timer is mixed into the RNG state so the likelihood of a parent
2033 and child process sharing state is significantly reduced.
2035 If an application already calls OPENSSL_init_crypto() explicitly using
2036 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2039 *Matthias St. Pierre*
2041 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2042 used even when parsing explicit parameters, when loading a encoded key
2043 or calling `EC_GROUP_new_from_ecpkparameters()`/
2044 `EC_GROUP_new_from_ecparameters()`.
2045 This prevents bypass of security hardening and performance gains,
2046 especially for curves with specialized EC_METHODs.
2047 By default, if a key encoded with explicit parameters is loaded and later
2048 encoded, the output is still encoded with explicit parameters, even if
2049 internally a "named" EC_GROUP is used for computation.
2053 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2054 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2055 NULL. After this change, only the cofactor parameter can be NULL. It also
2056 does some minimal sanity checks on the passed order.
2061 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2062 An attack is simple, if the first CMS_recipientInfo is valid but the
2063 second CMS_recipientInfo is chosen ciphertext. If the second
2064 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2065 encryption key will be replaced by garbage, and the message cannot be
2066 decoded, but if the RSA decryption fails, the correct encryption key is
2067 used and the recipient will not notice the attack.
2068 As a work around for this potential attack the length of the decrypted
2069 key must be equal to the cipher default key length, in case the
2070 certifiate is not given and all recipientInfo are tried out.
2071 The old behaviour can be re-enabled in the CMS code by setting the
2072 CMS_DEBUG_DECRYPT flag.
2077 * Early start up entropy quality from the DEVRANDOM seed source has been
2078 improved for older Linux systems. The RAND subsystem will wait for
2079 /dev/random to be producing output before seeding from /dev/urandom.
2080 The seeded state is stored for future library initialisations using
2081 a system global shared memory segment. The shared memory identifier
2082 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2083 the desired value. The default identifier is 114.
2087 * Correct the extended master secret constant on EBCDIC systems. Without this
2088 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2089 negotiate EMS will fail. Unfortunately this also means that TLS connections
2090 between EBCDIC systems with this fix, and EBCDIC systems without this
2091 fix will fail if they negotiate EMS.
2095 * Use Windows installation paths in the mingw builds
2097 Mingw isn't a POSIX environment per se, which means that Windows
2098 paths should be used for installation.
2103 * Changed DH_check to accept parameters with order q and 2q subgroups.
2104 With order 2q subgroups the bit 0 of the private key is not secret
2105 but DH_generate_key works around that by clearing bit 0 of the
2106 private key for those. This avoids leaking bit 0 of the private key.
2110 * Significantly reduce secure memory usage by the randomness pools.
2114 * Revert the DEVRANDOM_WAIT feature for Linux systems
2116 The DEVRANDOM_WAIT feature added a select() call to wait for the
2117 /dev/random device to become readable before reading from the
2118 /dev/urandom device.
2120 It turned out that this change had negative side effects on
2121 performance which were not acceptable. After some discussion it
2122 was decided to revert this feature and leave it up to the OS
2123 resp. the platform maintainer to ensure a proper initialization
2124 during early boot time.
2126 *Matthias St. Pierre*
2128 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2130 * Add build tests for C++. These are generated files that only do one
2131 thing, to include one public OpenSSL head file each. This tests that
2132 the public header files can be usefully included in a C++ application.
2134 This test isn't enabled by default. It can be enabled with the option
2135 'enable-buildtest-c++'.
2139 * Enable SHA3 pre-hashing for ECDSA and DSA.
2143 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2144 This changes the size when using the `genpkey` command when no size is given.
2145 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2146 generation commands to use 2048 bits by default.
2150 * Reorganize the manual pages to consistently have RETURN VALUES,
2151 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2152 util/fix-doc-nits accordingly.
2154 *Paul Yang, Joshua Lock*
2156 * Add the missing accessor EVP_PKEY_get0_engine()
2160 * Have commands like `s_client` and `s_server` output the signature scheme
2161 along with other cipher suite parameters when debugging.
2165 * Make OPENSSL_config() error agnostic again.
2169 * Do the error handling in RSA decryption constant time.
2173 * Prevent over long nonces in ChaCha20-Poly1305.
2175 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2176 for every encryption operation. RFC 7539 specifies that the nonce value
2177 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2178 and front pads the nonce with 0 bytes if it is less than 12
2179 bytes. However it also incorrectly allows a nonce to be set of up to 16
2180 bytes. In this case only the last 12 bytes are significant and any
2181 additional leading bytes are ignored.
2183 It is a requirement of using this cipher that nonce values are
2184 unique. Messages encrypted using a reused nonce value are susceptible to
2185 serious confidentiality and integrity attacks. If an application changes
2186 the default nonce length to be longer than 12 bytes and then makes a
2187 change to the leading bytes of the nonce expecting the new value to be a
2188 new unique nonce then such an application could inadvertently encrypt
2189 messages with a reused nonce.
2191 Additionally the ignored bytes in a long nonce are not covered by the
2192 integrity guarantee of this cipher. Any application that relies on the
2193 integrity of these ignored leading bytes of a long nonce may be further
2194 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2195 is safe because no such use sets such a long nonce value. However user
2196 applications that use this cipher directly and set a non-default nonce
2197 length to be longer than 12 bytes may be vulnerable.
2199 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2205 * Add DEVRANDOM_WAIT feature for Linux systems
2207 On older Linux systems where the getrandom() system call is not available,
2208 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2209 Contrary to getrandom(), the /dev/urandom device will not block during
2210 early boot when the kernel CSPRNG has not been seeded yet.
2212 To mitigate this known weakness, use select() to wait for /dev/random to
2213 become readable before reading from /dev/urandom.
2215 * Ensure that SM2 only uses SM3 as digest algorithm
2219 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2221 * Change the info callback signals for the start and end of a post-handshake
2222 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2223 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2224 confused by this and assume that a TLSv1.2 renegotiation has started. This
2225 can break KeyUpdate handling. Instead we no longer signal the start and end
2226 of a post handshake message exchange (although the messages themselves are
2227 still signalled). This could break some applications that were expecting
2228 the old signals. However without this KeyUpdate is not usable for many
2233 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2235 * Timing vulnerability in DSA signature generation
2237 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2238 timing side channel attack. An attacker could use variations in the signing
2239 algorithm to recover the private key.
2241 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2246 * Timing vulnerability in ECDSA signature generation
2248 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2249 timing side channel attack. An attacker could use variations in the signing
2250 algorithm to recover the private key.
2252 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2257 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2258 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2259 of two gigabytes and the error handling improved.
2261 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2262 categorized as a normal bug, not a security issue, because the DRBG reseeds
2263 automatically and is fully functional even without additional randomness
2264 provided by the application.
2266 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2268 * Add a new ClientHello callback. Provides a callback interface that gives
2269 the application the ability to adjust the nascent SSL object at the
2270 earliest stage of ClientHello processing, immediately after extensions have
2271 been collected but before they have been processed. In particular, this
2272 callback can adjust the supported TLS versions in response to the contents
2277 * Add SM2 base algorithm support.
2281 * s390x assembly pack: add (improved) hardware-support for the following
2282 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2283 aes-cfb/cfb8, aes-ecb.
2287 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2288 parameter is no longer accepted, as it leads to a corrupt table. NULL
2289 pem_str is reserved for alias entries only.
2293 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2294 step for prime curves. The new implementation is based on formulae from
2295 differential addition-and-doubling in homogeneous projective coordinates
2296 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2297 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2298 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2299 to work in projective coordinates.
2301 *Billy Bob Brumley, Nicola Tuveri*
2303 * Change generating and checking of primes so that the error rate of not
2304 being prime depends on the intended use based on the size of the input.
2305 For larger primes this will result in more rounds of Miller-Rabin.
2306 The maximal error rate for primes with more than 1080 bits is lowered
2309 *Kurt Roeckx, Annie Yousar*
2311 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2315 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2316 moving between systems, and to avoid confusion when a Windows build is
2317 done with mingw vs with MSVC. For POSIX installs, there's still a
2318 symlink or copy named 'tsget' to avoid that confusion as well.
2322 * Revert blinding in ECDSA sign and instead make problematic addition
2323 length-invariant. Switch even to fixed-length Montgomery multiplication.
2327 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2328 step for binary curves. The new implementation is based on formulae from
2329 differential addition-and-doubling in mixed Lopez-Dahab projective
2330 coordinates, modified to independently blind the operands.
2332 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2334 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2335 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2336 EC_METHODs to implement their own specialized "ladder step", to take
2337 advantage of more favorable coordinate systems or more efficient
2338 differential addition-and-doubling algorithms.
2340 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2342 * Modified the random device based seed sources to keep the relevant
2343 file descriptors open rather than reopening them on each access.
2344 This allows such sources to operate in a chroot() jail without
2345 the associated device nodes being available. This behaviour can be
2346 controlled using RAND_keep_random_devices_open().
2350 * Numerous side-channel attack mitigations have been applied. This may have
2351 performance impacts for some algorithms for the benefit of improved
2352 security. Specific changes are noted in this change log by their respective
2357 * AIX shared library support overhaul. Switch to AIX "natural" way of
2358 handling shared libraries, which means collecting shared objects of
2359 different versions and bitnesses in one common archive. This allows to
2360 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2361 doesn't affect the way 3rd party applications are linked, only how
2362 multi-version installation is managed.
2366 * Make ec_group_do_inverse_ord() more robust and available to other
2367 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2368 mitigations are applied to the fallback BN_mod_inverse().
2369 When using this function rather than BN_mod_inverse() directly, new
2370 EC cryptosystem implementations are then safer-by-default.
2374 * Add coordinate blinding for EC_POINT and implement projective
2375 coordinate blinding for generic prime curves as a countermeasure to
2376 chosen point SCA attacks.
2378 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2380 * Add blinding to ECDSA and DSA signatures to protect against side channel
2381 attacks discovered by Keegan Ryan (NCC Group).
2385 * Enforce checking in the `pkeyutl` command to ensure that the input
2386 length does not exceed the maximum supported digest length when performing
2387 a sign, verify or verifyrecover operation.
2391 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2392 I/O in combination with something like select() or poll() will hang. This
2393 can be turned off again using SSL_CTX_clear_mode().
2394 Many applications do not properly handle non-application data records, and
2395 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2396 around the problems in those applications, but can also break some.
2397 It's recommended to read the manpages about SSL_read(), SSL_write(),
2398 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2399 SSL_CTX_set_read_ahead() again.
2403 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2404 now allow empty (zero character) pass phrases.
2408 * Apply blinding to binary field modular inversion and remove patent
2409 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2413 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2414 binary and prime elliptic curves.
2418 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2419 constant time fixed point multiplication.
2423 * Revise elliptic curve scalar multiplication with timing attack
2424 defenses: ec_wNAF_mul redirects to a constant time implementation
2425 when computing fixed point and variable point multiplication (which
2426 in OpenSSL are mostly used with secret scalars in keygen, sign,
2427 ECDH derive operations).
2428 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2431 * Updated CONTRIBUTING
2435 * Updated DRBG / RAND to request nonce and additional low entropy
2436 randomness from the system.
2438 *Matthias St. Pierre*
2440 * Updated 'openssl rehash' to use OpenSSL consistent default.
2444 * Moved the load of the ssl_conf module to libcrypto, which helps
2445 loading engines that libssl uses before libssl is initialised.
2449 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2453 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2455 *Ingo Schwarze, Rich Salz*
2457 * Added output of accepting IP address and port for 'openssl s_server'
2461 * Added a new API for TLSv1.3 ciphersuites:
2462 SSL_CTX_set_ciphersuites()
2463 SSL_set_ciphersuites()
2467 * Memory allocation failures consistently add an error to the error
2472 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2473 in libcrypto when run as setuid/setgid.
2477 * Load any config file by default when libssl is used.
2481 * Added new public header file <openssl/rand_drbg.h> and documentation
2482 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2484 *Matthias St. Pierre*
2486 * QNX support removed (cannot find contributors to get their approval
2487 for the license change).
2491 * TLSv1.3 replay protection for early data has been implemented. See the
2492 SSL_read_early_data() man page for further details.
2496 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2497 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2498 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2499 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2500 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2501 configuration has been separated out. See the ciphers man page or the
2502 SSL_CTX_set_ciphersuites() man page for more information.
2506 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2507 in responder mode now supports the new "-multi" option, which
2508 spawns the specified number of child processes to handle OCSP
2509 requests. The "-timeout" option now also limits the OCSP
2510 responder's patience to wait to receive the full client request
2511 on a newly accepted connection. Child processes are respawned
2512 as needed, and the CA index file is automatically reloaded
2513 when changed. This makes it possible to run the "ocsp" responder
2514 as a long-running service, making the OpenSSL CA somewhat more
2515 feature-complete. In this mode, most diagnostic messages logged
2516 after entering the event loop are logged via syslog(3) rather than
2521 * Added support for X448 and Ed448. Heavily based on original work by
2526 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2527 objects loaded. This adds the functions OSSL_STORE_expect() and
2528 OSSL_STORE_find() as well as needed tools to construct searches and
2529 get the search data out of them.
2533 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2534 version of OpenSSL should review their configuration settings to ensure
2535 that they are still appropriate for TLSv1.3. For further information see:
2536 <https://wiki.openssl.org/index.php/TLS1.3>
2540 * Grand redesign of the OpenSSL random generator
2542 The default RAND method now utilizes an AES-CTR DRBG according to
2543 NIST standard SP 800-90Ar1. The new random generator is essentially
2544 a port of the default random generator from the OpenSSL FIPS 2.0
2545 object module. It is a hybrid deterministic random bit generator
2546 using an AES-CTR bit stream and which seeds and reseeds itself
2547 automatically using trusted system entropy sources.
2549 Some of its new features are:
2550 - Support for multiple DRBG instances with seed chaining.
2551 - The default RAND method makes use of a DRBG.
2552 - There is a public and private DRBG instance.
2553 - The DRBG instances are fork-safe.
2554 - Keep all global DRBG instances on the secure heap if it is enabled.
2555 - The public and private DRBG instance are per thread for lock free
2558 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2560 * Changed Configure so it only says what it does and doesn't dump
2561 so much data. Instead, ./configdata.pm should be used as a script
2562 to display all sorts of configuration data.
2566 * Added processing of "make variables" to Configure.
2570 * Added SHA512/224 and SHA512/256 algorithm support.
2574 * The last traces of Netware support, first removed in 1.1.0, have
2579 * Get rid of Makefile.shared, and in the process, make the processing
2580 of certain files (rc.obj, or the .def/.map/.opt files produced from
2581 the ordinal files) more visible and hopefully easier to trace and
2582 debug (or make silent).
2586 * Make it possible to have environment variable assignments as
2587 arguments to config / Configure.
2591 * Add multi-prime RSA (RFC 8017) support.
2595 * Add SM3 implemented according to GB/T 32905-2016
2596 *Jack Lloyd <jack.lloyd@ribose.com>,*
2597 *Ronald Tse <ronald.tse@ribose.com>,*
2598 *Erick Borsboom <erick.borsboom@ribose.com>*
2600 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2601 as documented in RFC6066.
2602 Based on a patch from Tomasz Moń
2604 *Filipe Raimundo da Silva*
2606 * Add SM4 implemented according to GB/T 32907-2016.
2607 *Jack Lloyd <jack.lloyd@ribose.com>,*
2608 *Ronald Tse <ronald.tse@ribose.com>,*
2609 *Erick Borsboom <erick.borsboom@ribose.com>*
2611 * Reimplement -newreq-nodes and ERR_error_string_n; the
2612 original author does not agree with the license change.
2616 * Add ARIA AEAD TLS support.
2620 * Some macro definitions to support VS6 have been removed. Visual
2621 Studio 6 has not worked since 1.1.0
2625 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2626 without clearing the errors.
2630 * Add "atfork" functions. If building on a system that without
2631 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2632 requirements. The RAND facility now uses/requires this.
2640 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2641 not possible to disable entirely. However, it's still possible to
2642 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2645 To disable, configure with 'no-ui-console'. 'no-ui' is still
2646 possible to use as an alias. Check at compile time with the
2647 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2648 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2652 * Add a STORE module, which implements a uniform and URI based reader of
2653 stores that can contain keys, certificates, CRLs and numerous other
2654 objects. The main API is loosely based on a few stdio functions,
2655 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2656 OSSL_STORE_error and OSSL_STORE_close.
2657 The implementation uses backends called "loaders" to implement arbitrary
2658 URI schemes. There is one built in "loader" for the 'file' scheme.
2662 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2663 then adjusted to work on FreeBSD 8.4 as well.
2664 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2665 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2669 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2670 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2671 error code calls like this:
2673 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2675 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2676 that can be encoded in C. For the foreseeable future, this will only
2679 *Richard Levitte and Tim Hudson*
2681 * Removed BSD cryptodev engine.
2685 * Add a build target 'build_all_generated', to build all generated files
2686 and only that. This can be used to prepare everything that requires
2687 things like perl for a system that lacks perl and then move everything
2688 to that system and do the rest of the build there.
2692 * In the UI interface, make it possible to duplicate the user data. This
2693 can be used by engines that need to retain the data for a longer time
2694 than just the call where this user data is passed.
2698 * Ignore the '-named_curve auto' value for compatibility of applications
2701 *Tomáš Mráz <tmraz@fedoraproject.org>*
2703 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2704 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2705 alerts across multiple records (some of which could be empty). In practice
2706 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2707 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2708 support this at all. Supporting it adds significant complexity to the
2709 record layer, and its removal is unlikely to cause interoperability
2714 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2715 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2716 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2721 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2722 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2724 *Richard Levitte, Andy Polyakov*
2726 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2731 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2732 platform rather than 'mingw'.
2736 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2737 success if they are asked to add an object which already exists
2738 in the store. This change cascades to other functions which load
2739 certificates and CRLs.
2743 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2744 facilitate stack unwinding even from assembly subroutines.
2748 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2749 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2753 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2754 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2755 which is the minimum version we support.
2759 * Certificate time validation (X509_cmp_time) enforces stricter
2760 compliance with RFC 5280. Fractional seconds and timezone offsets
2761 are no longer allowed.
2765 * Add support for ARIA
2769 * s_client will now send the Server Name Indication (SNI) extension by
2770 default unless the new "-noservername" option is used. The server name is
2771 based on the host provided to the "-connect" option unless overridden by
2772 using "-servername".
2776 * Add support for SipHash
2780 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2781 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2782 prevent issues where no progress is being made and the peer continually
2783 sends unrecognised record types, using up resources processing them.
2787 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2788 using the algorithm defined in
2789 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2793 * Heartbeat support has been removed; the ABI is changed for now.
2795 *Richard Levitte, Rich Salz*
2797 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2801 * The RSA "null" method, which was partially supported to avoid patent
2802 issues, has been replaced to always returns NULL.
2809 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2811 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2812 used even when parsing explicit parameters, when loading a encoded key
2813 or calling `EC_GROUP_new_from_ecpkparameters()`/
2814 `EC_GROUP_new_from_ecparameters()`.
2815 This prevents bypass of security hardening and performance gains,
2816 especially for curves with specialized EC_METHODs.
2817 By default, if a key encoded with explicit parameters is loaded and later
2818 encoded, the output is still encoded with explicit parameters, even if
2819 internally a "named" EC_GROUP is used for computation.
2823 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2824 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2825 NULL. After this change, only the cofactor parameter can be NULL. It also
2826 does some minimal sanity checks on the passed order.
2831 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2832 An attack is simple, if the first CMS_recipientInfo is valid but the
2833 second CMS_recipientInfo is chosen ciphertext. If the second
2834 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2835 encryption key will be replaced by garbage, and the message cannot be
2836 decoded, but if the RSA decryption fails, the correct encryption key is
2837 used and the recipient will not notice the attack.
2838 As a work around for this potential attack the length of the decrypted
2839 key must be equal to the cipher default key length, in case the
2840 certifiate is not given and all recipientInfo are tried out.
2841 The old behaviour can be re-enabled in the CMS code by setting the
2842 CMS_DEBUG_DECRYPT flag.
2847 * Use Windows installation paths in the mingw builds
2849 Mingw isn't a POSIX environment per se, which means that Windows
2850 paths should be used for installation.
2855 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2857 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2858 This changes the size when using the `genpkey` command when no size is given.
2859 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2860 generation commands to use 2048 bits by default.
2864 * Prevent over long nonces in ChaCha20-Poly1305.
2866 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2867 for every encryption operation. RFC 7539 specifies that the nonce value
2868 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2869 and front pads the nonce with 0 bytes if it is less than 12
2870 bytes. However it also incorrectly allows a nonce to be set of up to 16
2871 bytes. In this case only the last 12 bytes are significant and any
2872 additional leading bytes are ignored.
2874 It is a requirement of using this cipher that nonce values are
2875 unique. Messages encrypted using a reused nonce value are susceptible to
2876 serious confidentiality and integrity attacks. If an application changes
2877 the default nonce length to be longer than 12 bytes and then makes a
2878 change to the leading bytes of the nonce expecting the new value to be a
2879 new unique nonce then such an application could inadvertently encrypt
2880 messages with a reused nonce.
2882 Additionally the ignored bytes in a long nonce are not covered by the
2883 integrity guarantee of this cipher. Any application that relies on the
2884 integrity of these ignored leading bytes of a long nonce may be further
2885 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2886 is safe because no such use sets such a long nonce value. However user
2887 applications that use this cipher directly and set a non-default nonce
2888 length to be longer than 12 bytes may be vulnerable.
2890 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2896 * Added SCA hardening for modular field inversion in EC_GROUP through
2897 a new dedicated field_inv() pointer in EC_METHOD.
2898 This also addresses a leakage affecting conversions from projective
2899 to affine coordinates.
2901 *Billy Bob Brumley, Nicola Tuveri*
2903 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2904 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2908 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2912 * Remove the 'dist' target and add a tarball building script. The
2913 'dist' target has fallen out of use, and it shouldn't be
2914 necessary to configure just to create a source distribution.
2918 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2920 * Timing vulnerability in DSA signature generation
2922 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2923 timing side channel attack. An attacker could use variations in the signing
2924 algorithm to recover the private key.
2926 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2931 * Timing vulnerability in ECDSA signature generation
2933 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2934 timing side channel attack. An attacker could use variations in the signing
2935 algorithm to recover the private key.
2937 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2942 * Add coordinate blinding for EC_POINT and implement projective
2943 coordinate blinding for generic prime curves as a countermeasure to
2944 chosen point SCA attacks.
2946 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2948 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2950 * Client DoS due to large DH parameter
2952 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2953 malicious server can send a very large prime value to the client. This will
2954 cause the client to spend an unreasonably long period of time generating a
2955 key for this prime resulting in a hang until the client has finished. This
2956 could be exploited in a Denial Of Service attack.
2958 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2963 * Cache timing vulnerability in RSA Key Generation
2965 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2966 a cache timing side channel attack. An attacker with sufficient access to
2967 mount cache timing attacks during the RSA key generation process could
2968 recover the private key.
2970 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2971 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2976 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2977 parameter is no longer accepted, as it leads to a corrupt table. NULL
2978 pem_str is reserved for alias entries only.
2982 * Revert blinding in ECDSA sign and instead make problematic addition
2983 length-invariant. Switch even to fixed-length Montgomery multiplication.
2987 * Change generating and checking of primes so that the error rate of not
2988 being prime depends on the intended use based on the size of the input.
2989 For larger primes this will result in more rounds of Miller-Rabin.
2990 The maximal error rate for primes with more than 1080 bits is lowered
2993 *Kurt Roeckx, Annie Yousar*
2995 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2999 * Add blinding to ECDSA and DSA signatures to protect against side channel
3000 attacks discovered by Keegan Ryan (NCC Group).
3004 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3005 now allow empty (zero character) pass phrases.
3009 * Certificate time validation (X509_cmp_time) enforces stricter
3010 compliance with RFC 5280. Fractional seconds and timezone offsets
3011 are no longer allowed.
3015 * Fixed a text canonicalisation bug in CMS
3017 Where a CMS detached signature is used with text content the text goes
3018 through a canonicalisation process first prior to signing or verifying a
3019 signature. This process strips trailing space at the end of lines, converts
3020 line terminators to CRLF and removes additional trailing line terminators
3021 at the end of a file. A bug in the canonicalisation process meant that
3022 some characters, such as form-feed, were incorrectly treated as whitespace
3023 and removed. This is contrary to the specification (RFC5485). This fix
3024 could mean that detached text data signed with an earlier version of
3025 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3026 signed with a fixed OpenSSL may fail to verify with an earlier version of
3027 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3028 and use the "-binary" flag (for the "cms" command line application) or set
3029 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3033 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3035 * Constructed ASN.1 types with a recursive definition could exceed the stack
3037 Constructed ASN.1 types with a recursive definition (such as can be found
3038 in PKCS7) could eventually exceed the stack given malicious input with
3039 excessive recursion. This could result in a Denial Of Service attack. There
3040 are no such structures used within SSL/TLS that come from untrusted sources
3041 so this is considered safe.
3043 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3049 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3051 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3052 effectively reduced to only comparing the least significant bit of each
3053 byte. This allows an attacker to forge messages that would be considered as
3054 authenticated in an amount of tries lower than that guaranteed by the
3055 security claims of the scheme. The module can only be compiled by the
3056 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3058 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3064 * Add a build target 'build_all_generated', to build all generated files
3065 and only that. This can be used to prepare everything that requires
3066 things like perl for a system that lacks perl and then move everything
3067 to that system and do the rest of the build there.
3071 * Backport SSL_OP_NO_RENGOTIATION
3073 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3074 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3075 changes this is no longer possible in 1.1.0. Therefore the new
3076 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3077 1.1.0 to provide equivalent functionality.
3079 Note that if an application built against 1.1.0h headers (or above) is run
3080 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3081 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3085 * Removed the OS390-Unix config target. It relied on a script that doesn't
3090 * rsaz_1024_mul_avx2 overflow bug on x86_64
3092 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3093 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3094 Analysis suggests that attacks against RSA and DSA as a result of this
3095 defect would be very difficult to perform and are not believed likely.
3096 Attacks against DH1024 are considered just feasible, because most of the
3097 work necessary to deduce information about a private key may be performed
3098 offline. The amount of resources required for such an attack would be
3099 significant. However, for an attack on TLS to be meaningful, the server
3100 would have to share the DH1024 private key among multiple clients, which is
3101 no longer an option since CVE-2016-0701.
3103 This only affects processors that support the AVX2 but not ADX extensions
3104 like Intel Haswell (4th generation).
3106 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3107 was originally found via the OSS-Fuzz project.
3112 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3114 * bn_sqrx8x_internal carry bug on x86_64
3116 There is a carry propagating bug in the x86_64 Montgomery squaring
3117 procedure. No EC algorithms are affected. Analysis suggests that attacks
3118 against RSA and DSA as a result of this defect would be very difficult to
3119 perform and are not believed likely. Attacks against DH are considered just
3120 feasible (although very difficult) because most of the work necessary to
3121 deduce information about a private key may be performed offline. The amount
3122 of resources required for such an attack would be very significant and
3123 likely only accessible to a limited number of attackers. An attacker would
3124 additionally need online access to an unpatched system using the target
3125 private key in a scenario with persistent DH parameters and a private
3126 key that is shared between multiple clients.
3128 This only affects processors that support the BMI1, BMI2 and ADX extensions
3129 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3131 This issue was reported to OpenSSL by the OSS-Fuzz project.
3136 * Malformed X.509 IPAddressFamily could cause OOB read
3138 If an X.509 certificate has a malformed IPAddressFamily extension,
3139 OpenSSL could do a one-byte buffer overread. The most likely result
3140 would be an erroneous display of the certificate in text format.
3142 This issue was reported to OpenSSL by the OSS-Fuzz project.
3147 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3149 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3150 platform rather than 'mingw'.
3154 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3155 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3156 which is the minimum version we support.
3160 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3162 * Encrypt-Then-Mac renegotiation crash
3164 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3165 negotiated where it was not in the original handshake (or vice-versa) then
3166 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3167 and servers are affected.
3169 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3174 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3176 * Truncated packet could crash via OOB read
3178 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3179 cipher is being used, then a truncated packet can cause that host to
3180 perform an out-of-bounds read, usually resulting in a crash.
3182 This issue was reported to OpenSSL by Robert Święcki of Google.
3187 * Bad (EC)DHE parameters cause a client crash
3189 If a malicious server supplies bad parameters for a DHE or ECDHE key
3190 exchange then this can result in the client attempting to dereference a
3191 NULL pointer leading to a client crash. This could be exploited in a Denial
3194 This issue was reported to OpenSSL by Guido Vranken.
3199 * BN_mod_exp may produce incorrect results on x86_64
3201 There is a carry propagating bug in the x86_64 Montgomery squaring
3202 procedure. No EC algorithms are affected. Analysis suggests that attacks
3203 against RSA and DSA as a result of this defect would be very difficult to
3204 perform and are not believed likely. Attacks against DH are considered just
3205 feasible (although very difficult) because most of the work necessary to
3206 deduce information about a private key may be performed offline. The amount
3207 of resources required for such an attack would be very significant and
3208 likely only accessible to a limited number of attackers. An attacker would
3209 additionally need online access to an unpatched system using the target
3210 private key in a scenario with persistent DH parameters and a private
3211 key that is shared between multiple clients. For example this can occur by
3212 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3213 similar to CVE-2015-3193 but must be treated as a separate problem.
3215 This issue was reported to OpenSSL by the OSS-Fuzz project.
3220 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3222 * ChaCha20/Poly1305 heap-buffer-overflow
3224 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3225 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3226 crash. This issue is not considered to be exploitable beyond a DoS.
3228 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3233 * CMS Null dereference
3235 Applications parsing invalid CMS structures can crash with a NULL pointer
3236 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3237 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3238 structure callback if an attempt is made to free certain invalid encodings.
3239 Only CHOICE structures using a callback which do not handle NULL value are
3242 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3247 * Montgomery multiplication may produce incorrect results
3249 There is a carry propagating bug in the Broadwell-specific Montgomery
3250 multiplication procedure that handles input lengths divisible by, but
3251 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3252 and DH private keys are impossible. This is because the subroutine in
3253 question is not used in operations with the private key itself and an input
3254 of the attacker's direct choice. Otherwise the bug can manifest itself as
3255 transient authentication and key negotiation failures or reproducible
3256 erroneous outcome of public-key operations with specially crafted input.
3257 Among EC algorithms only Brainpool P-512 curves are affected and one
3258 presumably can attack ECDH key negotiation. Impact was not analyzed in
3259 detail, because pre-requisites for attack are considered unlikely. Namely
3260 multiple clients have to choose the curve in question and the server has to
3261 share the private key among them, neither of which is default behaviour.
3262 Even then only clients that chose the curve will be affected.
3264 This issue was publicly reported as transient failures and was not
3265 initially recognized as a security issue. Thanks to Richard Morgan for
3266 providing reproducible case.
3271 * Removed automatic addition of RPATH in shared libraries and executables,
3272 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3276 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3278 * Fix Use After Free for large message sizes
3280 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3281 message larger than approx 16k is received then the underlying buffer to
3282 store the incoming message is reallocated and moved. Unfortunately a
3283 dangling pointer to the old location is left which results in an attempt to
3284 write to the previously freed location. This is likely to result in a
3285 crash, however it could potentially lead to execution of arbitrary code.
3287 This issue only affects OpenSSL 1.1.0a.
3289 This issue was reported to OpenSSL by Robert Święcki.
3294 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3296 * OCSP Status Request extension unbounded memory growth
3298 A malicious client can send an excessively large OCSP Status Request
3299 extension. If that client continually requests renegotiation, sending a
3300 large OCSP Status Request extension each time, then there will be unbounded
3301 memory growth on the server. This will eventually lead to a Denial Of
3302 Service attack through memory exhaustion. Servers with a default
3303 configuration are vulnerable even if they do not support OCSP. Builds using
3304 the "no-ocsp" build time option are not affected.
3306 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3311 * SSL_peek() hang on empty record
3313 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3314 sends an empty record. This could be exploited by a malicious peer in a
3315 Denial Of Service attack.
3317 This issue was reported to OpenSSL by Alex Gaynor.
3322 * Excessive allocation of memory in tls_get_message_header() and
3323 dtls1_preprocess_fragment()
3325 A (D)TLS message includes 3 bytes for its length in the header for the
3326 message. This would allow for messages up to 16Mb in length. Messages of
3327 this length are excessive and OpenSSL includes a check to ensure that a
3328 peer is sending reasonably sized messages in order to avoid too much memory
3329 being consumed to service a connection. A flaw in the logic of version
3330 1.1.0 means that memory for the message is allocated too early, prior to
3331 the excessive message length check. Due to way memory is allocated in
3332 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3333 to service a connection. This could lead to a Denial of Service through
3334 memory exhaustion. However, the excessive message length check still takes
3335 place, and this would cause the connection to immediately fail. Assuming
3336 that the application calls SSL_free() on the failed connection in a timely
3337 manner then the 21Mb of allocated memory will then be immediately freed
3338 again. Therefore the excessive memory allocation will be transitory in
3339 nature. This then means that there is only a security impact if:
3341 1) The application does not call SSL_free() in a timely manner in the event
3342 that the connection fails
3344 2) The application is working in a constrained environment where there is
3345 very little free memory
3347 3) The attacker initiates multiple connection attempts such that there are
3348 multiple connections in a state where memory has been allocated for the
3349 connection; SSL_free() has not yet been called; and there is insufficient
3350 memory to service the multiple requests.
3352 Except in the instance of (1) above any Denial Of Service is likely to be
3353 transitory because as soon as the connection fails the memory is
3354 subsequently freed again in the SSL_free() call. However there is an
3355 increased risk during this period of application crashes due to the lack of
3356 memory - which would then mean a more serious Denial of Service.
3358 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3359 (CVE-2016-6307 and CVE-2016-6308)
3363 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3364 had to be removed. Primary reason is that vendor assembler can't
3365 assemble our modules with -KPIC flag. As result it, assembly
3366 support, was not even available as option. But its lack means
3367 lack of side-channel resistant code, which is incompatible with
3368 security by todays standards. Fortunately gcc is readily available
3369 prepackaged option, which we firmly point at...
3373 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3375 * Windows command-line tool supports UTF-8 opt-in option for arguments
3376 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3377 (to any value) allows Windows user to access PKCS#12 file generated
3378 with Windows CryptoAPI and protected with non-ASCII password, as well
3379 as files generated under UTF-8 locale on Linux also protected with
3384 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3385 have been disabled by default and removed from DEFAULT, just like RC4.
3386 See the RC4 item below to re-enable both.
3390 * The method for finding the storage location for the Windows RAND seed file
3391 has changed. First we check %RANDFILE%. If that is not set then we check
3392 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3393 all else fails we fall back to C:\.
3397 * The EVP_EncryptUpdate() function has had its return type changed from void
3398 to int. A return of 0 indicates and error while a return of 1 indicates
3403 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3404 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3405 off the constant time implementation for RSA, DSA and DH have been made
3406 no-ops and deprecated.
3410 * Windows RAND implementation was simplified to only get entropy by
3411 calling CryptGenRandom(). Various other RAND-related tickets
3414 *Joseph Wylie Yandle, Rich Salz*
3416 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3417 and `OPENSSL_LH_`, respectively. The old names are available
3418 with API compatibility. They new names are now completely documented.
3422 * Unify TYPE_up_ref(obj) methods signature.
3423 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3424 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3425 int (instead of void) like all others TYPE_up_ref() methods.
3426 So now these methods also check the return value of CRYPTO_atomic_add(),
3427 and the validity of object reference counter.
3429 *fdasilvayy@gmail.com*
3431 * With Windows Visual Studio builds, the .pdb files are installed
3432 alongside the installed libraries and executables. For a static
3433 library installation, ossl_static.pdb is the associate compiler
3434 generated .pdb file to be used when linking programs.
3438 * Remove openssl.spec. Packaging files belong with the packagers.
3442 * Automatic Darwin/OSX configuration has had a refresh, it will now
3443 recognise x86_64 architectures automatically. You can still decide
3444 to build for a different bitness with the environment variable
3445 KERNEL_BITS (can be 32 or 64), for example:
3447 KERNEL_BITS=32 ./config
3451 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3452 256 bit AES and HMAC with SHA256.
3456 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3460 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3464 * To enable users to have their own config files and build file templates,
3465 Configure looks in the directory indicated by the environment variable
3466 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3467 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3468 name and is used as is.
3472 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3473 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3474 X509_CERT_FILE_CTX was removed.
3478 * "shared" builds are now the default. To create only static libraries use
3479 the "no-shared" Configure option.
3483 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3484 All of these option have not worked for some while and are fundamental
3489 * Make various cleanup routines no-ops and mark them as deprecated. Most
3490 global cleanup functions are no longer required because they are handled
3491 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3492 Explicitly de-initing can cause problems (e.g. where a library that uses
3493 OpenSSL de-inits, but an application is still using it). The affected
3494 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3495 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3496 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3497 COMP_zlib_cleanup().
3501 * --strict-warnings no longer enables runtime debugging options
3502 such as REF_DEBUG. Instead, debug options are automatically
3503 enabled with '--debug' builds.
3505 *Andy Polyakov, Emilia Käsper*
3507 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3508 have been moved out of the public header files. New functions for managing
3509 these have been added.
3513 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3514 objects have been moved out of the public header files. New
3515 functions for managing these have been added.
3519 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3520 have been moved out of the public header files. New functions for managing
3521 these have been added.
3525 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3526 moved out of the public header files. New functions for managing these
3531 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3535 * Removed the mk1mf build scripts.
3539 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3540 it is always safe to #include a header now.
3544 * Removed the aged BC-32 config and all its supporting scripts
3548 * Removed support for Ultrix, Netware, and OS/2.
3552 * Add support for HKDF.
3554 *Alessandro Ghedini*
3556 * Add support for blake2b and blake2s
3560 * Added support for "pipelining". Ciphers that have the
3561 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3562 encryptions/decryptions simultaneously. There are currently no built-in
3563 ciphers with this property but the expectation is that engines will be able
3564 to offer it to significantly improve throughput. Support has been extended
3565 into libssl so that multiple records for a single connection can be
3566 processed in one go (for >=TLS 1.1).
3570 * Added the AFALG engine. This is an async capable engine which is able to
3571 offload work to the Linux kernel. In this initial version it only supports
3572 AES128-CBC. The kernel must be version 4.1.0 or greater.
3576 * OpenSSL now uses a new threading API. It is no longer necessary to
3577 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3578 are two supported threading models: pthreads and windows threads. It is
3579 also possible to configure OpenSSL at compile time for "no-threads". The
3580 old threading API should no longer be used. The functions have been
3581 replaced with "no-op" compatibility macros.
3583 *Alessandro Ghedini, Matt Caswell*
3585 * Modify behavior of ALPN to invoke callback after SNI/servername
3586 callback, such that updates to the SSL_CTX affect ALPN.
3590 * Add SSL_CIPHER queries for authentication and key-exchange.
3594 * Changes to the DEFAULT cipherlist:
3595 - Prefer (EC)DHE handshakes over plain RSA.
3596 - Prefer AEAD ciphers over legacy ciphers.
3597 - Prefer ECDSA over RSA when both certificates are available.
3598 - Prefer TLSv1.2 ciphers/PRF.
3599 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3604 * Change the ECC default curve list to be this, in order: x25519,
3605 secp256r1, secp521r1, secp384r1.
3609 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3610 disabled by default. They can be re-enabled using the
3611 enable-weak-ssl-ciphers option to Configure.
3615 * If the server has ALPN configured, but supports no protocols that the
3616 client advertises, send a fatal "no_application_protocol" alert.
3617 This behaviour is SHALL in RFC 7301, though it isn't universally
3618 implemented by other servers.
3622 * Add X25519 support.
3623 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3624 for public and private key encoding using the format documented in
3625 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3626 key generation and key derivation.
3628 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3633 * Deprecate SRP_VBASE_get_by_user.
3634 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3635 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3636 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3637 seed, even if the seed is configured.
3639 Users should use SRP_VBASE_get1_by_user instead. Note that in
3640 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3641 also that even though configuring the SRP seed attempts to hide
3642 invalid usernames by continuing the handshake with fake
3643 credentials, this behaviour is not constant time and no strong
3644 guarantees are made that the handshake is indistinguishable from
3645 that of a valid user.
3649 * Configuration change; it's now possible to build dynamic engines
3650 without having to build shared libraries and vice versa. This
3651 only applies to the engines in `engines/`, those in `crypto/engine/`
3652 will always be built into libcrypto (i.e. "static").
3654 Building dynamic engines is enabled by default; to disable, use
3655 the configuration option "disable-dynamic-engine".
3657 The only requirements for building dynamic engines are the
3658 presence of the DSO module and building with position independent
3659 code, so they will also automatically be disabled if configuring
3660 with "disable-dso" or "disable-pic".
3662 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3663 are also taken away from openssl/opensslconf.h, as they are
3668 * Configuration change; if there is a known flag to compile
3669 position independent code, it will always be applied on the
3670 libcrypto and libssl object files, and never on the application
3671 object files. This means other libraries that use routines from
3672 libcrypto / libssl can be made into shared libraries regardless
3673 of how OpenSSL was configured.
3675 If this isn't desirable, the configuration options "disable-pic"
3676 or "no-pic" can be used to disable the use of PIC. This will
3677 also disable building shared libraries and dynamic engines.
3681 * Removed JPAKE code. It was experimental and has no wide use.
3685 * The INSTALL_PREFIX Makefile variable has been renamed to
3686 DESTDIR. That makes for less confusion on what this variable
3687 is for. Also, the configuration option --install_prefix is
3692 * Heartbeat for TLS has been removed and is disabled by default
3693 for DTLS; configure with enable-heartbeats. Code that uses the
3694 old #define's might need to be updated.
3696 *Emilia Käsper, Rich Salz*
3698 * Rename REF_CHECK to REF_DEBUG.
3702 * New "unified" build system
3704 The "unified" build system is aimed to be a common system for all
3705 platforms we support. With it comes new support for VMS.
3707 This system builds supports building in a different directory tree
3708 than the source tree. It produces one Makefile (for unix family
3709 or lookalikes), or one descrip.mms (for VMS).
3711 The source of information to make the Makefile / descrip.mms is
3712 small files called 'build.info', holding the necessary
3713 information for each directory with source to compile, and a
3714 template in Configurations, like unix-Makefile.tmpl or
3717 With this change, the library names were also renamed on Windows
3718 and on VMS. They now have names that are closer to the standard
3719 on Unix, and include the major version number, and in certain
3720 cases, the architecture they are built for. See "Notes on shared
3721 libraries" in INSTALL.
3723 We rely heavily on the perl module Text::Template.
3727 * Added support for auto-initialisation and de-initialisation of the library.
3728 OpenSSL no longer requires explicit init or deinit routines to be called,
3729 except in certain circumstances. See the OPENSSL_init_crypto() and
3730 OPENSSL_init_ssl() man pages for further information.
3734 * The arguments to the DTLSv1_listen function have changed. Specifically the
3735 "peer" argument is now expected to be a BIO_ADDR object.
3737 * Rewrite of BIO networking library. The BIO library lacked consistent
3738 support of IPv6, and adding it required some more extensive
3739 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3740 which hold all types of addresses and chains of address information.
3741 It also introduces a new API, with functions like BIO_socket,
3742 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3743 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3744 have been adapted accordingly.
3748 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3753 * CRIME protection: disable compression by default, even if OpenSSL is
3754 compiled with zlib enabled. Applications can still enable compression
3755 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3756 using the SSL_CONF library to configure compression.
3760 * The signature of the session callback configured with
3761 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3762 was explicitly marked as `const unsigned char*` instead of
3767 * Always DPURIFY. Remove the use of uninitialized memory in the
3768 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3772 * Removed many obsolete configuration items, including
3773 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3774 MD2_CHAR, MD2_INT, MD2_LONG
3776 IDEA_SHORT, IDEA_LONG
3777 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3779 *Rich Salz, with advice from Andy Polyakov*
3781 * Many BN internals have been moved to an internal header file.
3783 *Rich Salz with help from Andy Polyakov*
3785 * Configuration and writing out the results from it has changed.
3786 Files such as Makefile include/openssl/opensslconf.h and are now
3787 produced through general templates, such as Makefile.in and
3788 crypto/opensslconf.h.in and some help from the perl module
3791 Also, the center of configuration information is no longer
3792 Makefile. Instead, Configure produces a perl module in
3793 configdata.pm which holds most of the config data (in the hash
3794 table %config), the target data that comes from the target
3795 configuration in one of the `Configurations/*.conf` files (in
3800 * To clarify their intended purposes, the Configure options
3801 --prefix and --openssldir change their semantics, and become more
3802 straightforward and less interdependent.
3804 --prefix shall be used exclusively to give the location INSTALLTOP
3805 where programs, scripts, libraries, include files and manuals are
3806 going to be installed. The default is now /usr/local.
3808 --openssldir shall be used exclusively to give the default
3809 location OPENSSLDIR where certificates, private keys, CRLs are
3810 managed. This is also where the default openssl.cnf gets
3812 If the directory given with this option is a relative path, the
3813 values of both the --prefix value and the --openssldir value will
3814 be combined to become OPENSSLDIR.
3815 The default for --openssldir is INSTALLTOP/ssl.
3817 Anyone who uses --openssldir to specify where OpenSSL is to be
3818 installed MUST change to use --prefix instead.
3822 * The GOST engine was out of date and therefore it has been removed. An up
3823 to date GOST engine is now being maintained in an external repository.
3824 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3825 support for GOST ciphersuites (these are only activated if a GOST engine
3830 * EGD is no longer supported by default; use enable-egd when
3833 *Ben Kaduk and Rich Salz*
3835 * The distribution now has Makefile.in files, which are used to
3836 create Makefile's when Configure is run. *Configure must be run
3837 before trying to build now.*
3841 * The return value for SSL_CIPHER_description() for error conditions
3846 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3848 Obtaining and performing DNSSEC validation of TLSA records is
3849 the application's responsibility. The application provides
3850 the TLSA records of its choice to OpenSSL, and these are then
3851 used to authenticate the peer.
3853 The TLSA records need not even come from DNS. They can, for
3854 example, be used to implement local end-entity certificate or
3855 trust-anchor "pinning", where the "pin" data takes the form
3856 of TLSA records, which can augment or replace verification
3857 based on the usual WebPKI public certification authorities.
3861 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3862 continues to support deprecated interfaces in default builds.
3863 However, applications are strongly advised to compile their
3864 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3865 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3866 or the 1.1.0 releases.
3868 In environments in which all applications have been ported to
3869 not use any deprecated interfaces OpenSSL's Configure script
3870 should be used with the --api=1.1.0 option to entirely remove
3871 support for the deprecated features from the library and
3872 unconditionally disable them in the installed headers.
3873 Essentially the same effect can be achieved with the "no-deprecated"
3874 argument to Configure, except that this will always restrict
3875 the build to just the latest API, rather than a fixed API
3878 As applications are ported to future revisions of the API,
3879 they should update their compile-time OPENSSL_API_COMPAT define
3880 accordingly, but in most cases should be able to continue to
3881 compile with later releases.
3883 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3884 0x10000000L and 0x00908000L, respectively. However those
3885 versions did not support the OPENSSL_API_COMPAT feature, and
3886 so applications are not typically tested for explicit support
3887 of just the undeprecated features of either release.
3891 * Add support for setting the minimum and maximum supported protocol.
3892 It can bet set via the SSL_set_min_proto_version() and
3893 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3894 MaxProtocol. It's recommended to use the new APIs to disable
3895 protocols instead of disabling individual protocols using
3896 SSL_set_options() or SSL_CONF's Protocol. This change also
3897 removes support for disabling TLS 1.2 in the OpenSSL TLS
3898 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3902 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3906 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3907 and integrates ECDSA and ECDH functionality into EC. Implementations can
3908 now redirect key generation and no longer need to convert to or from
3911 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3912 include the ec.h header file instead.
3916 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3917 ciphers who are no longer supported and drops support the ephemeral RSA key
3918 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3922 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3923 opaque. For HMAC_CTX, the following constructors and destructors
3926 HMAC_CTX *HMAC_CTX_new(void);
3927 void HMAC_CTX_free(HMAC_CTX *ctx);
3929 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3930 destroy such methods has been added. See EVP_MD_meth_new(3) and
3931 EVP_CIPHER_meth_new(3) for documentation.
3934 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3935 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3936 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3937 an already created structure.
3938 2) For consistency with the majority of our object creators and
3939 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3940 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3941 for deprecated builds.
3945 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3946 cryptographic operations to be performed asynchronously as long as an
3947 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3948 further details. Libssl has also had this capability integrated with the
3949 introduction of the new mode SSL_MODE_ASYNC and associated error
3950 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3951 pages. This work was developed in partnership with Intel Corp.
3955 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3956 always enabled now. If you want to disable the support you should
3957 exclude it using the list of supported ciphers. This also means that the
3958 "-no_ecdhe" option has been removed from s_server.
3962 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3963 SSL_{CTX_}set1_curves() which can set a list.
3967 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3968 curve you want to support using SSL_{CTX_}set1_curves().
3972 * State machine rewrite. The state machine code has been significantly
3973 refactored in order to remove much duplication of code and solve issues
3974 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3975 further details). This change does have some associated API changes.
3976 Notably the SSL_state() function has been removed and replaced by
3977 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3978 SSL_set_state() has been removed altogether. The previous handshake states
3979 defined in ssl.h and ssl3.h have also been removed.
3983 * All instances of the string "ssleay" in the public API were replaced
3984 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3985 Some error codes related to internal RSA_eay API's were renamed.
3989 * The demo files in crypto/threads were moved to demo/threads.
3993 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3996 *Matt Caswell, Rich Salz*
3998 * New ASN.1 embed macro.
4000 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4001 structure is not allocated: it is part of the parent. That is instead of
4009 This reduces memory fragmentation and make it impossible to accidentally
4010 set a mandatory field to NULL.
4012 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4013 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4014 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4019 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4023 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4024 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4025 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4026 DES and RC4 ciphersuites.
4030 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4031 This changes the decoding behaviour for some invalid messages,
4032 though the change is mostly in the more lenient direction, and
4033 legacy behaviour is preserved as much as possible.
4037 * Fix no-stdio build.
4038 *David Woodhouse <David.Woodhouse@intel.com> and also*
4039 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4041 * New testing framework
4042 The testing framework has been largely rewritten and is now using
4043 perl and the perl modules Test::Harness and an extended variant of
4044 Test::More called OpenSSL::Test to do its work. All test scripts in
4045 test/ have been rewritten into test recipes, and all direct calls to
4046 executables in test/Makefile have become individual recipes using the
4047 simplified testing OpenSSL::Test::Simple.
4049 For documentation on our testing modules, do:
4051 perldoc test/testlib/OpenSSL/Test/Simple.pm
4052 perldoc test/testlib/OpenSSL/Test.pm
4056 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4057 are used; the latter aborts on memory leaks (usually checked on exit).
4058 Some undocumented "set malloc, etc., hooks" functions were removed
4059 and others were changed. All are now documented.
4063 * In DSA_generate_parameters_ex, if the provided seed is too short,
4066 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4068 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4069 from RFC4279, RFC4785, RFC5487, RFC5489.
4071 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4072 original RSA_PSK patch.
4076 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4077 era flag was never set throughout the codebase (only read). Also removed
4078 SSL3_FLAGS_POP_BUFFER which was only used if
4079 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4083 * Changed the default name options in the "ca", "crl", "req" and "x509"
4084 to be "oneline" instead of "compat".
4088 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4089 not aware of clients that still exhibit this bug, and the workaround
4090 hasn't been working properly for a while.
4094 * The return type of BIO_number_read() and BIO_number_written() as well as
4095 the corresponding num_read and num_write members in the BIO structure has
4096 changed from unsigned long to uint64_t. On platforms where an unsigned
4097 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4102 * Given the pervasive nature of TLS extensions it is inadvisable to run
4103 OpenSSL without support for them. It also means that maintaining
4104 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4105 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4109 * Removed support for the two export grade static DH ciphersuites
4110 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4111 were newly added (along with a number of other static DH ciphersuites) to
4112 1.0.2. However the two export ones have *never* worked since they were
4113 introduced. It seems strange in any case to be adding new export
4114 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4118 * Version negotiation has been rewritten. In particular SSLv23_method(),
4119 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4120 and turned into macros which simply call the new preferred function names
4121 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4122 should use the new names instead. Also as part of this change the ssl23.h
4123 header file has been removed.
4127 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4128 code and the associated standard is no longer considered fit-for-purpose.
4132 * RT2547 was closed. When generating a private key, try to make the
4133 output file readable only by the owner. This behavior change might
4134 be noticeable when interacting with other software.
4136 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4141 * Added HTTP GET support to the ocsp command.
4145 * Changed default digest for the dgst and enc commands from MD5 to
4150 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4154 * Added support for TLS extended master secret from
4155 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4156 initial patch which was a great help during development.
4160 * All libssl internal structures have been removed from the public header
4161 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4162 now redundant). Users should not attempt to access internal structures
4163 directly. Instead they should use the provided API functions.
4167 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4168 Access to deprecated functions can be re-enabled by running config with
4169 "enable-deprecated". In addition applications wishing to use deprecated
4170 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4171 will, by default, disable some transitive includes that previously existed
4172 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4176 * Added support for OCB mode. OpenSSL has been granted a patent license
4177 compatible with the OpenSSL license for use of OCB. Details are available
4178 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4179 for OCB can be removed by calling config with no-ocb.
4183 * SSLv2 support has been removed. It still supports receiving a SSLv2
4184 compatible client hello.
4188 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4189 done while fixing the error code for the key-too-small case.
4191 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4193 * CA.sh has been removed; use CA.pl instead.
4197 * Removed old DES API.
4201 * Remove various unsupported platforms:
4207 Sinix/ReliantUNIX RM400
4212 16-bit platforms such as WIN16
4216 * Clean up OPENSSL_NO_xxx #define's
4217 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4218 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4219 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4220 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4221 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4222 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4223 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4224 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4225 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4226 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4230 * Cleaned up dead code
4231 Remove all but one '#ifdef undef' which is to be looked at.
4235 * Clean up calling of xxx_free routines.
4236 Just like free(), fix most of the xxx_free routines to accept
4237 NULL. Remove the non-null checks from callers. Save much code.
4241 * Add secure heap for storage of private keys (when possible).
4242 Add BIO_s_secmem(), CBIGNUM, etc.
4243 Contributed by Akamai Technologies under our Corporate CLA.
4247 * Experimental support for a new, fast, unbiased prime candidate generator,
4248 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4250 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4252 * New output format NSS in the sess_id command line tool. This allows
4253 exporting the session id and the master key in NSS keylog format.
4255 *Martin Kaiser <martin@kaiser.cx>*
4257 * Harmonize version and its documentation. -f flag is used to display
4260 *mancha <mancha1@zoho.com>*
4262 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4263 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4265 *mancha <mancha1@zoho.com>*
4267 * Fix some double frees. These are not thought to be exploitable.
4269 *mancha <mancha1@zoho.com>*
4271 * A missing bounds check in the handling of the TLS heartbeat extension
4272 can be used to reveal up to 64k of memory to a connected client or
4275 Thanks for Neel Mehta of Google Security for discovering this bug and to
4276 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4277 preparing the fix ([CVE-2014-0160])
4279 *Adam Langley, Bodo Moeller*
4281 * Fix for the attack described in the paper "Recovering OpenSSL
4282 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4283 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4284 <http://eprint.iacr.org/2014/140>
4286 Thanks to Yuval Yarom and Naomi Benger for discovering this
4287 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4289 *Yuval Yarom and Naomi Benger*
4291 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4292 this fixes a limitation in previous versions of OpenSSL.
4296 * Experimental encrypt-then-mac support.
4298 Experimental support for encrypt then mac from
4299 draft-gutmann-tls-encrypt-then-mac-02.txt
4301 To enable it set the appropriate extension number (0x42 for the test
4302 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4304 For non-compliant peers (i.e. just about everything) this should have no
4307 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4311 * Add EVP support for key wrapping algorithms, to avoid problems with
4312 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4313 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4314 algorithms and include tests cases.
4318 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4323 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4324 MGF1 digest and OAEP label.
4328 * Make openssl verify return errors.
4330 *Chris Palmer <palmer@google.com> and Ben Laurie*
4332 * New function ASN1_TIME_diff to calculate the difference between two
4333 ASN1_TIME structures or one structure and the current time.
4337 * Update fips_test_suite to support multiple command line options. New
4338 test to induce all self test errors in sequence and check expected
4343 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4344 sign or verify all in one operation.
4348 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4349 test programs and fips_test_suite. Includes functionality to parse
4350 the minimal script output of fipsalgest.pl directly.
4354 * Add authorisation parameter to FIPS_module_mode_set().
4358 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4362 * Use separate DRBG fields for internal and external flags. New function
4363 FIPS_drbg_health_check() to perform on demand health checking. Add
4364 generation tests to fips_test_suite with reduced health check interval to
4365 demonstrate periodic health checking. Add "nodh" option to
4366 fips_test_suite to skip very slow DH test.
4370 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4375 * More extensive health check for DRBG checking many more failure modes.
4376 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4377 combination: call this in fips_test_suite.
4381 * Add support for canonical generation of DSA parameter 'g'. See
4384 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4385 POST to handle HMAC cases.
4389 * Add functions FIPS_module_version() and FIPS_module_version_text()
4390 to return numerical and string versions of the FIPS module number.
4394 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4395 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4396 outside the validated module in the FIPS capable OpenSSL.
4400 * Minor change to DRBG entropy callback semantics. In some cases
4401 there is no multiple of the block length between min_len and
4402 max_len. Allow the callback to return more than max_len bytes
4403 of entropy but discard any extra: it is the callback's responsibility
4404 to ensure that the extra data discarded does not impact the
4405 requested amount of entropy.
4409 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4410 information in FIPS186-3, SP800-57 and SP800-131A.
4414 * CCM support via EVP. Interface is very similar to GCM case except we
4415 must supply all data in one chunk (i.e. no update, final) and the
4416 message length must be supplied if AAD is used. Add algorithm test
4421 * Initial version of POST overhaul. Add POST callback to allow the status
4422 of POST to be monitored and/or failures induced. Modify fips_test_suite
4423 to use callback. Always run all selftests even if one fails.
4427 * XTS support including algorithm test driver in the fips_gcmtest program.
4428 Note: this does increase the maximum key length from 32 to 64 bytes but
4429 there should be no binary compatibility issues as existing applications
4430 will never use XTS mode.
4434 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4435 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4436 performs algorithm blocking for unapproved PRNG types. Also do not
4437 set PRNG type in FIPS_mode_set(): leave this to the application.
4438 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4439 the standard OpenSSL PRNG: set additional data to a date time vector.
4443 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4444 This shouldn't present any incompatibility problems because applications
4445 shouldn't be using these directly and any that are will need to rethink
4446 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4450 * Extensive self tests and health checking required by SP800-90 DRBG.
4451 Remove strength parameter from FIPS_drbg_instantiate and always
4452 instantiate at maximum supported strength.
4456 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4460 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4464 * New function DH_compute_key_padded() to compute a DH key and pad with
4465 leading zeroes if needed: this complies with SP800-56A et al.
4469 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4470 anything, incomplete, subject to change and largely untested at present.
4474 * Modify fipscanisteronly build option to only build the necessary object
4475 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4479 * Add experimental option FIPSSYMS to give all symbols in
4480 fipscanister.o and FIPS or fips prefix. This will avoid
4481 conflicts with future versions of OpenSSL. Add perl script
4482 util/fipsas.pl to preprocess assembly language source files
4483 and rename any affected symbols.
4487 * Add selftest checks and algorithm block of non-fips algorithms in
4488 FIPS mode. Remove DES2 from selftests.
4492 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4493 return internal method without any ENGINE dependencies. Add new
4494 tiny fips sign and verify functions.
4498 * New build option no-ec2m to disable characteristic 2 code.
4502 * New build option "fipscanisteronly". This only builds fipscanister.o
4503 and (currently) associated fips utilities. Uses the file Makefile.fips
4504 instead of Makefile.org as the prototype.
4508 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4509 Update fips_gcmtest to use IV generator.
4513 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4514 setting output buffer to NULL. The `*Final` function must be
4515 called although it will not retrieve any additional data. The tag
4516 can be set or retrieved with a ctrl. The IV length is by default 12
4517 bytes (96 bits) but can be set to an alternative value. If the IV
4518 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4523 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4524 underlying do_cipher function handles all cipher semantics itself
4525 including padding and finalisation. This is useful if (for example)
4526 an ENGINE cipher handles block padding itself. The behaviour of
4527 do_cipher is subtly changed if this flag is set: the return value
4528 is the number of characters written to the output buffer (zero is
4529 no longer an error code) or a negative error code. Also if the
4530 input buffer is NULL and length 0 finalisation should be performed.
4534 * If a candidate issuer certificate is already part of the constructed
4535 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4539 * Improve forward-security support: add functions
4541 void SSL_CTX_set_not_resumable_session_callback(
4542 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4543 void SSL_set_not_resumable_session_callback(
4544 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4546 for use by SSL/TLS servers; the callback function will be called whenever a
4547 new session is created, and gets to decide whether the session may be
4548 cached to make it resumable (return 0) or not (return 1). (As by the
4549 SSL/TLS protocol specifications, the session_id sent by the server will be
4550 empty to indicate that the session is not resumable; also, the server will
4551 not generate RFC 4507 (RFC 5077) session tickets.)
4553 A simple reasonable callback implementation is to return is_forward_secure.
4554 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4555 by the SSL/TLS server library, indicating whether it can provide forward
4558 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4560 * New -verify_name option in command line utilities to set verification
4565 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4566 Add CMAC pkey methods.
4570 * Experimental renegotiation in s_server -www mode. If the client
4571 browses /reneg connection is renegotiated. If /renegcert it is
4572 renegotiated requesting a certificate.
4576 * Add an "external" session cache for debugging purposes to s_server. This
4577 should help trace issues which normally are only apparent in deployed
4578 multi-process servers.
4582 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4583 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4584 BIO_set_cipher() and some obscure PEM functions were changed so they
4585 can now return an error. The RAND changes required a change to the
4586 RAND_METHOD structure.
4590 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4591 a gcc attribute to warn if the result of a function is ignored. This
4592 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4593 whose return value is often ignored.
4597 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4598 These allow SCTs (signed certificate timestamps) to be requested and
4599 validated when establishing a connection.
4601 *Rob Percival <robpercival@google.com>*
4606 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4608 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4609 used even when parsing explicit parameters, when loading a encoded key
4610 or calling `EC_GROUP_new_from_ecpkparameters()`/
4611 `EC_GROUP_new_from_ecparameters()`.
4612 This prevents bypass of security hardening and performance gains,
4613 especially for curves with specialized EC_METHODs.
4614 By default, if a key encoded with explicit parameters is loaded and later
4615 encoded, the output is still encoded with explicit parameters, even if
4616 internally a "named" EC_GROUP is used for computation.
4620 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4621 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4622 NULL. After this change, only the cofactor parameter can be NULL. It also
4623 does some minimal sanity checks on the passed order.
4628 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4629 An attack is simple, if the first CMS_recipientInfo is valid but the
4630 second CMS_recipientInfo is chosen ciphertext. If the second
4631 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4632 encryption key will be replaced by garbage, and the message cannot be
4633 decoded, but if the RSA decryption fails, the correct encryption key is
4634 used and the recipient will not notice the attack.
4635 As a work around for this potential attack the length of the decrypted
4636 key must be equal to the cipher default key length, in case the
4637 certifiate is not given and all recipientInfo are tried out.
4638 The old behaviour can be re-enabled in the CMS code by setting the
4639 CMS_DEBUG_DECRYPT flag.
4644 * Document issue with installation paths in diverse Windows builds
4646 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4647 binaries and run-time config file.
4652 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4654 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4655 This changes the size when using the `genpkey` command when no size is given.
4656 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4657 generation commands to use 2048 bits by default.
4661 * Add FIPS support for Android Arm 64-bit
4663 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4664 Module in Version 2.0.10. For some reason, the corresponding target
4665 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4666 built with FIPS support on Android Arm 64-bit. This omission has been
4669 *Matthias St. Pierre*
4671 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4673 * 0-byte record padding oracle
4675 If an application encounters a fatal protocol error and then calls
4676 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4677 then OpenSSL can respond differently to the calling application if a 0 byte
4678 record is received with invalid padding compared to if a 0 byte record is
4679 received with an invalid MAC. If the application then behaves differently
4680 based on that in a way that is detectable to the remote peer, then this
4681 amounts to a padding oracle that could be used to decrypt data.
4683 In order for this to be exploitable "non-stitched" ciphersuites must be in
4684 use. Stitched ciphersuites are optimised implementations of certain
4685 commonly used ciphersuites. Also the application must call SSL_shutdown()
4686 twice even if a protocol error has occurred (applications should not do
4687 this but some do anyway).
4689 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4690 Aviram, with additional investigation by Steven Collison and Andrew
4691 Hourselt. It was reported to OpenSSL on 10th December 2018.
4696 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4700 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4702 * Microarchitecture timing vulnerability in ECC scalar multiplication
4704 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4705 shown to be vulnerable to a microarchitecture timing side channel attack.
4706 An attacker with sufficient access to mount local timing attacks during
4707 ECDSA signature generation could recover the private key.
4709 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4710 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4716 * Timing vulnerability in DSA signature generation
4718 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4719 timing side channel attack. An attacker could use variations in the signing
4720 algorithm to recover the private key.
4722 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4727 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4728 Module, accidentally introduced while backporting security fixes from the
4729 development branch and hindering the use of ECC in FIPS mode.
4733 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4735 * Client DoS due to large DH parameter
4737 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4738 malicious server can send a very large prime value to the client. This will
4739 cause the client to spend an unreasonably long period of time generating a
4740 key for this prime resulting in a hang until the client has finished. This
4741 could be exploited in a Denial Of Service attack.
4743 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4748 * Cache timing vulnerability in RSA Key Generation
4750 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4751 a cache timing side channel attack. An attacker with sufficient access to
4752 mount cache timing attacks during the RSA key generation process could
4753 recover the private key.
4755 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4756 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4761 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4762 parameter is no longer accepted, as it leads to a corrupt table. NULL
4763 pem_str is reserved for alias entries only.
4767 * Revert blinding in ECDSA sign and instead make problematic addition
4768 length-invariant. Switch even to fixed-length Montgomery multiplication.
4772 * Change generating and checking of primes so that the error rate of not
4773 being prime depends on the intended use based on the size of the input.
4774 For larger primes this will result in more rounds of Miller-Rabin.
4775 The maximal error rate for primes with more than 1080 bits is lowered
4778 *Kurt Roeckx, Annie Yousar*
4780 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4784 * Add blinding to ECDSA and DSA signatures to protect against side channel
4785 attacks discovered by Keegan Ryan (NCC Group).
4789 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4790 now allow empty (zero character) pass phrases.
4794 * Certificate time validation (X509_cmp_time) enforces stricter
4795 compliance with RFC 5280. Fractional seconds and timezone offsets
4796 are no longer allowed.
4800 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4802 * Constructed ASN.1 types with a recursive definition could exceed the stack
4804 Constructed ASN.1 types with a recursive definition (such as can be found
4805 in PKCS7) could eventually exceed the stack given malicious input with
4806 excessive recursion. This could result in a Denial Of Service attack. There
4807 are no such structures used within SSL/TLS that come from untrusted sources
4808 so this is considered safe.
4810 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4816 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4818 * Read/write after SSL object in error state
4820 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4821 mechanism. The intent was that if a fatal error occurred during a handshake
4822 then OpenSSL would move into the error state and would immediately fail if
4823 you attempted to continue the handshake. This works as designed for the
4824 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4825 SSL_connect()), however due to a bug it does not work correctly if
4826 SSL_read() or SSL_write() is called directly. In that scenario, if the
4827 handshake fails then a fatal error will be returned in the initial function
4828 call. If SSL_read()/SSL_write() is subsequently called by the application
4829 for the same SSL object then it will succeed and the data is passed without
4830 being decrypted/encrypted directly from the SSL/TLS record layer.
4832 In order to exploit this issue an application bug would have to be present
4833 that resulted in a call to SSL_read()/SSL_write() being issued after having
4834 already received a fatal error.
4836 This issue was reported to OpenSSL by David Benjamin (Google).
4841 * rsaz_1024_mul_avx2 overflow bug on x86_64
4843 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4844 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4845 Analysis suggests that attacks against RSA and DSA as a result of this
4846 defect would be very difficult to perform and are not believed likely.
4847 Attacks against DH1024 are considered just feasible, because most of the
4848 work necessary to deduce information about a private key may be performed
4849 offline. The amount of resources required for such an attack would be
4850 significant. However, for an attack on TLS to be meaningful, the server
4851 would have to share the DH1024 private key among multiple clients, which is
4852 no longer an option since CVE-2016-0701.
4854 This only affects processors that support the AVX2 but not ADX extensions
4855 like Intel Haswell (4th generation).
4857 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4858 was originally found via the OSS-Fuzz project.
4863 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4865 * bn_sqrx8x_internal carry bug on x86_64
4867 There is a carry propagating bug in the x86_64 Montgomery squaring
4868 procedure. No EC algorithms are affected. Analysis suggests that attacks
4869 against RSA and DSA as a result of this defect would be very difficult to
4870 perform and are not believed likely. Attacks against DH are considered just
4871 feasible (although very difficult) because most of the work necessary to
4872 deduce information about a private key may be performed offline. The amount
4873 of resources required for such an attack would be very significant and
4874 likely only accessible to a limited number of attackers. An attacker would
4875 additionally need online access to an unpatched system using the target
4876 private key in a scenario with persistent DH parameters and a private
4877 key that is shared between multiple clients.
4879 This only affects processors that support the BMI1, BMI2 and ADX extensions
4880 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4882 This issue was reported to OpenSSL by the OSS-Fuzz project.
4887 * Malformed X.509 IPAddressFamily could cause OOB read
4889 If an X.509 certificate has a malformed IPAddressFamily extension,
4890 OpenSSL could do a one-byte buffer overread. The most likely result
4891 would be an erroneous display of the certificate in text format.
4893 This issue was reported to OpenSSL by the OSS-Fuzz project.
4898 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4900 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4901 platform rather than 'mingw'.
4905 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4907 * Truncated packet could crash via OOB read
4909 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4910 cipher is being used, then a truncated packet can cause that host to
4911 perform an out-of-bounds read, usually resulting in a crash.
4913 This issue was reported to OpenSSL by Robert Święcki of Google.
4918 * BN_mod_exp may produce incorrect results on x86_64
4920 There is a carry propagating bug in the x86_64 Montgomery squaring
4921 procedure. No EC algorithms are affected. Analysis suggests that attacks
4922 against RSA and DSA as a result of this defect would be very difficult to
4923 perform and are not believed likely. Attacks against DH are considered just
4924 feasible (although very difficult) because most of the work necessary to
4925 deduce information about a private key may be performed offline. The amount
4926 of resources required for such an attack would be very significant and
4927 likely only accessible to a limited number of attackers. An attacker would
4928 additionally need online access to an unpatched system using the target
4929 private key in a scenario with persistent DH parameters and a private
4930 key that is shared between multiple clients. For example this can occur by
4931 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4932 similar to CVE-2015-3193 but must be treated as a separate problem.
4934 This issue was reported to OpenSSL by the OSS-Fuzz project.
4939 * Montgomery multiplication may produce incorrect results
4941 There is a carry propagating bug in the Broadwell-specific Montgomery
4942 multiplication procedure that handles input lengths divisible by, but
4943 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4944 and DH private keys are impossible. This is because the subroutine in
4945 question is not used in operations with the private key itself and an input
4946 of the attacker's direct choice. Otherwise the bug can manifest itself as
4947 transient authentication and key negotiation failures or reproducible
4948 erroneous outcome of public-key operations with specially crafted input.
4949 Among EC algorithms only Brainpool P-512 curves are affected and one
4950 presumably can attack ECDH key negotiation. Impact was not analyzed in
4951 detail, because pre-requisites for attack are considered unlikely. Namely
4952 multiple clients have to choose the curve in question and the server has to
4953 share the private key among them, neither of which is default behaviour.
4954 Even then only clients that chose the curve will be affected.
4956 This issue was publicly reported as transient failures and was not
4957 initially recognized as a security issue. Thanks to Richard Morgan for
4958 providing reproducible case.
4963 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4964 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4965 prevent issues where no progress is being made and the peer continually
4966 sends unrecognised record types, using up resources processing them.
4970 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4972 * Missing CRL sanity check
4974 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4975 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4976 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4978 This issue only affects the OpenSSL 1.0.2i
4983 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4985 * OCSP Status Request extension unbounded memory growth
4987 A malicious client can send an excessively large OCSP Status Request
4988 extension. If that client continually requests renegotiation, sending a
4989 large OCSP Status Request extension each time, then there will be unbounded
4990 memory growth on the server. This will eventually lead to a Denial Of
4991 Service attack through memory exhaustion. Servers with a default
4992 configuration are vulnerable even if they do not support OCSP. Builds using
4993 the "no-ocsp" build time option are not affected.
4995 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5000 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5003 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5009 * OOB write in MDC2_Update()
5011 An overflow can occur in MDC2_Update() either if called directly or
5012 through the EVP_DigestUpdate() function using MDC2. If an attacker
5013 is able to supply very large amounts of input data after a previous
5014 call to EVP_EncryptUpdate() with a partial block then a length check
5015 can overflow resulting in a heap corruption.
5017 The amount of data needed is comparable to SIZE_MAX which is impractical
5020 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5025 * Malformed SHA512 ticket DoS
5027 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5028 DoS attack where a malformed ticket will result in an OOB read which will
5031 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5032 a custom server callback and ticket lookup mechanism.
5034 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5039 * OOB write in BN_bn2dec()
5041 The function BN_bn2dec() does not check the return value of BN_div_word().
5042 This can cause an OOB write if an application uses this function with an
5043 overly large BIGNUM. This could be a problem if an overly large certificate
5044 or CRL is printed out from an untrusted source. TLS is not affected because
5045 record limits will reject an oversized certificate before it is parsed.
5047 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5052 * OOB read in TS_OBJ_print_bio()
5054 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5055 the total length the OID text representation would use and not the amount
5056 of data written. This will result in OOB reads when large OIDs are
5059 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5064 * Pointer arithmetic undefined behaviour
5066 Avoid some undefined pointer arithmetic
5068 A common idiom in the codebase is to check limits in the following manner:
5071 Where "p" points to some malloc'd data of SIZE bytes and
5074 "len" here could be from some externally supplied data (e.g. from a TLS
5077 The rules of C pointer arithmetic are such that "p + len" is only well
5078 defined where len <= SIZE. Therefore the above idiom is actually
5079 undefined behaviour.
5081 For example this could cause problems if some malloc implementation
5082 provides an address for "p" such that "p + len" actually overflows for
5083 values of len that are too big and therefore p + len < limit.
5085 This issue was reported to OpenSSL by Guido Vranken
5090 * Constant time flag not preserved in DSA signing
5092 Operations in the DSA signing algorithm should run in constant time in
5093 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5094 implementation means that a non-constant time codepath is followed for
5095 certain operations. This has been demonstrated through a cache-timing
5096 attack to be sufficient for an attacker to recover the private DSA key.
5098 This issue was reported by César Pereida (Aalto University), Billy Brumley
5099 (Tampere University of Technology), and Yuval Yarom (The University of
5100 Adelaide and NICTA).
5105 * DTLS buffered message DoS
5107 In a DTLS connection where handshake messages are delivered out-of-order
5108 those messages that OpenSSL is not yet ready to process will be buffered
5109 for later use. Under certain circumstances, a flaw in the logic means that
5110 those messages do not get removed from the buffer even though the handshake
5111 has been completed. An attacker could force up to approx. 15 messages to
5112 remain in the buffer when they are no longer required. These messages will
5113 be cleared when the DTLS connection is closed. The default maximum size for
5114 a message is 100k. Therefore the attacker could force an additional 1500k
5115 to be consumed per connection. By opening many simulataneous connections an
5116 attacker could cause a DoS attack through memory exhaustion.
5118 This issue was reported to OpenSSL by Quan Luo.
5123 * DTLS replay protection DoS
5125 A flaw in the DTLS replay attack protection mechanism means that records
5126 that arrive for future epochs update the replay protection "window" before
5127 the MAC for the record has been validated. This could be exploited by an
5128 attacker by sending a record for the next epoch (which does not have to
5129 decrypt or have a valid MAC), with a very large sequence number. This means
5130 that all subsequent legitimate packets are dropped causing a denial of
5131 service for a specific DTLS connection.
5133 This issue was reported to OpenSSL by the OCAP audit team.
5138 * Certificate message OOB reads
5140 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5141 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5142 theoretical DoS risk but this has not been observed in practice on common
5145 The messages affected are client certificate, client certificate request
5146 and server certificate. As a result the attack can only be performed
5147 against a client or a server which enables client authentication.
5149 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5154 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5156 * Prevent padding oracle in AES-NI CBC MAC check
5158 A MITM attacker can use a padding oracle attack to decrypt traffic
5159 when the connection uses an AES CBC cipher and the server support
5162 This issue was introduced as part of the fix for Lucky 13 padding
5163 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5164 constant time by making sure that always the same bytes are read and
5165 compared against either the MAC or padding bytes. But it no longer
5166 checked that there was enough data to have both the MAC and padding
5169 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5174 * Fix EVP_EncodeUpdate overflow
5176 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5177 Base64 encoding of binary data. If an attacker is able to supply very large
5178 amounts of input data then a length check can overflow resulting in a heap
5181 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5182 the `PEM_write_bio*` family of functions. These are mainly used within the
5183 OpenSSL command line applications, so any application which processes data
5184 from an untrusted source and outputs it as a PEM file should be considered
5185 vulnerable to this issue. User applications that call these APIs directly
5186 with large amounts of untrusted data may also be vulnerable.
5188 This issue was reported by Guido Vranken.
5193 * Fix EVP_EncryptUpdate overflow
5195 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5196 is able to supply very large amounts of input data after a previous call to
5197 EVP_EncryptUpdate() with a partial block then a length check can overflow
5198 resulting in a heap corruption. Following an analysis of all OpenSSL
5199 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5200 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5201 the first called function after an EVP_EncryptInit(), and therefore that
5202 specific call must be safe. The second form is where the length passed to
5203 EVP_EncryptUpdate() can be seen from the code to be some small value and
5204 therefore there is no possibility of an overflow. Since all instances are
5205 one of these two forms, it is believed that there can be no overflows in
5206 internal code due to this problem. It should be noted that
5207 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5208 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5209 of these calls have also been analysed too and it is believed there are no
5210 instances in internal usage where an overflow could occur.
5212 This issue was reported by Guido Vranken.
5217 * Prevent ASN.1 BIO excessive memory allocation
5219 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5220 a short invalid encoding can cause allocation of large amounts of memory
5221 potentially consuming excessive resources or exhausting memory.
5223 Any application parsing untrusted data through d2i BIO functions is
5224 affected. The memory based functions such as d2i_X509() are *not* affected.
5225 Since the memory based functions are used by the TLS library, TLS
5226 applications are not affected.
5228 This issue was reported by Brian Carpenter.
5235 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5236 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5237 in arbitrary stack data being returned in the buffer.
5239 This issue was reported by Guido Vranken.
5244 * Modify behavior of ALPN to invoke callback after SNI/servername
5245 callback, such that updates to the SSL_CTX affect ALPN.
5249 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5254 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5255 methods are enabled and ssl2 is disabled the methods return NULL.
5259 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5261 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5262 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5263 provide any "EXPORT" or "LOW" strength ciphers.
5267 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5268 is by default disabled at build-time. Builds that are not configured with
5269 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5270 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5271 will need to explicitly call either of:
5273 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5275 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5277 as appropriate. Even if either of those is used, or the application
5278 explicitly uses the version-specific SSLv2_method() or its client and
5279 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5280 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5281 ciphers, and SSLv2 56-bit DES are no longer available.
5286 * Fix a double-free in DSA code
5288 A double free bug was discovered when OpenSSL parses malformed DSA private
5289 keys and could lead to a DoS attack or memory corruption for applications
5290 that receive DSA private keys from untrusted sources. This scenario is
5293 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5299 * Disable SRP fake user seed to address a server memory leak.
5301 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5303 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5304 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5305 was changed to ignore the "fake user" SRP seed, even if the seed
5308 Users should use SRP_VBASE_get1_by_user instead. Note that in
5309 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5310 also that even though configuring the SRP seed attempts to hide
5311 invalid usernames by continuing the handshake with fake
5312 credentials, this behaviour is not constant time and no strong
5313 guarantees are made that the handshake is indistinguishable from
5314 that of a valid user.
5319 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5321 In the BN_hex2bn function the number of hex digits is calculated using an
5322 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5323 large values of `i` this can result in `bn_expand` not allocating any
5324 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5325 field as NULL leading to a subsequent NULL ptr deref. For very large values
5326 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5327 In this case memory is allocated to the internal BIGNUM data field, but it
5328 is insufficiently sized leading to heap corruption. A similar issue exists
5329 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5330 is ever called by user applications with very large untrusted hex/dec data.
5331 This is anticipated to be a rare occurrence.
5333 All OpenSSL internal usage of these functions use data that is not expected
5334 to be untrusted, e.g. config file data or application command line
5335 arguments. If user developed applications generate config file data based
5336 on untrusted data then it is possible that this could also lead to security
5337 consequences. This is also anticipated to be rare.
5339 This issue was reported to OpenSSL by Guido Vranken.
5344 * Fix memory issues in `BIO_*printf` functions
5346 The internal `fmtstr` function used in processing a "%s" format string in
5347 the `BIO_*printf` functions could overflow while calculating the length of a
5348 string and cause an OOB read when printing very long strings.
5350 Additionally the internal `doapr_outch` function can attempt to write to an
5351 OOB memory location (at an offset from the NULL pointer) in the event of a
5352 memory allocation failure. In 1.0.2 and below this could be caused where
5353 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5354 could be in processing a very long "%s" format string. Memory leaks can
5357 The first issue may mask the second issue dependent on compiler behaviour.
5358 These problems could enable attacks where large amounts of untrusted data
5359 is passed to the `BIO_*printf` functions. If applications use these functions
5360 in this way then they could be vulnerable. OpenSSL itself uses these
5361 functions when printing out human-readable dumps of ASN.1 data. Therefore
5362 applications that print this data could be vulnerable if the data is from
5363 untrusted sources. OpenSSL command line applications could also be
5364 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5365 as command line arguments.
5367 Libssl is not considered directly vulnerable. Additionally certificates etc
5368 received via remote connections via libssl are also unlikely to be able to
5369 trigger these issues because of message size limits enforced within libssl.
5371 This issue was reported to OpenSSL Guido Vranken.
5376 * Side channel attack on modular exponentiation
5378 A side-channel attack was found which makes use of cache-bank conflicts on
5379 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5380 of RSA keys. The ability to exploit this issue is limited as it relies on
5381 an attacker who has control of code in a thread running on the same
5382 hyper-threaded core as the victim thread which is performing decryptions.
5384 This issue was reported to OpenSSL by Yuval Yarom, The University of
5385 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5386 Nadia Heninger, University of Pennsylvania with more information at
5387 <http://cachebleed.info>.
5392 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5393 if no keysize is specified with default_bits. This fixes an
5394 omission in an earlier change that changed all RSA/DSA key generation
5395 commands to use 2048 bits by default.
5399 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5401 * DH small subgroups
5403 Historically OpenSSL only ever generated DH parameters based on "safe"
5404 primes. More recently (in version 1.0.2) support was provided for
5405 generating X9.42 style parameter files such as those required for RFC 5114
5406 support. The primes used in such files may not be "safe". Where an
5407 application is using DH configured with parameters based on primes that are
5408 not "safe" then an attacker could use this fact to find a peer's private
5409 DH exponent. This attack requires that the attacker complete multiple
5410 handshakes in which the peer uses the same private DH exponent. For example
5411 this could be used to discover a TLS server's private DH exponent if it's
5412 reusing the private DH exponent or it's using a static DH ciphersuite.
5414 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5415 TLS. It is not on by default. If the option is not set then the server
5416 reuses the same private DH exponent for the life of the server process and
5417 would be vulnerable to this attack. It is believed that many popular
5418 applications do set this option and would therefore not be at risk.
5420 The fix for this issue adds an additional check where a "q" parameter is
5421 available (as is the case in X9.42 based parameters). This detects the
5422 only known attack, and is the only possible defense for static DH
5423 ciphersuites. This could have some performance impact.
5425 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5426 default and cannot be disabled. This could have some performance impact.
5428 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5433 * SSLv2 doesn't block disabled ciphers
5435 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5436 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5437 been disabled, provided that the SSLv2 protocol was not also disabled via
5440 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5441 and Sebastian Schinzel.
5446 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5448 * BN_mod_exp may produce incorrect results on x86_64
5450 There is a carry propagating bug in the x86_64 Montgomery squaring
5451 procedure. No EC algorithms are affected. Analysis suggests that attacks
5452 against RSA and DSA as a result of this defect would be very difficult to
5453 perform and are not believed likely. Attacks against DH are considered just
5454 feasible (although very difficult) because most of the work necessary to
5455 deduce information about a private key may be performed offline. The amount
5456 of resources required for such an attack would be very significant and
5457 likely only accessible to a limited number of attackers. An attacker would
5458 additionally need online access to an unpatched system using the target
5459 private key in a scenario with persistent DH parameters and a private
5460 key that is shared between multiple clients. For example this can occur by
5461 default in OpenSSL DHE based SSL/TLS ciphersuites.
5463 This issue was reported to OpenSSL by Hanno Böck.
5468 * Certificate verify crash with missing PSS parameter
5470 The signature verification routines will crash with a NULL pointer
5471 dereference if presented with an ASN.1 signature using the RSA PSS
5472 algorithm and absent mask generation function parameter. Since these
5473 routines are used to verify certificate signature algorithms this can be
5474 used to crash any certificate verification operation and exploited in a
5475 DoS attack. Any application which performs certificate verification is
5476 vulnerable including OpenSSL clients and servers which enable client
5479 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5484 * X509_ATTRIBUTE memory leak
5486 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5487 memory. This structure is used by the PKCS#7 and CMS routines so any
5488 application which reads PKCS#7 or CMS data from untrusted sources is
5489 affected. SSL/TLS is not affected.
5491 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5497 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5498 This changes the decoding behaviour for some invalid messages,
5499 though the change is mostly in the more lenient direction, and
5500 legacy behaviour is preserved as much as possible.
5504 * In DSA_generate_parameters_ex, if the provided seed is too short,
5507 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5509 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5511 * Alternate chains certificate forgery
5513 During certificate verification, OpenSSL will attempt to find an
5514 alternative certificate chain if the first attempt to build such a chain
5515 fails. An error in the implementation of this logic can mean that an
5516 attacker could cause certain checks on untrusted certificates to be
5517 bypassed, such as the CA flag, enabling them to use a valid leaf
5518 certificate to act as a CA and "issue" an invalid certificate.
5520 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5525 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5527 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5528 incompatibility in the handling of HMAC. The previous ABI has now been
5533 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5535 * Malformed ECParameters causes infinite loop
5537 When processing an ECParameters structure OpenSSL enters an infinite loop
5538 if the curve specified is over a specially malformed binary polynomial
5541 This can be used to perform denial of service against any
5542 system which processes public keys, certificate requests or
5543 certificates. This includes TLS clients and TLS servers with
5544 client authentication enabled.
5546 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5551 * Exploitable out-of-bounds read in X509_cmp_time
5553 X509_cmp_time does not properly check the length of the ASN1_TIME
5554 string and can read a few bytes out of bounds. In addition,
5555 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5558 An attacker can use this to craft malformed certificates and CRLs of
5559 various sizes and potentially cause a segmentation fault, resulting in
5560 a DoS on applications that verify certificates or CRLs. TLS clients
5561 that verify CRLs are affected. TLS clients and servers with client
5562 authentication enabled may be affected if they use custom verification
5565 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5566 independently by Hanno Böck.
5571 * PKCS7 crash with missing EnvelopedContent
5573 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5574 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5575 with missing content and trigger a NULL pointer dereference on parsing.
5577 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5578 structures from untrusted sources are affected. OpenSSL clients and
5579 servers are not affected.
5581 This issue was reported to OpenSSL by Michal Zalewski (Google).
5586 * CMS verify infinite loop with unknown hash function
5588 When verifying a signedData message the CMS code can enter an infinite loop
5589 if presented with an unknown hash function OID. This can be used to perform
5590 denial of service against any system which verifies signedData messages using
5592 This issue was reported to OpenSSL by Johannes Bauer.
5597 * Race condition handling NewSessionTicket
5599 If a NewSessionTicket is received by a multi-threaded client when attempting to
5600 reuse a previous ticket then a race condition can occur potentially leading to
5601 a double free of the ticket data.
5606 * Only support 256-bit or stronger elliptic curves with the
5607 'ecdh_auto' setting (server) or by default (client). Of supported
5608 curves, prefer P-256 (both).
5612 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5614 * ClientHello sigalgs DoS fix
5616 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5617 invalid signature algorithms extension a NULL pointer dereference will
5618 occur. This can be exploited in a DoS attack against the server.
5620 This issue was was reported to OpenSSL by David Ramos of Stanford
5624 *Stephen Henson and Matt Caswell*
5626 * Multiblock corrupted pointer fix
5628 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5629 feature only applies on 64 bit x86 architecture platforms that support AES
5630 NI instructions. A defect in the implementation of "multiblock" can cause
5631 OpenSSL's internal write buffer to become incorrectly set to NULL when
5632 using non-blocking IO. Typically, when the user application is using a
5633 socket BIO for writing, this will only result in a failed connection.
5634 However if some other BIO is used then it is likely that a segmentation
5635 fault will be triggered, thus enabling a potential DoS attack.
5637 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5642 * Segmentation fault in DTLSv1_listen fix
5644 The DTLSv1_listen function is intended to be stateless and processes the
5645 initial ClientHello from many peers. It is common for user code to loop
5646 over the call to DTLSv1_listen until a valid ClientHello is received with
5647 an associated cookie. A defect in the implementation of DTLSv1_listen means
5648 that state is preserved in the SSL object from one invocation to the next
5649 that can lead to a segmentation fault. Errors processing the initial
5650 ClientHello can trigger this scenario. An example of such an error could be
5651 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5654 This issue was reported to OpenSSL by Per Allansson.
5659 * Segmentation fault in ASN1_TYPE_cmp fix
5661 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5662 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5663 certificate signature algorithm consistency this can be used to crash any
5664 certificate verification operation and exploited in a DoS attack. Any
5665 application which performs certificate verification is vulnerable including
5666 OpenSSL clients and servers which enable client authentication.
5671 * Segmentation fault for invalid PSS parameters fix
5673 The signature verification routines will crash with a NULL pointer
5674 dereference if presented with an ASN.1 signature using the RSA PSS
5675 algorithm and invalid parameters. Since these routines are used to verify
5676 certificate signature algorithms this can be used to crash any
5677 certificate verification operation and exploited in a DoS attack. Any
5678 application which performs certificate verification is vulnerable including
5679 OpenSSL clients and servers which enable client authentication.
5681 This issue was was reported to OpenSSL by Brian Carpenter.
5686 * ASN.1 structure reuse memory corruption fix
5688 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5689 memory corruption via an invalid write. Such reuse is and has been
5690 strongly discouraged and is believed to be rare.
5692 Applications that parse structures containing CHOICE or ANY DEFINED BY
5693 components may be affected. Certificate parsing (d2i_X509 and related
5694 functions) are however not affected. OpenSSL clients and servers are
5700 * PKCS7 NULL pointer dereferences fix
5702 The PKCS#7 parsing code does not handle missing outer ContentInfo
5703 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5704 missing content and trigger a NULL pointer dereference on parsing.
5706 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5707 otherwise parse PKCS#7 structures from untrusted sources are
5708 affected. OpenSSL clients and servers are not affected.
5710 This issue was reported to OpenSSL by Michal Zalewski (Google).
5715 * DoS via reachable assert in SSLv2 servers fix
5717 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5718 servers that both support SSLv2 and enable export cipher suites by sending
5719 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5721 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5722 (OpenSSL development team).
5727 * Empty CKE with client auth and DHE fix
5729 If client auth is used then a server can seg fault in the event of a DHE
5730 ciphersuite being selected and a zero length ClientKeyExchange message
5731 being sent by the client. This could be exploited in a DoS attack.
5736 * Handshake with unseeded PRNG fix
5738 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5739 with an unseeded PRNG. The conditions are:
5740 - The client is on a platform where the PRNG has not been seeded
5741 automatically, and the user has not seeded manually
5742 - A protocol specific client method version has been used (i.e. not
5743 SSL_client_methodv23)
5744 - A ciphersuite is used that does not require additional random data from
5745 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5747 If the handshake succeeds then the client random that has been used will
5748 have been generated from a PRNG with insufficient entropy and therefore the
5749 output may be predictable.
5751 For example using the following command with an unseeded openssl will
5752 succeed on an unpatched platform:
5754 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5759 * Use After Free following d2i_ECPrivatekey error fix
5761 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5762 could cause a use after free condition. This, in turn, could cause a double
5763 free in several private key parsing functions (such as d2i_PrivateKey
5764 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5765 for applications that receive EC private keys from untrusted
5766 sources. This scenario is considered rare.
5768 This issue was discovered by the BoringSSL project and fixed in their
5774 * X509_to_X509_REQ NULL pointer deref fix
5776 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5777 the certificate key is invalid. This function is rarely used in practice.
5779 This issue was discovered by Brian Carpenter.
5784 * Removed the export ciphers from the DEFAULT ciphers
5788 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5790 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5791 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5792 So far those who have to target multiple platforms would compromise
5793 and argue that binary targeting say ARMv5 would still execute on
5794 ARMv8. "Universal" build resolves this compromise by providing
5795 near-optimal performance even on newer platforms.
5799 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5800 (other platforms pending).
5802 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5804 * Add support for the SignedCertificateTimestampList certificate and
5805 OCSP response extensions from RFC6962.
5809 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5810 for corner cases. (Certain input points at infinity could lead to
5811 bogus results, with non-infinity inputs mapped to infinity too.)
5815 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5816 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5817 common cases are optimized and there still is room for further
5818 improvements. Vector Permutation AES for Altivec is also added.
5822 * Add support for little-endian ppc64 Linux target.
5824 *Marcelo Cerri (IBM)*
5826 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5827 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5828 are optimized and there still is room for further improvements.
5829 Both 32- and 64-bit modes are supported.
5831 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5833 * Improved ARMv7 NEON support.
5837 * Support for SPARC Architecture 2011 crypto extensions, first
5838 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5839 SHA256/512, MD5, GHASH and modular exponentiation.
5841 *Andy Polyakov, David Miller*
5843 * Accelerated modular exponentiation for Intel processors, a.k.a.
5846 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5848 * Support for new and upcoming Intel processors, including AVX2,
5849 BMI and SHA ISA extensions. This includes additional "stitched"
5850 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5853 This work was sponsored by Intel Corp.
5857 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5858 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5859 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5863 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5864 this fixes a limitation in previous versions of OpenSSL.
5868 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5869 MGF1 digest and OAEP label.
5873 * Add EVP support for key wrapping algorithms, to avoid problems with
5874 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5875 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5876 algorithms and include tests cases.
5880 * Add functions to allocate and set the fields of an ECDSA_METHOD
5883 *Douglas E. Engert, Steve Henson*
5885 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5886 difference in days and seconds between two tm or ASN1_TIME structures.
5890 * Add -rev test option to s_server to just reverse order of characters
5891 received by client and send back to server. Also prints an abbreviated
5892 summary of the connection parameters.
5896 * New option -brief for s_client and s_server to print out a brief summary
5897 of connection parameters.
5901 * Add callbacks for arbitrary TLS extensions.
5903 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5905 * New option -crl_download in several openssl utilities to download CRLs
5906 from CRLDP extension in certificates.
5910 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5914 * New function X509_CRL_diff to generate a delta CRL from the difference
5915 of two full CRLs. Add support to "crl" utility.
5919 * New functions to set lookup_crls function and to retrieve
5920 X509_STORE from X509_STORE_CTX.
5924 * Print out deprecated issuer and subject unique ID fields in
5929 * Extend OCSP I/O functions so they can be used for simple general purpose
5930 HTTP as well as OCSP. New wrapper function which can be used to download
5931 CRLs using the OCSP API.
5935 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5939 * `SSL_CONF*` functions. These provide a common framework for application
5940 configuration using configuration files or command lines.
5944 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5945 message callback and prints the results. Needs compile time option
5946 "enable-ssl-trace". New options to s_client and s_server to enable
5951 * New ctrl and macro to retrieve supported points extensions.
5952 Print out extension in s_server and s_client.
5956 * New functions to retrieve certificate signature and signature
5961 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5966 * New Suite B modes for TLS code. These use and enforce the requirements
5967 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5968 only use Suite B curves. The Suite B modes can be set by using the
5969 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5973 * New chain verification flags for Suite B levels of security. Check
5974 algorithms are acceptable when flags are set in X509_verify_cert.
5978 * Make tls1_check_chain return a set of flags indicating checks passed
5979 by a certificate chain. Add additional tests to handle client
5980 certificates: checks for matching certificate type and issuer name
5985 * If an attempt is made to use a signature algorithm not in the peer
5986 preference list abort the handshake. If client has no suitable
5987 signature algorithms in response to a certificate request do not
5988 use the certificate.
5992 * If server EC tmp key is not in client preference list abort handshake.
5996 * Add support for certificate stores in CERT structure. This makes it
5997 possible to have different stores per SSL structure or one store in
5998 the parent SSL_CTX. Include distinct stores for certificate chain
5999 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6000 to build and store a certificate chain in CERT structure: returning
6001 an error if the chain cannot be built: this will allow applications
6002 to test if a chain is correctly configured.
6004 Note: if the CERT based stores are not set then the parent SSL_CTX
6005 store is used to retain compatibility with existing behaviour.
6009 * New function ssl_set_client_disabled to set a ciphersuite disabled
6010 mask based on the current session, check mask when sending client
6011 hello and checking the requested ciphersuite.
6015 * New ctrls to retrieve and set certificate types in a certificate
6016 request message. Print out received values in s_client. If certificate
6017 types is not set with custom values set sensible values based on
6018 supported signature algorithms.
6022 * Support for distinct client and server supported signature algorithms.
6026 * Add certificate callback. If set this is called whenever a certificate
6027 is required by client or server. An application can decide which
6028 certificate chain to present based on arbitrary criteria: for example
6029 supported signature algorithms. Add very simple example to s_server.
6030 This fixes many of the problems and restrictions of the existing client
6031 certificate callback: for example you can now clear an existing
6032 certificate and specify the whole chain.
6036 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6037 the certificate can be used for (if anything). Set valid_flags field
6038 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6039 to have similar checks in it.
6041 Add new "cert_flags" field to CERT structure and include a "strict mode".
6042 This enforces some TLS certificate requirements (such as only permitting
6043 certificate signature algorithms contained in the supported algorithms
6044 extension) which some implementations ignore: this option should be used
6045 with caution as it could cause interoperability issues.
6049 * Update and tidy signature algorithm extension processing. Work out
6050 shared signature algorithms based on preferences and peer algorithms
6051 and print them out in s_client and s_server. Abort handshake if no
6052 shared signature algorithms.
6056 * Add new functions to allow customised supported signature algorithms
6057 for SSL and SSL_CTX structures. Add options to s_client and s_server
6062 * New function SSL_certs_clear() to delete all references to certificates
6063 from an SSL structure. Before this once a certificate had been added
6064 it couldn't be removed.
6068 * Integrate hostname, email address and IP address checking with certificate
6069 verification. New verify options supporting checking in openssl utility.
6073 * Fixes and wildcard matching support to hostname and email checking
6074 functions. Add manual page.
6076 *Florian Weimer (Red Hat Product Security Team)*
6078 * New functions to check a hostname email or IP address against a
6079 certificate. Add options x509 utility to print results of checks against
6084 * Fix OCSP checking.
6086 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6088 * Initial experimental support for explicitly trusted non-root CAs.
6089 OpenSSL still tries to build a complete chain to a root but if an
6090 intermediate CA has a trust setting included that is used. The first
6091 setting is used: whether to trust (e.g., -addtrust option to the x509
6096 * Add -trusted_first option which attempts to find certificates in the
6097 trusted store even if an untrusted chain is also supplied.
6101 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6102 platform support for Linux and Android.
6106 * Support for linux-x32, ILP32 environment in x86_64 framework.
6110 * Experimental multi-implementation support for FIPS capable OpenSSL.
6111 When in FIPS mode the approved implementations are used as normal,
6112 when not in FIPS mode the internal unapproved versions are used instead.
6113 This means that the FIPS capable OpenSSL isn't forced to use the
6114 (often lower performance) FIPS implementations outside FIPS mode.
6118 * Transparently support X9.42 DH parameters when calling
6119 PEM_read_bio_DHparameters. This means existing applications can handle
6120 the new parameter format automatically.
6124 * Initial experimental support for X9.42 DH parameter format: mainly
6125 to support use of 'q' parameter for RFC5114 parameters.
6129 * Add DH parameters from RFC5114 including test data to dhtest.
6133 * Support for automatic EC temporary key parameter selection. If enabled
6134 the most preferred EC parameters are automatically used instead of
6135 hardcoded fixed parameters. Now a server just has to call:
6136 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6137 support ECDH and use the most appropriate parameters.
6141 * Enhance and tidy EC curve and point format TLS extension code. Use
6142 static structures instead of allocation if default values are used.
6143 New ctrls to set curves we wish to support and to retrieve shared curves.
6144 Print out shared curves in s_server. New options to s_server and s_client
6145 to set list of supported curves.
6149 * New ctrls to retrieve supported signature algorithms and
6150 supported curve values as an array of NIDs. Extend openssl utility
6151 to print out received values.
6155 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6156 between NIDs and the more common NIST names such as "P-256". Enhance
6157 ecparam utility and ECC method to recognise the NIST names for curves.
6161 * Enhance SSL/TLS certificate chain handling to support different
6162 chains for each certificate instead of one chain in the parent SSL_CTX.
6166 * Support for fixed DH ciphersuite client authentication: where both
6167 server and client use DH certificates with common parameters.
6171 * Support for fixed DH ciphersuites: those requiring DH server
6176 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6178 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6179 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6180 X509_CINF_get_signature were reverted post internal team review.
6185 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6187 * OCSP Status Request extension unbounded memory growth
6189 A malicious client can send an excessively large OCSP Status Request
6190 extension. If that client continually requests renegotiation, sending a
6191 large OCSP Status Request extension each time, then there will be unbounded
6192 memory growth on the server. This will eventually lead to a Denial Of
6193 Service attack through memory exhaustion. Servers with a default
6194 configuration are vulnerable even if they do not support OCSP. Builds using
6195 the "no-ocsp" build time option are not affected.
6197 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6202 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6205 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6211 * OOB write in MDC2_Update()
6213 An overflow can occur in MDC2_Update() either if called directly or
6214 through the EVP_DigestUpdate() function using MDC2. If an attacker
6215 is able to supply very large amounts of input data after a previous
6216 call to EVP_EncryptUpdate() with a partial block then a length check
6217 can overflow resulting in a heap corruption.
6219 The amount of data needed is comparable to SIZE_MAX which is impractical
6222 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6227 * Malformed SHA512 ticket DoS
6229 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6230 DoS attack where a malformed ticket will result in an OOB read which will
6233 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6234 a custom server callback and ticket lookup mechanism.
6236 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6241 * OOB write in BN_bn2dec()
6243 The function BN_bn2dec() does not check the return value of BN_div_word().
6244 This can cause an OOB write if an application uses this function with an
6245 overly large BIGNUM. This could be a problem if an overly large certificate
6246 or CRL is printed out from an untrusted source. TLS is not affected because
6247 record limits will reject an oversized certificate before it is parsed.
6249 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6254 * OOB read in TS_OBJ_print_bio()
6256 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6257 the total length the OID text representation would use and not the amount
6258 of data written. This will result in OOB reads when large OIDs are
6261 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6266 * Pointer arithmetic undefined behaviour
6268 Avoid some undefined pointer arithmetic
6270 A common idiom in the codebase is to check limits in the following manner:
6273 Where "p" points to some malloc'd data of SIZE bytes and
6276 "len" here could be from some externally supplied data (e.g. from a TLS
6279 The rules of C pointer arithmetic are such that "p + len" is only well
6280 defined where len <= SIZE. Therefore the above idiom is actually
6281 undefined behaviour.
6283 For example this could cause problems if some malloc implementation
6284 provides an address for "p" such that "p + len" actually overflows for
6285 values of len that are too big and therefore p + len < limit.
6287 This issue was reported to OpenSSL by Guido Vranken
6292 * Constant time flag not preserved in DSA signing
6294 Operations in the DSA signing algorithm should run in constant time in
6295 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6296 implementation means that a non-constant time codepath is followed for
6297 certain operations. This has been demonstrated through a cache-timing
6298 attack to be sufficient for an attacker to recover the private DSA key.
6300 This issue was reported by César Pereida (Aalto University), Billy Brumley
6301 (Tampere University of Technology), and Yuval Yarom (The University of
6302 Adelaide and NICTA).
6307 * DTLS buffered message DoS
6309 In a DTLS connection where handshake messages are delivered out-of-order
6310 those messages that OpenSSL is not yet ready to process will be buffered
6311 for later use. Under certain circumstances, a flaw in the logic means that
6312 those messages do not get removed from the buffer even though the handshake
6313 has been completed. An attacker could force up to approx. 15 messages to
6314 remain in the buffer when they are no longer required. These messages will
6315 be cleared when the DTLS connection is closed. The default maximum size for
6316 a message is 100k. Therefore the attacker could force an additional 1500k
6317 to be consumed per connection. By opening many simulataneous connections an
6318 attacker could cause a DoS attack through memory exhaustion.
6320 This issue was reported to OpenSSL by Quan Luo.
6325 * DTLS replay protection DoS
6327 A flaw in the DTLS replay attack protection mechanism means that records
6328 that arrive for future epochs update the replay protection "window" before
6329 the MAC for the record has been validated. This could be exploited by an
6330 attacker by sending a record for the next epoch (which does not have to
6331 decrypt or have a valid MAC), with a very large sequence number. This means
6332 that all subsequent legitimate packets are dropped causing a denial of
6333 service for a specific DTLS connection.
6335 This issue was reported to OpenSSL by the OCAP audit team.
6340 * Certificate message OOB reads
6342 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6343 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6344 theoretical DoS risk but this has not been observed in practice on common
6347 The messages affected are client certificate, client certificate request
6348 and server certificate. As a result the attack can only be performed
6349 against a client or a server which enables client authentication.
6351 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6356 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6358 * Prevent padding oracle in AES-NI CBC MAC check
6360 A MITM attacker can use a padding oracle attack to decrypt traffic
6361 when the connection uses an AES CBC cipher and the server support
6364 This issue was introduced as part of the fix for Lucky 13 padding
6365 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6366 constant time by making sure that always the same bytes are read and
6367 compared against either the MAC or padding bytes. But it no longer
6368 checked that there was enough data to have both the MAC and padding
6371 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6376 * Fix EVP_EncodeUpdate overflow
6378 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6379 Base64 encoding of binary data. If an attacker is able to supply very large
6380 amounts of input data then a length check can overflow resulting in a heap
6383 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6384 the `PEM_write_bio*` family of functions. These are mainly used within the
6385 OpenSSL command line applications, so any application which processes data
6386 from an untrusted source and outputs it as a PEM file should be considered
6387 vulnerable to this issue. User applications that call these APIs directly
6388 with large amounts of untrusted data may also be vulnerable.
6390 This issue was reported by Guido Vranken.
6395 * Fix EVP_EncryptUpdate overflow
6397 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6398 is able to supply very large amounts of input data after a previous call to
6399 EVP_EncryptUpdate() with a partial block then a length check can overflow
6400 resulting in a heap corruption. Following an analysis of all OpenSSL
6401 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6402 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6403 the first called function after an EVP_EncryptInit(), and therefore that
6404 specific call must be safe. The second form is where the length passed to
6405 EVP_EncryptUpdate() can be seen from the code to be some small value and
6406 therefore there is no possibility of an overflow. Since all instances are
6407 one of these two forms, it is believed that there can be no overflows in
6408 internal code due to this problem. It should be noted that
6409 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6410 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6411 of these calls have also been analysed too and it is believed there are no
6412 instances in internal usage where an overflow could occur.
6414 This issue was reported by Guido Vranken.
6419 * Prevent ASN.1 BIO excessive memory allocation
6421 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6422 a short invalid encoding can casuse allocation of large amounts of memory
6423 potentially consuming excessive resources or exhausting memory.
6425 Any application parsing untrusted data through d2i BIO functions is
6426 affected. The memory based functions such as d2i_X509() are *not* affected.
6427 Since the memory based functions are used by the TLS library, TLS
6428 applications are not affected.
6430 This issue was reported by Brian Carpenter.
6437 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6438 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6439 in arbitrary stack data being returned in the buffer.
6441 This issue was reported by Guido Vranken.
6446 * Modify behavior of ALPN to invoke callback after SNI/servername
6447 callback, such that updates to the SSL_CTX affect ALPN.
6451 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6456 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6457 methods are enabled and ssl2 is disabled the methods return NULL.
6461 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6463 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6464 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6465 provide any "EXPORT" or "LOW" strength ciphers.
6469 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6470 is by default disabled at build-time. Builds that are not configured with
6471 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6472 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6473 will need to explicitly call either of:
6475 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6477 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6479 as appropriate. Even if either of those is used, or the application
6480 explicitly uses the version-specific SSLv2_method() or its client and
6481 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6482 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6483 ciphers, and SSLv2 56-bit DES are no longer available.
6488 * Fix a double-free in DSA code
6490 A double free bug was discovered when OpenSSL parses malformed DSA private
6491 keys and could lead to a DoS attack or memory corruption for applications
6492 that receive DSA private keys from untrusted sources. This scenario is
6495 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6501 * Disable SRP fake user seed to address a server memory leak.
6503 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6505 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6506 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6507 was changed to ignore the "fake user" SRP seed, even if the seed
6510 Users should use SRP_VBASE_get1_by_user instead. Note that in
6511 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6512 also that even though configuring the SRP seed attempts to hide
6513 invalid usernames by continuing the handshake with fake
6514 credentials, this behaviour is not constant time and no strong
6515 guarantees are made that the handshake is indistinguishable from
6516 that of a valid user.
6521 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6523 In the BN_hex2bn function the number of hex digits is calculated using an
6524 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6525 large values of `i` this can result in `bn_expand` not allocating any
6526 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6527 field as NULL leading to a subsequent NULL ptr deref. For very large values
6528 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6529 In this case memory is allocated to the internal BIGNUM data field, but it
6530 is insufficiently sized leading to heap corruption. A similar issue exists
6531 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6532 is ever called by user applications with very large untrusted hex/dec data.
6533 This is anticipated to be a rare occurrence.
6535 All OpenSSL internal usage of these functions use data that is not expected
6536 to be untrusted, e.g. config file data or application command line
6537 arguments. If user developed applications generate config file data based
6538 on untrusted data then it is possible that this could also lead to security
6539 consequences. This is also anticipated to be rare.
6541 This issue was reported to OpenSSL by Guido Vranken.
6546 * Fix memory issues in `BIO_*printf` functions
6548 The internal `fmtstr` function used in processing a "%s" format string in
6549 the `BIO_*printf` functions could overflow while calculating the length of a
6550 string and cause an OOB read when printing very long strings.
6552 Additionally the internal `doapr_outch` function can attempt to write to an
6553 OOB memory location (at an offset from the NULL pointer) in the event of a
6554 memory allocation failure. In 1.0.2 and below this could be caused where
6555 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6556 could be in processing a very long "%s" format string. Memory leaks can
6559 The first issue may mask the second issue dependent on compiler behaviour.
6560 These problems could enable attacks where large amounts of untrusted data
6561 is passed to the `BIO_*printf` functions. If applications use these functions
6562 in this way then they could be vulnerable. OpenSSL itself uses these
6563 functions when printing out human-readable dumps of ASN.1 data. Therefore
6564 applications that print this data could be vulnerable if the data is from
6565 untrusted sources. OpenSSL command line applications could also be
6566 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6567 as command line arguments.
6569 Libssl is not considered directly vulnerable. Additionally certificates etc
6570 received via remote connections via libssl are also unlikely to be able to
6571 trigger these issues because of message size limits enforced within libssl.
6573 This issue was reported to OpenSSL Guido Vranken.
6578 * Side channel attack on modular exponentiation
6580 A side-channel attack was found which makes use of cache-bank conflicts on
6581 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6582 of RSA keys. The ability to exploit this issue is limited as it relies on
6583 an attacker who has control of code in a thread running on the same
6584 hyper-threaded core as the victim thread which is performing decryptions.
6586 This issue was reported to OpenSSL by Yuval Yarom, The University of
6587 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6588 Nadia Heninger, University of Pennsylvania with more information at
6589 <http://cachebleed.info>.
6594 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6595 if no keysize is specified with default_bits. This fixes an
6596 omission in an earlier change that changed all RSA/DSA key generation
6597 commands to use 2048 bits by default.
6601 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6603 * Protection for DH small subgroup attacks
6605 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6606 switched on by default and cannot be disabled. This could have some
6611 * SSLv2 doesn't block disabled ciphers
6613 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6614 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6615 been disabled, provided that the SSLv2 protocol was not also disabled via
6618 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6619 and Sebastian Schinzel.
6624 * Reject DH handshakes with parameters shorter than 1024 bits.
6628 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6630 * Certificate verify crash with missing PSS parameter
6632 The signature verification routines will crash with a NULL pointer
6633 dereference if presented with an ASN.1 signature using the RSA PSS
6634 algorithm and absent mask generation function parameter. Since these
6635 routines are used to verify certificate signature algorithms this can be
6636 used to crash any certificate verification operation and exploited in a
6637 DoS attack. Any application which performs certificate verification is
6638 vulnerable including OpenSSL clients and servers which enable client
6641 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6646 * X509_ATTRIBUTE memory leak
6648 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6649 memory. This structure is used by the PKCS#7 and CMS routines so any
6650 application which reads PKCS#7 or CMS data from untrusted sources is
6651 affected. SSL/TLS is not affected.
6653 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6659 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6660 This changes the decoding behaviour for some invalid messages,
6661 though the change is mostly in the more lenient direction, and
6662 legacy behaviour is preserved as much as possible.
6666 * In DSA_generate_parameters_ex, if the provided seed is too short,
6667 use a random seed, as already documented.
6669 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6671 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6673 * Alternate chains certificate forgery
6675 During certificate verfification, OpenSSL will attempt to find an
6676 alternative certificate chain if the first attempt to build such a chain
6677 fails. An error in the implementation of this logic can mean that an
6678 attacker could cause certain checks on untrusted certificates to be
6679 bypassed, such as the CA flag, enabling them to use a valid leaf
6680 certificate to act as a CA and "issue" an invalid certificate.
6682 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6688 * Race condition handling PSK identify hint
6690 If PSK identity hints are received by a multi-threaded client then
6691 the values are wrongly updated in the parent SSL_CTX structure. This can
6692 result in a race condition potentially leading to a double free of the
6698 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6700 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6701 incompatibility in the handling of HMAC. The previous ABI has now been
6704 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6706 * Malformed ECParameters causes infinite loop
6708 When processing an ECParameters structure OpenSSL enters an infinite loop
6709 if the curve specified is over a specially malformed binary polynomial
6712 This can be used to perform denial of service against any
6713 system which processes public keys, certificate requests or
6714 certificates. This includes TLS clients and TLS servers with
6715 client authentication enabled.
6717 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6722 * Exploitable out-of-bounds read in X509_cmp_time
6724 X509_cmp_time does not properly check the length of the ASN1_TIME
6725 string and can read a few bytes out of bounds. In addition,
6726 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6729 An attacker can use this to craft malformed certificates and CRLs of
6730 various sizes and potentially cause a segmentation fault, resulting in
6731 a DoS on applications that verify certificates or CRLs. TLS clients
6732 that verify CRLs are affected. TLS clients and servers with client
6733 authentication enabled may be affected if they use custom verification
6736 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6737 independently by Hanno Böck.
6742 * PKCS7 crash with missing EnvelopedContent
6744 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6745 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6746 with missing content and trigger a NULL pointer dereference on parsing.
6748 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6749 structures from untrusted sources are affected. OpenSSL clients and
6750 servers are not affected.
6752 This issue was reported to OpenSSL by Michal Zalewski (Google).
6757 * CMS verify infinite loop with unknown hash function
6759 When verifying a signedData message the CMS code can enter an infinite loop
6760 if presented with an unknown hash function OID. This can be used to perform
6761 denial of service against any system which verifies signedData messages using
6763 This issue was reported to OpenSSL by Johannes Bauer.
6768 * Race condition handling NewSessionTicket
6770 If a NewSessionTicket is received by a multi-threaded client when attempting to
6771 reuse a previous ticket then a race condition can occur potentially leading to
6772 a double free of the ticket data.
6777 * Reject DH handshakes with parameters shorter than 768 bits.
6779 *Kurt Roeckx and Emilia Kasper*
6781 * dhparam: generate 2048-bit parameters by default.
6783 *Kurt Roeckx and Emilia Kasper*
6785 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6787 * Segmentation fault in ASN1_TYPE_cmp fix
6789 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6790 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6791 certificate signature algorithm consistency this can be used to crash any
6792 certificate verification operation and exploited in a DoS attack. Any
6793 application which performs certificate verification is vulnerable including
6794 OpenSSL clients and servers which enable client authentication.
6799 * ASN.1 structure reuse memory corruption fix
6801 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6802 memory corruption via an invalid write. Such reuse is and has been
6803 strongly discouraged and is believed to be rare.
6805 Applications that parse structures containing CHOICE or ANY DEFINED BY
6806 components may be affected. Certificate parsing (d2i_X509 and related
6807 functions) are however not affected. OpenSSL clients and servers are
6813 * PKCS7 NULL pointer dereferences fix
6815 The PKCS#7 parsing code does not handle missing outer ContentInfo
6816 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6817 missing content and trigger a NULL pointer dereference on parsing.
6819 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6820 otherwise parse PKCS#7 structures from untrusted sources are
6821 affected. OpenSSL clients and servers are not affected.
6823 This issue was reported to OpenSSL by Michal Zalewski (Google).
6828 * DoS via reachable assert in SSLv2 servers fix
6830 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6831 servers that both support SSLv2 and enable export cipher suites by sending
6832 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6834 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6835 (OpenSSL development team).
6840 * Use After Free following d2i_ECPrivatekey error fix
6842 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6843 could cause a use after free condition. This, in turn, could cause a double
6844 free in several private key parsing functions (such as d2i_PrivateKey
6845 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6846 for applications that receive EC private keys from untrusted
6847 sources. This scenario is considered rare.
6849 This issue was discovered by the BoringSSL project and fixed in their
6855 * X509_to_X509_REQ NULL pointer deref fix
6857 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6858 the certificate key is invalid. This function is rarely used in practice.
6860 This issue was discovered by Brian Carpenter.
6865 * Removed the export ciphers from the DEFAULT ciphers
6869 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6871 * Build fixes for the Windows and OpenVMS platforms
6873 *Matt Caswell and Richard Levitte*
6875 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6877 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6878 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6879 dereference. This could lead to a Denial Of Service attack. Thanks to
6880 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6885 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6886 dtls1_buffer_record function under certain conditions. In particular this
6887 could occur if an attacker sent repeated DTLS records with the same
6888 sequence number but for the next epoch. The memory leak could be exploited
6889 by an attacker in a Denial of Service attack through memory exhaustion.
6890 Thanks to Chris Mueller for reporting this issue.
6895 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6896 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6897 method would be set to NULL which could later result in a NULL pointer
6898 dereference. Thanks to Frank Schmirler for reporting this issue.
6903 * Abort handshake if server key exchange message is omitted for ephemeral
6906 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6907 reporting this issue.
6912 * Remove non-export ephemeral RSA code on client and server. This code
6913 violated the TLS standard by allowing the use of temporary RSA keys in
6914 non-export ciphersuites and could be used by a server to effectively
6915 downgrade the RSA key length used to a value smaller than the server
6916 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6917 INRIA or reporting this issue.
6922 * Fixed issue where DH client certificates are accepted without verification.
6923 An OpenSSL server will accept a DH certificate for client authentication
6924 without the certificate verify message. This effectively allows a client to
6925 authenticate without the use of a private key. This only affects servers
6926 which trust a client certificate authority which issues certificates
6927 containing DH keys: these are extremely rare and hardly ever encountered.
6928 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6934 * Ensure that the session ID context of an SSL is updated when its
6935 SSL_CTX is updated via SSL_set_SSL_CTX.
6937 The session ID context is typically set from the parent SSL_CTX,
6938 and can vary with the CTX.
6942 * Fix various certificate fingerprint issues.
6944 By using non-DER or invalid encodings outside the signed portion of a
6945 certificate the fingerprint can be changed without breaking the signature.
6946 Although no details of the signed portion of the certificate can be changed
6947 this can cause problems with some applications: e.g. those using the
6948 certificate fingerprint for blacklists.
6950 1. Reject signatures with non zero unused bits.
6952 If the BIT STRING containing the signature has non zero unused bits reject
6953 the signature. All current signature algorithms require zero unused bits.
6955 2. Check certificate algorithm consistency.
6957 Check the AlgorithmIdentifier inside TBS matches the one in the
6958 certificate signature. NB: this will result in signature failure
6959 errors for some broken certificates.
6961 Thanks to Konrad Kraszewski from Google for reporting this issue.
6963 3. Check DSA/ECDSA signatures use DER.
6965 Re-encode DSA/ECDSA signatures and compare with the original received
6966 signature. Return an error if there is a mismatch.
6968 This will reject various cases including garbage after signature
6969 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6970 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6971 (negative or with leading zeroes).
6973 Further analysis was conducted and fixes were developed by Stephen Henson
6974 of the OpenSSL core team.
6980 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6981 results on some platforms, including x86_64. This bug occurs at random
6982 with a very low probability, and is not known to be exploitable in any
6983 way, though its exact impact is difficult to determine. Thanks to Pieter
6984 Wuille (Blockstream) who reported this issue and also suggested an initial
6985 fix. Further analysis was conducted by the OpenSSL development team and
6986 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6987 the OpenSSL core team.
6992 * Do not resume sessions on the server if the negotiated protocol
6993 version does not match the session's version. Resuming with a different
6994 version, while not strictly forbidden by the RFC, is of questionable
6995 sanity and breaks all known clients.
6997 *David Benjamin, Emilia Käsper*
6999 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7000 early CCS messages during renegotiation. (Note that because
7001 renegotiation is encrypted, this early CCS was not exploitable.)
7005 * Tighten client-side session ticket handling during renegotiation:
7006 ensure that the client only accepts a session ticket if the server sends
7007 the extension anew in the ServerHello. Previously, a TLS client would
7008 reuse the old extension state and thus accept a session ticket if one was
7009 announced in the initial ServerHello.
7011 Similarly, ensure that the client requires a session ticket if one
7012 was advertised in the ServerHello. Previously, a TLS client would
7013 ignore a missing NewSessionTicket message.
7017 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7021 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7022 sends a carefully crafted handshake message, to cause OpenSSL to fail
7023 to free up to 64k of memory causing a memory leak. This could be
7024 exploited in a Denial Of Service attack. This issue affects OpenSSL
7025 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7026 whether SRTP is used or configured. Implementations of OpenSSL that
7027 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7029 The fix was developed by the OpenSSL team.
7034 * Session Ticket Memory Leak.
7036 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7037 integrity of that ticket is first verified. In the event of a session
7038 ticket integrity check failing, OpenSSL will fail to free memory
7039 causing a memory leak. By sending a large number of invalid session
7040 tickets an attacker could exploit this issue in a Denial Of Service
7046 * Build option no-ssl3 is incomplete.
7048 When OpenSSL is configured with "no-ssl3" as a build option, servers
7049 could accept and complete a SSL 3.0 handshake, and clients could be
7050 configured to send them.
7053 *Akamai and the OpenSSL team*
7055 * Add support for TLS_FALLBACK_SCSV.
7056 Client applications doing fallback retries should call
7057 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7060 *Adam Langley, Bodo Moeller*
7062 * Add additional DigestInfo checks.
7064 Re-encode DigestInto in DER and check against the original when
7065 verifying RSA signature: this will reject any improperly encoded
7066 DigestInfo structures.
7068 Note: this is a precautionary measure and no attacks are currently known.
7072 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7074 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7075 SRP code can be overrun an internal buffer. Add sanity check that
7076 g, A, B < N to SRP code.
7078 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7079 Group for discovering this issue.
7084 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7085 TLS 1.0 instead of higher protocol versions when the ClientHello message
7086 is badly fragmented. This allows a man-in-the-middle attacker to force a
7087 downgrade to TLS 1.0 even if both the server and the client support a
7088 higher protocol version, by modifying the client's TLS records.
7090 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7091 researching this issue.
7096 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7097 to a denial of service attack. A malicious server can crash the client
7098 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7099 ciphersuite and sending carefully crafted handshake messages.
7101 Thanks to Felix Gröbert (Google) for discovering and researching this
7107 * By sending carefully crafted DTLS packets an attacker could cause openssl
7108 to leak memory. This can be exploited through a Denial of Service attack.
7109 Thanks to Adam Langley for discovering and researching this issue.
7114 * An attacker can force openssl to consume large amounts of memory whilst
7115 processing DTLS handshake messages. This can be exploited through a
7116 Denial of Service attack.
7117 Thanks to Adam Langley for discovering and researching this issue.
7122 * An attacker can force an error condition which causes openssl to crash
7123 whilst processing DTLS packets due to memory being freed twice. This
7124 can be exploited through a Denial of Service attack.
7125 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7131 * If a multithreaded client connects to a malicious server using a resumed
7132 session and the server sends an ec point format extension it could write
7133 up to 255 bytes to freed memory.
7135 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7141 * A malicious server can crash an OpenSSL client with a null pointer
7142 dereference (read) by specifying an SRP ciphersuite even though it was not
7143 properly negotiated with the client. This can be exploited through a
7144 Denial of Service attack.
7146 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7147 discovering and researching this issue.
7152 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7153 X509_name_oneline, X509_name_print_ex et al. to leak some information
7154 from the stack. Applications may be affected if they echo pretty printing
7155 output to the attacker.
7157 Thanks to Ivan Fratric (Google) for discovering this issue.
7160 *Emilia Käsper, and Steve Henson*
7162 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7163 for corner cases. (Certain input points at infinity could lead to
7164 bogus results, with non-infinity inputs mapped to infinity too.)
7168 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7170 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7171 handshake can force the use of weak keying material in OpenSSL
7172 SSL/TLS clients and servers.
7174 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7175 researching this issue. ([CVE-2014-0224])
7177 *KIKUCHI Masashi, Steve Henson*
7179 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7180 OpenSSL DTLS client the code can be made to recurse eventually crashing
7183 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7186 *Imre Rad, Steve Henson*
7188 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7189 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7190 client or server. This is potentially exploitable to run arbitrary
7191 code on a vulnerable client or server.
7193 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7195 *Jüri Aedla, Steve Henson*
7197 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7198 are subject to a denial of service attack.
7200 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7201 this issue. ([CVE-2014-3470])
7203 *Felix Gröbert, Ivan Fratric, Steve Henson*
7205 * Harmonize version and its documentation. -f flag is used to display
7208 *mancha <mancha1@zoho.com>*
7210 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7211 in i2d_ECPrivateKey.
7213 *mancha <mancha1@zoho.com>*
7215 * Fix some double frees. These are not thought to be exploitable.
7217 *mancha <mancha1@zoho.com>*
7219 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7221 * A missing bounds check in the handling of the TLS heartbeat extension
7222 can be used to reveal up to 64k of memory to a connected client or
7225 Thanks for Neel Mehta of Google Security for discovering this bug and to
7226 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7227 preparing the fix ([CVE-2014-0160])
7229 *Adam Langley, Bodo Moeller*
7231 * Fix for the attack described in the paper "Recovering OpenSSL
7232 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7233 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7234 <http://eprint.iacr.org/2014/140>
7236 Thanks to Yuval Yarom and Naomi Benger for discovering this
7237 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7239 *Yuval Yarom and Naomi Benger*
7241 * TLS pad extension: draft-agl-tls-padding-03
7243 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7244 TLS client Hello record length value would otherwise be > 255 and
7245 less that 512 pad with a dummy extension containing zeroes so it
7246 is at least 512 bytes long.
7248 *Adam Langley, Steve Henson*
7250 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7252 * Fix for TLS record tampering bug. A carefully crafted invalid
7253 handshake could crash OpenSSL with a NULL pointer exception.
7254 Thanks to Anton Johansson for reporting this issues.
7257 * Keep original DTLS digest and encryption contexts in retransmission
7258 structures so we can use the previous session parameters if they need
7259 to be resent. ([CVE-2013-6450])
7263 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7264 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7265 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7266 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7267 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7268 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7270 *Rob Stradling, Adam Langley*
7272 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7274 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7275 supporting platforms or when small records were transferred.
7277 *Andy Polyakov, Steve Henson*
7279 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7281 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7283 This addresses the flaw in CBC record processing discovered by
7284 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7285 at: <http://www.isg.rhul.ac.uk/tls/>
7287 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7288 Security Group at Royal Holloway, University of London
7289 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7290 Emilia Käsper for the initial patch.
7293 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7295 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7296 ciphersuites which can be exploited in a denial of service attack.
7297 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7298 and detecting this bug and to Wolfgang Ettlinger
7299 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7304 * Return an error when checking OCSP signatures when key is NULL.
7305 This fixes a DoS attack. ([CVE-2013-0166])
7309 * Make openssl verify return errors.
7311 *Chris Palmer <palmer@google.com> and Ben Laurie*
7313 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7314 the right response is stapled. Also change SSL_get_certificate()
7315 so it returns the certificate actually sent.
7316 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7318 *Rob Stradling <rob.stradling@comodo.com>*
7320 * Fix possible deadlock when decoding public keys.
7324 * Don't use TLS 1.0 record version number in initial client hello
7329 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7331 * Sanity check record length before skipping explicit IV in TLS
7332 1.2, 1.1 and DTLS to fix DoS attack.
7334 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7335 fuzzing as a service testing platform.
7340 * Initialise tkeylen properly when encrypting CMS messages.
7341 Thanks to Solar Designer of Openwall for reporting this issue.
7345 * In FIPS mode don't try to use composite ciphers as they are not
7350 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7352 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7353 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7354 mean any application compiled against OpenSSL 1.0.0 headers setting
7355 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7356 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7357 0x10000000L Any application which was previously compiled against
7358 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7359 will need to be recompiled as a result. Letting be results in
7360 inability to disable specifically TLS 1.1 and in client context,
7361 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7365 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7366 disable just protocol X, but all protocols above X *if* there are
7367 protocols *below* X still enabled. In more practical terms it means
7368 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7369 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7370 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7375 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7377 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7378 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7379 in CRYPTO_realloc_clean.
7381 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7382 issue and to Adam Langley <agl@chromium.org> for fixing it.
7385 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7387 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7391 * Workarounds for some broken servers that "hang" if a client hello
7392 record length exceeds 255 bytes.
7394 1. Do not use record version number > TLS 1.0 in initial client
7395 hello: some (but not all) hanging servers will now work.
7396 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7397 the number of ciphers sent in the client hello. This should be
7398 set to an even number, such as 50, for example by passing:
7399 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7400 Most broken servers should now work.
7401 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7402 TLS 1.2 client support entirely.
7406 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7410 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7412 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7413 STRING form instead of a DigestInfo.
7417 * The format used for MDC2 RSA signatures is inconsistent between EVP
7418 and the RSA_sign/RSA_verify functions. This was made more apparent when
7419 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7420 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7421 the correct format in RSA_verify so both forms transparently work.
7425 * Some servers which support TLS 1.0 can choke if we initially indicate
7426 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7427 encrypted premaster secret. As a workaround use the maximum permitted
7428 client version in client hello, this should keep such servers happy
7429 and still work with previous versions of OpenSSL.
7433 * Add support for TLS/DTLS heartbeats.
7435 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7437 * Add support for SCTP.
7439 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7441 * Improved PRNG seeding for VOS.
7443 *Paul Green <Paul.Green@stratus.com>*
7445 * Extensive assembler packs updates, most notably:
7447 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7448 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7449 - x86_64: bit-sliced AES implementation;
7450 - ARM: NEON support, contemporary platforms optimizations;
7451 - s390x: z196 support;
7452 - `*`: GHASH and GF(2^m) multiplication implementations;
7456 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7457 (removal of unnecessary code)
7459 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7461 * Add TLS key material exporter from RFC 5705.
7465 * Add DTLS-SRTP negotiation from RFC 5764.
7469 * Add Next Protocol Negotiation,
7470 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7471 disabled with a no-npn flag to config or Configure. Code donated
7474 *Adam Langley <agl@google.com> and Ben Laurie*
7476 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7477 NIST-P256, NIST-P521, with constant-time single point multiplication on
7478 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7479 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7480 Code made available under Apache License version 2.0.
7482 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7483 line to include this in your build of OpenSSL, and run "make depend" (or
7484 "make update"). This enables the following EC_METHODs:
7486 EC_GFp_nistp224_method()
7487 EC_GFp_nistp256_method()
7488 EC_GFp_nistp521_method()
7490 EC_GROUP_new_by_curve_name() will automatically use these (while
7491 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7494 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7496 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7497 all platforms. Move ssize_t definition from e_os.h to the public
7498 header file e_os2.h as it now appears in public header file cms.h
7502 * New -sigopt option to the ca, req and x509 utilities. Additional
7503 signature parameters can be passed using this option and in
7508 * Add RSA PSS signing function. This will generate and set the
7509 appropriate AlgorithmIdentifiers for PSS based on those in the
7510 corresponding EVP_MD_CTX structure. No application support yet.
7514 * Support for companion algorithm specific ASN1 signing routines.
7515 New function ASN1_item_sign_ctx() signs a pre-initialised
7516 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7517 the appropriate parameters.
7521 * Add new algorithm specific ASN1 verification initialisation function
7522 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7523 handling will be the same no matter what EVP_PKEY_METHOD is used.
7524 Add a PSS handler to support verification of PSS signatures: checked
7525 against a number of sample certificates.
7529 * Add signature printing for PSS. Add PSS OIDs.
7531 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7533 * Add algorithm specific signature printing. An individual ASN1 method
7534 can now print out signatures instead of the standard hex dump.
7536 More complex signatures (e.g. PSS) can print out more meaningful
7537 information. Include DSA version that prints out the signature
7542 * Password based recipient info support for CMS library: implementing
7547 * Split password based encryption into PBES2 and PBKDF2 functions. This
7548 neatly separates the code into cipher and PBE sections and is required
7549 for some algorithms that split PBES2 into separate pieces (such as
7550 password based CMS).
7554 * Session-handling fixes:
7555 - Fix handling of connections that are resuming with a session ID,
7556 but also support Session Tickets.
7557 - Fix a bug that suppressed issuing of a new ticket if the client
7558 presented a ticket with an expired session.
7559 - Try to set the ticket lifetime hint to something reasonable.
7560 - Make tickets shorter by excluding irrelevant information.
7561 - On the client side, don't ignore renewed tickets.
7563 *Adam Langley, Bodo Moeller (Google)*
7565 * Fix PSK session representation.
7569 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7571 This work was sponsored by Intel.
7575 * Add GCM support to TLS library. Some custom code is needed to split
7576 the IV between the fixed (from PRF) and explicit (from TLS record)
7577 portions. This adds all GCM ciphersuites supported by RFC5288 and
7578 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7579 add a special AESGCM string for GCM only.
7583 * Expand range of ctrls for AES GCM. Permit setting invocation
7584 field on decrypt and retrieval of invocation field only on encrypt.
7588 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7589 As required by RFC5289 these ciphersuites cannot be used if for
7590 versions of TLS earlier than 1.2.
7594 * For FIPS capable OpenSSL interpret a NULL default public key method
7595 as unset and return the appropriate default but do *not* set the default.
7596 This means we can return the appropriate method in applications that
7597 switch between FIPS and non-FIPS modes.
7601 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7602 ENGINE is used then we cannot handle that in the FIPS module so we
7603 keep original code iff non-FIPS operations are allowed.
7607 * Add -attime option to openssl utilities.
7609 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7611 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7615 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7616 FIPS EC methods unconditionally for now.
7620 * New build option no-ec2m to disable characteristic 2 code.
7624 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7625 all cases can be covered as some introduce binary incompatibilities.
7629 * Redirect RSA operations to FIPS module including keygen,
7630 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7634 * Add similar low level API blocking to ciphers.
7638 * Low level digest APIs are not approved in FIPS mode: any attempt
7639 to use these will cause a fatal error. Applications that *really* want
7640 to use them can use the `private_*` version instead.
7644 * Redirect cipher operations to FIPS module for FIPS builds.
7648 * Redirect digest operations to FIPS module for FIPS builds.
7652 * Update build system to add "fips" flag which will link in fipscanister.o
7653 for static and shared library builds embedding a signature if needed.
7657 * Output TLS supported curves in preference order instead of numerical
7658 order. This is currently hardcoded for the highest order curves first.
7659 This should be configurable so applications can judge speed vs strength.
7663 * Add TLS v1.2 server support for client authentication.
7667 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7672 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7673 FIPS modules versions.
7677 * Add TLS v1.2 client side support for client authentication. Keep cache
7678 of handshake records longer as we don't know the hash algorithm to use
7679 until after the certificate request message is received.
7683 * Initial TLS v1.2 client support. Add a default signature algorithms
7684 extension including all the algorithms we support. Parse new signature
7685 format in client key exchange. Relax some ECC signing restrictions for
7686 TLS v1.2 as indicated in RFC5246.
7690 * Add server support for TLS v1.2 signature algorithms extension. Switch
7691 to new signature format when needed using client digest preference.
7692 All server ciphersuites should now work correctly in TLS v1.2. No client
7693 support yet and no support for client certificates.
7697 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7698 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7699 ciphersuites. At present only RSA key exchange ciphersuites work with
7700 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7701 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7702 and version checking.
7706 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7707 with this defined it will not be affected by any changes to ssl internal
7708 structures. Add several utility functions to allow openssl application
7709 to work with OPENSSL_NO_SSL_INTERN defined.
7713 * A long standing patch to add support for SRP from EdelWeb (Peter
7714 Sylvester and Christophe Renou) was integrated.
7715 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7716 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7719 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7723 * Permit abbreviated handshakes when renegotiating using the function
7724 SSL_renegotiate_abbreviated().
7726 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7728 * Add call to ENGINE_register_all_complete() to
7729 ENGINE_load_builtin_engines(), so some implementations get used
7730 automatically instead of needing explicit application support.
7734 * Add support for TLS key exporter as described in RFC5705.
7736 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7738 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7739 a few changes are required:
7741 Add SSL_OP_NO_TLSv1_1 flag.
7742 Add TLSv1_1 methods.
7743 Update version checking logic to handle version 1.1.
7744 Add explicit IV handling (ported from DTLS code).
7745 Add command line options to s_client/s_server.
7752 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7754 * X509_ATTRIBUTE memory leak
7756 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7757 memory. This structure is used by the PKCS#7 and CMS routines so any
7758 application which reads PKCS#7 or CMS data from untrusted sources is
7759 affected. SSL/TLS is not affected.
7761 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7767 * Race condition handling PSK identify hint
7769 If PSK identity hints are received by a multi-threaded client then
7770 the values are wrongly updated in the parent SSL_CTX structure. This can
7771 result in a race condition potentially leading to a double free of the
7777 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7779 * Malformed ECParameters causes infinite loop
7781 When processing an ECParameters structure OpenSSL enters an infinite loop
7782 if the curve specified is over a specially malformed binary polynomial
7785 This can be used to perform denial of service against any
7786 system which processes public keys, certificate requests or
7787 certificates. This includes TLS clients and TLS servers with
7788 client authentication enabled.
7790 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7795 * Exploitable out-of-bounds read in X509_cmp_time
7797 X509_cmp_time does not properly check the length of the ASN1_TIME
7798 string and can read a few bytes out of bounds. In addition,
7799 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7802 An attacker can use this to craft malformed certificates and CRLs of
7803 various sizes and potentially cause a segmentation fault, resulting in
7804 a DoS on applications that verify certificates or CRLs. TLS clients
7805 that verify CRLs are affected. TLS clients and servers with client
7806 authentication enabled may be affected if they use custom verification
7809 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7810 independently by Hanno Böck.
7815 * PKCS7 crash with missing EnvelopedContent
7817 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7818 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7819 with missing content and trigger a NULL pointer dereference on parsing.
7821 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7822 structures from untrusted sources are affected. OpenSSL clients and
7823 servers are not affected.
7825 This issue was reported to OpenSSL by Michal Zalewski (Google).
7830 * CMS verify infinite loop with unknown hash function
7832 When verifying a signedData message the CMS code can enter an infinite loop
7833 if presented with an unknown hash function OID. This can be used to perform
7834 denial of service against any system which verifies signedData messages using
7836 This issue was reported to OpenSSL by Johannes Bauer.
7841 * Race condition handling NewSessionTicket
7843 If a NewSessionTicket is received by a multi-threaded client when attempting to
7844 reuse a previous ticket then a race condition can occur potentially leading to
7845 a double free of the ticket data.
7850 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7852 * Segmentation fault in ASN1_TYPE_cmp fix
7854 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7855 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7856 certificate signature algorithm consistency this can be used to crash any
7857 certificate verification operation and exploited in a DoS attack. Any
7858 application which performs certificate verification is vulnerable including
7859 OpenSSL clients and servers which enable client authentication.
7864 * ASN.1 structure reuse memory corruption fix
7866 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7867 memory corruption via an invalid write. Such reuse is and has been
7868 strongly discouraged and is believed to be rare.
7870 Applications that parse structures containing CHOICE or ANY DEFINED BY
7871 components may be affected. Certificate parsing (d2i_X509 and related
7872 functions) are however not affected. OpenSSL clients and servers are
7878 * PKCS7 NULL pointer dereferences fix
7880 The PKCS#7 parsing code does not handle missing outer ContentInfo
7881 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7882 missing content and trigger a NULL pointer dereference on parsing.
7884 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7885 otherwise parse PKCS#7 structures from untrusted sources are
7886 affected. OpenSSL clients and servers are not affected.
7888 This issue was reported to OpenSSL by Michal Zalewski (Google).
7893 * DoS via reachable assert in SSLv2 servers fix
7895 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7896 servers that both support SSLv2 and enable export cipher suites by sending
7897 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7899 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7900 (OpenSSL development team).
7905 * Use After Free following d2i_ECPrivatekey error fix
7907 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7908 could cause a use after free condition. This, in turn, could cause a double
7909 free in several private key parsing functions (such as d2i_PrivateKey
7910 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7911 for applications that receive EC private keys from untrusted
7912 sources. This scenario is considered rare.
7914 This issue was discovered by the BoringSSL project and fixed in their
7920 * X509_to_X509_REQ NULL pointer deref fix
7922 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7923 the certificate key is invalid. This function is rarely used in practice.
7925 This issue was discovered by Brian Carpenter.
7930 * Removed the export ciphers from the DEFAULT ciphers
7934 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7936 * Build fixes for the Windows and OpenVMS platforms
7938 *Matt Caswell and Richard Levitte*
7940 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7942 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7943 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7944 dereference. This could lead to a Denial Of Service attack. Thanks to
7945 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7950 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7951 dtls1_buffer_record function under certain conditions. In particular this
7952 could occur if an attacker sent repeated DTLS records with the same
7953 sequence number but for the next epoch. The memory leak could be exploited
7954 by an attacker in a Denial of Service attack through memory exhaustion.
7955 Thanks to Chris Mueller for reporting this issue.
7960 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7961 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7962 method would be set to NULL which could later result in a NULL pointer
7963 dereference. Thanks to Frank Schmirler for reporting this issue.
7968 * Abort handshake if server key exchange message is omitted for ephemeral
7971 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7972 reporting this issue.
7977 * Remove non-export ephemeral RSA code on client and server. This code
7978 violated the TLS standard by allowing the use of temporary RSA keys in
7979 non-export ciphersuites and could be used by a server to effectively
7980 downgrade the RSA key length used to a value smaller than the server
7981 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7982 INRIA or reporting this issue.
7987 * Fixed issue where DH client certificates are accepted without verification.
7988 An OpenSSL server will accept a DH certificate for client authentication
7989 without the certificate verify message. This effectively allows a client to
7990 authenticate without the use of a private key. This only affects servers
7991 which trust a client certificate authority which issues certificates
7992 containing DH keys: these are extremely rare and hardly ever encountered.
7993 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7999 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8000 results on some platforms, including x86_64. This bug occurs at random
8001 with a very low probability, and is not known to be exploitable in any
8002 way, though its exact impact is difficult to determine. Thanks to Pieter
8003 Wuille (Blockstream) who reported this issue and also suggested an initial
8004 fix. Further analysis was conducted by the OpenSSL development team and
8005 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8006 the OpenSSL core team.
8011 * Fix various certificate fingerprint issues.
8013 By using non-DER or invalid encodings outside the signed portion of a
8014 certificate the fingerprint can be changed without breaking the signature.
8015 Although no details of the signed portion of the certificate can be changed
8016 this can cause problems with some applications: e.g. those using the
8017 certificate fingerprint for blacklists.
8019 1. Reject signatures with non zero unused bits.
8021 If the BIT STRING containing the signature has non zero unused bits reject
8022 the signature. All current signature algorithms require zero unused bits.
8024 2. Check certificate algorithm consistency.
8026 Check the AlgorithmIdentifier inside TBS matches the one in the
8027 certificate signature. NB: this will result in signature failure
8028 errors for some broken certificates.
8030 Thanks to Konrad Kraszewski from Google for reporting this issue.
8032 3. Check DSA/ECDSA signatures use DER.
8034 Reencode DSA/ECDSA signatures and compare with the original received
8035 signature. Return an error if there is a mismatch.
8037 This will reject various cases including garbage after signature
8038 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8039 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8040 (negative or with leading zeroes).
8042 Further analysis was conducted and fixes were developed by Stephen Henson
8043 of the OpenSSL core team.
8049 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8051 * Session Ticket Memory Leak.
8053 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8054 integrity of that ticket is first verified. In the event of a session
8055 ticket integrity check failing, OpenSSL will fail to free memory
8056 causing a memory leak. By sending a large number of invalid session
8057 tickets an attacker could exploit this issue in a Denial Of Service
8063 * Build option no-ssl3 is incomplete.
8065 When OpenSSL is configured with "no-ssl3" as a build option, servers
8066 could accept and complete a SSL 3.0 handshake, and clients could be
8067 configured to send them.
8070 *Akamai and the OpenSSL team*
8072 * Add support for TLS_FALLBACK_SCSV.
8073 Client applications doing fallback retries should call
8074 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8077 *Adam Langley, Bodo Moeller*
8079 * Add additional DigestInfo checks.
8081 Reencode DigestInto in DER and check against the original when
8082 verifying RSA signature: this will reject any improperly encoded
8083 DigestInfo structures.
8085 Note: this is a precautionary measure and no attacks are currently known.
8089 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8091 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8092 to a denial of service attack. A malicious server can crash the client
8093 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8094 ciphersuite and sending carefully crafted handshake messages.
8096 Thanks to Felix Gröbert (Google) for discovering and researching this
8102 * By sending carefully crafted DTLS packets an attacker could cause openssl
8103 to leak memory. This can be exploited through a Denial of Service attack.
8104 Thanks to Adam Langley for discovering and researching this issue.
8109 * An attacker can force openssl to consume large amounts of memory whilst
8110 processing DTLS handshake messages. This can be exploited through a
8111 Denial of Service attack.
8112 Thanks to Adam Langley for discovering and researching this issue.
8117 * An attacker can force an error condition which causes openssl to crash
8118 whilst processing DTLS packets due to memory being freed twice. This
8119 can be exploited through a Denial of Service attack.
8120 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8126 * If a multithreaded client connects to a malicious server using a resumed
8127 session and the server sends an ec point format extension it could write
8128 up to 255 bytes to freed memory.
8130 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8136 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8137 X509_name_oneline, X509_name_print_ex et al. to leak some information
8138 from the stack. Applications may be affected if they echo pretty printing
8139 output to the attacker.
8141 Thanks to Ivan Fratric (Google) for discovering this issue.
8144 *Emilia Käsper, and Steve Henson*
8146 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8147 for corner cases. (Certain input points at infinity could lead to
8148 bogus results, with non-infinity inputs mapped to infinity too.)
8152 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8154 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8155 handshake can force the use of weak keying material in OpenSSL
8156 SSL/TLS clients and servers.
8158 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8159 researching this issue. ([CVE-2014-0224])
8161 *KIKUCHI Masashi, Steve Henson*
8163 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8164 OpenSSL DTLS client the code can be made to recurse eventually crashing
8167 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8170 *Imre Rad, Steve Henson*
8172 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8173 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8174 client or server. This is potentially exploitable to run arbitrary
8175 code on a vulnerable client or server.
8177 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8179 *Jüri Aedla, Steve Henson*
8181 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8182 are subject to a denial of service attack.
8184 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8185 this issue. ([CVE-2014-3470])
8187 *Felix Gröbert, Ivan Fratric, Steve Henson*
8189 * Harmonize version and its documentation. -f flag is used to display
8192 *mancha <mancha1@zoho.com>*
8194 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8195 in i2d_ECPrivateKey.
8197 *mancha <mancha1@zoho.com>*
8199 * Fix some double frees. These are not thought to be exploitable.
8201 *mancha <mancha1@zoho.com>*
8203 * Fix for the attack described in the paper "Recovering OpenSSL
8204 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8205 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8206 <http://eprint.iacr.org/2014/140>
8208 Thanks to Yuval Yarom and Naomi Benger for discovering this
8209 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8211 *Yuval Yarom and Naomi Benger*
8213 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8215 * Keep original DTLS digest and encryption contexts in retransmission
8216 structures so we can use the previous session parameters if they need
8217 to be resent. ([CVE-2013-6450])
8221 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8222 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8223 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8224 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8225 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8226 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8228 *Rob Stradling, Adam Langley*
8230 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8232 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8234 This addresses the flaw in CBC record processing discovered by
8235 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8236 at: <http://www.isg.rhul.ac.uk/tls/>
8238 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8239 Security Group at Royal Holloway, University of London
8240 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8241 Emilia Käsper for the initial patch.
8244 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8246 * Return an error when checking OCSP signatures when key is NULL.
8247 This fixes a DoS attack. ([CVE-2013-0166])
8251 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8252 the right response is stapled. Also change SSL_get_certificate()
8253 so it returns the certificate actually sent.
8254 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8255 (This is a backport)
8257 *Rob Stradling <rob.stradling@comodo.com>*
8259 * Fix possible deadlock when decoding public keys.
8263 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8265 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8268 * Sanity check record length before skipping explicit IV in DTLS
8271 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8272 fuzzing as a service testing platform.
8277 * Initialise tkeylen properly when encrypting CMS messages.
8278 Thanks to Solar Designer of Openwall for reporting this issue.
8282 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8284 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8285 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8286 in CRYPTO_realloc_clean.
8288 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8289 issue and to Adam Langley <agl@chromium.org> for fixing it.
8292 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8294 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8296 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8297 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8298 content decryption and always return the same error. Note: this attack
8299 needs on average 2^20 messages so it only affects automated senders. The
8300 old behaviour can be re-enabled in the CMS code by setting the
8301 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8302 an MMA defence is not necessary.
8303 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8304 this issue. ([CVE-2012-0884])
8308 * Fix CVE-2011-4619: make sure we really are receiving a
8309 client hello before rejecting multiple SGC restarts. Thanks to
8310 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8314 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8316 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8317 Thanks to Antonio Martin, Enterprise Secure Access Research and
8318 Development, Cisco Systems, Inc. for discovering this bug and
8319 preparing a fix. ([CVE-2012-0050])
8323 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8325 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8326 of the Vaudenay padding oracle attack on CBC mode encryption
8327 which enables an efficient plaintext recovery attack against
8328 the OpenSSL implementation of DTLS. Their attack exploits timing
8329 differences arising during decryption processing. A research
8330 paper describing this attack can be found at:
8331 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8332 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8333 Security Group at Royal Holloway, University of London
8334 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8335 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8336 for preparing the fix. ([CVE-2011-4108])
8338 *Robin Seggelmann, Michael Tuexen*
8340 * Clear bytes used for block padding of SSL 3.0 records.
8343 *Adam Langley (Google)*
8345 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8346 Kadianakis <desnacked@gmail.com> for discovering this issue and
8347 Adam Langley for preparing the fix. ([CVE-2011-4619])
8349 *Adam Langley (Google)*
8351 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8353 *Andrey Kulikov <amdeich@gmail.com>*
8355 * Prevent malformed RFC3779 data triggering an assertion failure.
8356 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8357 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8359 *Rob Austein <sra@hactrn.net>*
8361 * Improved PRNG seeding for VOS.
8363 *Paul Green <Paul.Green@stratus.com>*
8365 * Fix ssl_ciph.c set-up race.
8367 *Adam Langley (Google)*
8369 * Fix spurious failures in ecdsatest.c.
8371 *Emilia Käsper (Google)*
8373 * Fix the BIO_f_buffer() implementation (which was mixing different
8374 interpretations of the `..._len` fields).
8376 *Adam Langley (Google)*
8378 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8379 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8380 threads won't reuse the same blinding coefficients.
8382 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8383 lock to call BN_BLINDING_invert_ex, and avoids one use of
8384 BN_BLINDING_update for each BN_BLINDING structure (previously,
8385 the last update always remained unused).
8387 *Emilia Käsper (Google)*
8389 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8391 *Bob Buckholz (Google)*
8393 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8395 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8396 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8398 *Kaspar Brand <ossl@velox.ch>*
8400 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8401 for multi-threaded use of ECDH. ([CVE-2011-3210])
8403 *Adam Langley (Google)*
8405 * Fix x509_name_ex_d2i memory leak on bad inputs.
8409 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8410 signature public key algorithm by using OID xref utilities instead.
8411 Before this you could only use some ECC ciphersuites with SHA1 only.
8415 * Add protection against ECDSA timing attacks as mentioned in the paper
8416 by Billy Bob Brumley and Nicola Tuveri, see:
8417 <http://eprint.iacr.org/2011/232.pdf>
8419 *Billy Bob Brumley and Nicola Tuveri*
8421 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8423 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8425 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8427 * Fix bug in string printing code: if *any* escaping is enabled we must
8428 escape the escape character (backslash) or the resulting string is
8433 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8435 * Disable code workaround for ancient and obsolete Netscape browsers
8436 and servers: an attacker can use it in a ciphersuite downgrade attack.
8437 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8441 * Fixed J-PAKE implementation error, originally discovered by
8442 Sebastien Martini, further info and confirmation from Stefan
8443 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8447 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8449 * Fix extension code to avoid race conditions which can result in a buffer
8450 overrun vulnerability: resumed sessions must not be modified as they can
8451 be shared by multiple threads. CVE-2010-3864
8455 * Fix WIN32 build system to correctly link an ENGINE directory into
8460 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8462 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8465 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8467 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8469 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8470 context. The operation can be customised via the ctrl mechanism in
8471 case ENGINEs want to include additional functionality.
8475 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8479 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8480 output hashes compatible with older versions of OpenSSL.
8482 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8484 * Fix compression algorithm handling: if resuming a session use the
8485 compression algorithm of the resumed session instead of determining
8486 it from client hello again. Don't allow server to change algorithm.
8490 * Add load_crls() function to commands tidying load_certs() too. Add option
8491 to verify utility to allow additional CRLs to be included.
8495 * Update OCSP request code to permit adding custom headers to the request:
8496 some responders need this.
8500 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8503 *Julia Lawall <julia@diku.dk>*
8505 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8506 needlessly dereferenced structures, used obsolete functions and
8507 didn't handle all updated verify codes correctly.
8511 * Disable MD2 in the default configuration.
8515 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8516 indicate the initial BIO being pushed or popped. This makes it possible
8517 to determine whether the BIO is the one explicitly called or as a result
8518 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8519 it handles reference counts correctly and doesn't zero out the I/O bio
8520 when it is not being explicitly popped. WARNING: applications which
8521 included workarounds for the old buggy behaviour will need to be modified
8522 or they could free up already freed BIOs.
8526 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8527 renaming to all platforms (within the 0.9.8 branch, this was
8528 done conditionally on Netware platforms to avoid a name clash).
8530 *Guenter <lists@gknw.net>*
8532 * Add ECDHE and PSK support to DTLS.
8534 *Michael Tuexen <tuexen@fh-muenster.de>*
8536 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8541 * Add "missing" function EVP_MD_flags() (without this the only way to
8542 retrieve a digest flags is by accessing the structure directly. Update
8543 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8544 or cipher is registered as in the "from" argument. Print out all
8545 registered digests in the dgst usage message instead of manually
8546 attempting to work them out.
8550 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8551 this allows the use of compression and extensions. Change default cipher
8552 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8553 by default unless an application cipher string requests it.
8557 * Alter match criteria in PKCS12_parse(). It used to try to use local
8558 key ids to find matching certificates and keys but some PKCS#12 files
8559 don't follow the (somewhat unwritten) rules and this strategy fails.
8560 Now just gather all certificates together and the first private key
8561 then look for the first certificate that matches the key.
8565 * Support use of registered digest and cipher names for dgst and cipher
8566 commands instead of having to add each one as a special case. So now
8573 openssl dgst -sha256 foo
8575 and this works for ENGINE based algorithms too.
8579 * Update Gost ENGINE to support parameter files.
8581 *Victor B. Wagner <vitus@cryptocom.ru>*
8583 * Support GeneralizedTime in ca utility.
8585 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8587 * Enhance the hash format used for certificate directory links. The new
8588 form uses the canonical encoding (meaning equivalent names will work
8589 even if they aren't identical) and uses SHA1 instead of MD5. This form
8590 is incompatible with the older format and as a result c_rehash should
8591 be used to rebuild symbolic links.
8595 * Make PKCS#8 the default write format for private keys, replacing the
8596 traditional format. This form is standardised, more secure and doesn't
8597 include an implicit MD5 dependency.
8601 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8602 committed to OpenSSL should pass this lot as a minimum.
8606 * Add session ticket override functionality for use by EAP-FAST.
8608 *Jouni Malinen <j@w1.fi>*
8610 * Modify HMAC functions to return a value. Since these can be implemented
8611 in an ENGINE errors can occur.
8615 * Type-checked OBJ_bsearch_ex.
8619 * Type-checked OBJ_bsearch. Also some constification necessitated
8620 by type-checking. Still to come: TXT_DB, bsearch(?),
8621 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8626 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8627 seconds to a tm structure directly, instead of going through OS
8628 specific date routines. This avoids any issues with OS routines such
8629 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8630 and X509_time_adj_ex() to cover the extended range. The existing
8631 X509_time_adj() is still usable and will no longer have any date issues.
8635 * Delta CRL support. New use deltas option which will attempt to locate
8636 and search any appropriate delta CRLs available.
8638 This work was sponsored by Google.
8642 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8643 code and add additional score elements. Validate alternate CRL paths
8644 as part of the CRL checking and indicate a new error "CRL path validation
8645 error" in this case. Applications wanting additional details can use
8646 the verify callback and check the new "parent" field. If this is not
8647 NULL CRL path validation is taking place. Existing applications won't
8648 see this because it requires extended CRL support which is off by
8651 This work was sponsored by Google.
8655 * Support for freshest CRL extension.
8657 This work was sponsored by Google.
8661 * Initial indirect CRL support. Currently only supported in the CRLs
8662 passed directly and not via lookup. Process certificate issuer
8663 CRL entry extension and lookup CRL entries by bother issuer name
8664 and serial number. Check and process CRL issuer entry in IDP extension.
8666 This work was sponsored by Google.
8670 * Add support for distinct certificate and CRL paths. The CRL issuer
8671 certificate is validated separately in this case. Only enabled if
8672 an extended CRL support flag is set: this flag will enable additional
8673 CRL functionality in future.
8675 This work was sponsored by Google.
8679 * Add support for policy mappings extension.
8681 This work was sponsored by Google.
8685 * Fixes to pathlength constraint, self issued certificate handling,
8686 policy processing to align with RFC3280 and PKITS tests.
8688 This work was sponsored by Google.
8692 * Support for name constraints certificate extension. DN, email, DNS
8693 and URI types are currently supported.
8695 This work was sponsored by Google.
8699 * To cater for systems that provide a pointer-based thread ID rather
8700 than numeric, deprecate the current numeric thread ID mechanism and
8701 replace it with a structure and associated callback type. This
8702 mechanism allows a numeric "hash" to be extracted from a thread ID in
8703 either case, and on platforms where pointers are larger than 'long',
8704 mixing is done to help ensure the numeric 'hash' is usable even if it
8705 can't be guaranteed unique. The default mechanism is to use "&errno"
8706 as a pointer-based thread ID to distinguish between threads.
8708 Applications that want to provide their own thread IDs should now use
8709 CRYPTO_THREADID_set_callback() to register a callback that will call
8710 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8712 Note that ERR_remove_state() is now deprecated, because it is tied
8713 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8714 to free the current thread's error state should be replaced by
8715 ERR_remove_thread_state(NULL).
8717 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8718 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8719 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8720 application was previously providing a numeric thread callback that
8721 was inappropriate for distinguishing threads, then uniqueness might
8722 have been obtained with &errno that happened immediately in the
8723 intermediate development versions of OpenSSL; this is no longer the
8724 case, the numeric thread callback will now override the automatic use
8727 *Geoff Thorpe, with help from Bodo Moeller*
8729 * Initial support for different CRL issuing certificates. This covers a
8730 simple case where the self issued certificates in the chain exist and
8731 the real CRL issuer is higher in the existing chain.
8733 This work was sponsored by Google.
8737 * Removed effectively defunct crypto/store from the build.
8741 * Revamp of STACK to provide stronger type-checking. Still to come:
8742 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8743 ASN1_STRING, CONF_VALUE.
8747 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8748 RAM on SSL connections. This option can save about 34k per idle SSL.
8752 * Revamp of LHASH to provide stronger type-checking. Still to come:
8753 STACK, TXT_DB, bsearch, qsort.
8757 * Initial support for Cryptographic Message Syntax (aka CMS) based
8758 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8759 support for data, signedData, compressedData, digestedData and
8760 encryptedData, envelopedData types included. Scripts to check against
8761 RFC4134 examples draft and interop and consistency checks of many
8762 content types and variants.
8766 * Add options to enc utility to support use of zlib compression BIO.
8770 * Extend mk1mf to support importing of options and assembly language
8771 files from Configure script, currently only included in VC-WIN32.
8772 The assembly language rules can now optionally generate the source
8773 files from the associated perl scripts.
8777 * Implement remaining functionality needed to support GOST ciphersuites.
8778 Interop testing has been performed using CryptoPro implementations.
8780 *Victor B. Wagner <vitus@cryptocom.ru>*
8782 * s390x assembler pack.
8786 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8791 * Implement Opaque PRF Input TLS extension as specified in
8792 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8793 official specification yet and no extension type assignment by
8794 IANA exists, this extension (for now) will have to be explicitly
8795 enabled when building OpenSSL by providing the extension number
8796 to use. For example, specify an option
8798 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8800 to the "config" or "Configure" script to enable the extension,
8801 assuming extension number 0x9527 (which is a completely arbitrary
8802 and unofficial assignment based on the MD5 hash of the Internet
8803 Draft). Note that by doing so, you potentially lose
8804 interoperability with other TLS implementations since these might
8805 be using the same extension number for other purposes.
8807 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8808 opaque PRF input value to use in the handshake. This will create
8809 an internal copy of the length-'len' string at 'src', and will
8810 return non-zero for success.
8812 To get more control and flexibility, provide a callback function
8815 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8816 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8820 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8823 Callback function 'cb' will be called in handshakes, and is
8824 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8825 Argument 'arg' is for application purposes (the value as given to
8826 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8827 be provided to the callback function). The callback function
8828 has to return non-zero to report success: usually 1 to use opaque
8829 PRF input just if possible, or 2 to enforce use of the opaque PRF
8830 input. In the latter case, the library will abort the handshake
8831 if opaque PRF input is not successfully negotiated.
8833 Arguments 'peerinput' and 'len' given to the callback function
8834 will always be NULL and 0 in the case of a client. A server will
8835 see the client's opaque PRF input through these variables if
8836 available (NULL and 0 otherwise). Note that if the server
8837 provides an opaque PRF input, the length must be the same as the
8838 length of the client's opaque PRF input.
8840 Note that the callback function will only be called when creating
8841 a new session (session resumption can resume whatever was
8842 previously negotiated), and will not be called in SSL 2.0
8843 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8844 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8845 for applications that need to enforce opaque PRF input.
8849 * Update ssl code to support digests other than SHA1+MD5 for handshake
8852 *Victor B. Wagner <vitus@cryptocom.ru>*
8854 * Add RFC4507 support to OpenSSL. This includes the corrections in
8855 RFC4507bis. The encrypted ticket format is an encrypted encoded
8856 SSL_SESSION structure, that way new session features are automatically
8859 If a client application caches session in an SSL_SESSION structure
8860 support is transparent because tickets are now stored in the encoded
8863 The SSL_CTX structure automatically generates keys for ticket
8864 protection in servers so again support should be possible
8865 with no application modification.
8867 If a client or server wishes to disable RFC4507 support then the option
8868 SSL_OP_NO_TICKET can be set.
8870 Add a TLS extension debugging callback to allow the contents of any client
8871 or server extensions to be examined.
8873 This work was sponsored by Google.
8877 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8878 OpenSSL should now compile cleanly on gcc 4.2
8880 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8882 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8883 support including streaming MAC support: this is required for GOST
8884 ciphersuite support.
8886 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8888 * Add option -stream to use PKCS#7 streaming in smime utility. New
8889 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8890 to output in BER and PEM format.
8894 * Experimental support for use of HMAC via EVP_PKEY interface. This
8895 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8896 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8897 ENGINE support for HMAC keys which are unextractable. New -mac and
8898 -macopt options to dgst utility.
8902 * New option -sigopt to dgst utility. Update dgst to use
8903 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8904 alternative signing parameters such as X9.31 or PSS in the dgst
8909 * Change ssl_cipher_apply_rule(), the internal function that does
8910 the work each time a ciphersuite string requests enabling
8911 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8912 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8913 the order of disabled ciphersuites such that those ciphersuites
8914 that most recently went from enabled to disabled not only stay
8915 in order with respect to each other, but also have higher priority
8916 than other disabled ciphersuites the next time ciphersuites are
8919 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8920 the same ciphersuites as with "HIGH" alone, but in a specific
8921 order where the PSK ciphersuites come first (since they are the
8922 most recently disabled ciphersuites when "HIGH" is parsed).
8924 Also, change ssl_create_cipher_list() (using this new
8925 functionality) such that between otherwise identical
8926 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8931 * Change ssl_create_cipher_list() so that it automatically
8932 arranges the ciphersuites in reasonable order before starting
8933 to process the rule string. Thus, the definition for "DEFAULT"
8934 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8935 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8936 This makes it much easier to arrive at a reasonable default order
8937 in applications for which anonymous ciphers are OK (meaning
8938 that you can't actually use DEFAULT).
8940 *Bodo Moeller; suggested by Victor Duchovni*
8942 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8943 processing) into multiple integers instead of setting
8944 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8945 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8946 (These masks as well as the individual bit definitions are hidden
8947 away into the non-exported interface ssl/ssl_locl.h, so this
8948 change to the definition of the SSL_CIPHER structure shouldn't
8949 affect applications.) This give us more bits for each of these
8950 categories, so there is no longer a need to coagulate AES128 and
8951 AES256 into a single algorithm bit, and to coagulate Camellia128
8952 and Camellia256 into a single algorithm bit, which has led to all
8955 Thus, among other things, the kludge introduced in 0.9.7m and
8956 0.9.8e for masking out AES256 independently of AES128 or masking
8957 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8959 With the change, we also introduce new ciphersuite aliases that
8960 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8965 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8966 Use the leftmost N bytes of the signature input if the input is
8967 larger than the prime q (with N being the size in bytes of q).
8971 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8972 it yet and it is largely untested.
8976 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8980 * Initial incomplete changes to avoid need for function casts in OpenSSL
8981 some compilers (gcc 4.2 and later) reject their use. Safestack is
8982 reimplemented. Update ASN1 to avoid use of legacy functions.
8986 * Win32/64 targets are linked with Winsock2.
8990 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8991 to external functions. This can be used to increase CRL handling
8992 efficiency especially when CRLs are very large by (for example) storing
8993 the CRL revoked certificates in a database.
8997 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8998 new CRLs added to a directory can be used. New command line option
8999 -verify_return_error to s_client and s_server. This causes real errors
9000 to be returned by the verify callback instead of carrying on no matter
9001 what. This reflects the way a "real world" verify callback would behave.
9005 * GOST engine, supporting several GOST algorithms and public key formats.
9006 Kindly donated by Cryptocom.
9010 * Partial support for Issuing Distribution Point CRL extension. CRLs
9011 partitioned by DP are handled but no indirect CRL or reason partitioning
9012 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9013 selected via a scoring technique which handles IDP and AKID in CRLs.
9017 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9018 will ultimately be used for all verify operations: this will remove the
9019 X509_STORE dependency on certificate verification and allow alternative
9020 lookup methods. X509_STORE based implementations of these two callbacks.
9024 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9025 Modify get_crl() to find a valid (unexpired) CRL if possible.
9029 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9030 this would be called X509_CRL_cmp() but that name is already used by
9031 a function that just compares CRL issuer names. Cache several CRL
9032 extensions in X509_CRL structure and cache CRLDP in X509.
9036 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9037 this maps equivalent X509_NAME structures into a consistent structure.
9038 Name comparison can then be performed rapidly using memcmp().
9042 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9047 * Allow digests to supply their own micalg string for S/MIME type using
9048 the ctrl EVP_MD_CTRL_MICALG.
9052 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9053 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9054 ctrl. It can then customise the structure before and/or after signing
9059 * New function OBJ_add_sigid() to allow application defined signature OIDs
9060 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9061 to free up any added signature OIDs.
9065 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9066 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9067 digest and cipher tables. New options added to openssl utility:
9068 list-message-digest-algorithms and list-cipher-algorithms.
9072 * Change the array representation of binary polynomials: the list
9073 of degrees of non-zero coefficients is now terminated with -1.
9074 Previously it was terminated with 0, which was also part of the
9075 value; thus, the array representation was not applicable to
9076 polynomials where t^0 has coefficient zero. This change makes
9077 the array representation useful in a more general context.
9081 * Various modifications and fixes to SSL/TLS cipher string
9082 handling. For ECC, the code now distinguishes between fixed ECDH
9083 with RSA certificates on the one hand and with ECDSA certificates
9084 on the other hand, since these are separate ciphersuites. The
9085 unused code for Fortezza ciphersuites has been removed.
9087 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9088 (not "ECDHE"). For consistency with the code for DH
9089 certificates, use of ECDH certificates is now considered ECDH
9090 authentication, not RSA or ECDSA authentication (the latter is
9091 merely the CA's signing algorithm and not actively used in the
9094 The temporary ciphersuite alias "ECCdraft" is no longer
9095 available, and ECC ciphersuites are no longer excluded from "ALL"
9096 and "DEFAULT". The following aliases now exist for RFC 4492
9097 ciphersuites, most of these by analogy with the DH case:
9099 kECDHr - ECDH cert, signed with RSA
9100 kECDHe - ECDH cert, signed with ECDSA
9101 kECDH - ECDH cert (signed with either RSA or ECDSA)
9102 kEECDH - ephemeral ECDH
9103 ECDH - ECDH cert or ephemeral ECDH
9109 AECDH - anonymous ECDH
9110 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9114 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9115 Use correct micalg parameters depending on digest(s) in signed message.
9119 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9120 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9124 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9125 an engine to register a method. Add ENGINE lookups for methods and
9126 functional reference processing.
9130 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9131 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9136 * New -resign option to smime utility. This adds one or more signers
9137 to an existing PKCS#7 signedData structure. Also -md option to use an
9138 alternative message digest algorithm for signing.
9142 * Tidy up PKCS#7 routines and add new functions to make it easier to
9143 create PKCS7 structures containing multiple signers. Update smime
9144 application to support multiple signers.
9148 * New -macalg option to pkcs12 utility to allow setting of an alternative
9153 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9154 Reorganize PBE internals to lookup from a static table using NIDs,
9155 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9156 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9157 PRF which will be automatically used with PBES2.
9161 * Replace the algorithm specific calls to generate keys in "req" with the
9166 * Update PKCS#7 enveloped data routines to use new API. This is now
9167 supported by any public key method supporting the encrypt operation. A
9168 ctrl is added to allow the public key algorithm to examine or modify
9169 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9174 * Add a ctrl to asn1 method to allow a public key algorithm to express
9175 a default digest type to use. In most cases this will be SHA1 but some
9176 algorithms (such as GOST) need to specify an alternative digest. The
9177 return value indicates how strong the preference is 1 means optional and
9178 2 is mandatory (that is it is the only supported type). Modify
9179 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9180 use the default md. Update openssl utilities to use the default digest
9181 type for signing if it is not explicitly indicated.
9185 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9186 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9187 signing method from the key type. This effectively removes the link
9188 between digests and public key types.
9192 * Add an OID cross reference table and utility functions. Its purpose is to
9193 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9194 rsaEncryption. This will allow some of the algorithm specific hackery
9195 needed to use the correct OID to be removed.
9199 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9200 structures for PKCS7_sign(). They are now set up by the relevant public
9205 * Add provisional EC pkey method with support for ECDSA and ECDH.
9209 * Add support for key derivation (agreement) in the API, DH method and
9214 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9215 public and private key formats. As a side effect these add additional
9216 command line functionality not previously available: DSA signatures can be
9217 generated and verified using pkeyutl and DH key support and generation in
9224 *Oliver Tappe <zooey@hirschkaefer.de>*
9226 * New make target "install_html_docs" installs HTML renditions of the
9229 *Oliver Tappe <zooey@hirschkaefer.de>*
9231 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9232 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9233 support key and parameter generation and add initial key generation
9234 functionality for RSA.
9238 * Add functions for main EVP_PKEY_method operations. The undocumented
9239 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9240 `EVP_PKEY_{encrypt,decrypt}_old`.
9244 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9245 key API, doesn't do much yet.
9249 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9250 public key algorithms. New option to openssl utility:
9251 "list-public-key-algorithms" to print out info.
9255 * Implement the Supported Elliptic Curves Extension for
9256 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9260 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9261 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9265 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9266 utilities such as rsa, dsa, dsaparam etc except they process any key
9271 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9272 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9273 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9278 * Initial support for pluggable public key ASN1.
9279 De-spaghettify the public key ASN1 handling. Move public and private
9280 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9281 algorithm specific handling to a single module within the relevant
9282 algorithm directory. Add functions to allow (near) opaque processing
9283 of public and private key structures.
9287 * Implement the Supported Point Formats Extension for
9288 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9292 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9293 for the psk identity [hint] and the psk callback functions to the
9294 SSL_SESSION, SSL and SSL_CTX structure.
9297 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9301 SSL_CTX_use_psk_identity_hint
9302 SSL_get_psk_identity_hint
9303 SSL_get_psk_identity
9304 SSL_use_psk_identity_hint
9306 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9308 * Add RFC 3161 compliant time stamp request creation, response generation
9309 and response verification functionality.
9311 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9313 * Add initial support for TLS extensions, specifically for the server_name
9314 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9315 have new members for a host name. The SSL data structure has an
9316 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9317 stored in that context to allow for session resumption, even after the
9318 SSL has been switched to a new SSL_CTX in reaction to a client's
9319 server_name extension.
9321 New functions (subject to change):
9323 SSL_get_servername()
9324 SSL_get_servername_type()
9327 New CTRL codes and macros (subject to change):
9329 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9330 - SSL_CTX_set_tlsext_servername_callback()
9331 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9332 - SSL_CTX_set_tlsext_servername_arg()
9333 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9335 openssl s_client has a new '-servername ...' option.
9337 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9338 '-key2 ...', '-servername_fatal' (subject to change). This allows
9339 testing the HostName extension for a specific single host name ('-cert'
9340 and '-key' remain fallbacks for handshakes without HostName
9341 negotiation). If the unrecognized_name alert has to be sent, this by
9342 default is a warning; it becomes fatal with the '-servername_fatal'
9345 *Peter Sylvester, Remy Allais, Christophe Renou*
9347 * Whirlpool hash implementation is added.
9351 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9352 bn(64,32). Because of instruction set limitations it doesn't have
9353 any negative impact on performance. This was done mostly in order
9354 to make it possible to share assembler modules, such as bn_mul_mont
9355 implementations, between 32- and 64-bit builds without hassle.
9359 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9360 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9365 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9366 dedicated Montgomery multiplication procedure, is introduced.
9367 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9368 "64-bit" performance on certain 32-bit targets.
9372 * New option SSL_OP_NO_COMP to disable use of compression selectively
9373 in SSL structures. New SSL ctrl to set maximum send fragment size.
9374 Save memory by setting the I/O buffer sizes dynamically instead of
9375 using the maximum available value.
9379 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9380 in addition to the text details.
9384 * Very, very preliminary EXPERIMENTAL support for printing of general
9385 ASN1 structures. This currently produces rather ugly output and doesn't
9386 handle several customised structures at all.
9390 * Integrated support for PVK file format and some related formats such
9391 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9392 these in the 'rsa' and 'dsa' utilities.
9396 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9400 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9401 place for the (very old) "NETSCAPE" format certificates which are now
9402 handled using new ASN1 code equivalents.
9406 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9407 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9408 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9412 * Modify CRL distribution points extension code to print out previously
9413 unsupported fields. Enhance extension setting code to allow setting of
9418 * Add print and set support for Issuing Distribution Point CRL extension.
9422 * Change 'Configure' script to enable Camellia by default.
9429 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9431 * When rejecting SSL/TLS records due to an incorrect version number, never
9432 update s->server with a new major version number. As of
9433 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9434 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9435 the previous behavior could result in a read attempt at NULL when
9436 receiving specific incorrect SSL/TLS records once record payload
9437 protection is active. ([CVE-2010-0740])
9439 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9441 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9442 could be crashed if the relevant tables were not present (e.g. chrooted).
9444 *Tomas Hoger <thoger@redhat.com>*
9446 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9448 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9450 *Martin Olsson, Neel Mehta*
9452 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9453 accommodate for stack sorting, always a write lock!).
9457 * On some versions of WIN32 Heap32Next is very slow. This can cause
9458 excessive delays in the RAND_poll(): over a minute. As a workaround
9459 include a time check in the inner Heap32Next loop too.
9463 * The code that handled flushing of data in SSL/TLS originally used the
9464 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9465 the problem outlined in PR#1949. The fix suggested there however can
9466 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9467 of Apache). So instead simplify the code to flush unconditionally.
9468 This should be fine since flushing with no data to flush is a no op.
9472 * Handle TLS versions 2.0 and later properly and correctly use the
9473 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9474 off ancient servers have a habit of sticking around for a while...
9478 * Modify compression code so it frees up structures without using the
9479 ex_data callbacks. This works around a problem where some applications
9480 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9481 restarting) then use compression (e.g. SSL with compression) later.
9482 This results in significant per-connection memory leaks and
9483 has caused some security issues including CVE-2008-1678 and
9488 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9489 change when encrypting or decrypting.
9493 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9494 connect and renegotiate with servers which do not support RI.
9495 Until RI is more widely deployed this option is enabled by default.
9499 * Add "missing" ssl ctrls to clear options and mode.
9503 * If client attempts to renegotiate and doesn't support RI respond with
9504 a no_renegotiation alert as required by RFC5746. Some renegotiating
9505 TLS clients will continue a connection gracefully when they receive
9506 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9507 waiting for a server hello which it will never receive. Now we treat a
9508 received no_renegotiation alert as a fatal error. This is because
9509 applications requesting a renegotiation might well expect it to succeed
9510 and would have no code in place to handle the server denying it so the
9511 only safe thing to do is to terminate the connection.
9515 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9516 peer supports secure renegotiation and 0 otherwise. Print out peer
9517 renegotiation support in s_client/s_server.
9521 * Replace the highly broken and deprecated SPKAC certification method with
9522 the updated NID creation version. This should correctly handle UTF8.
9526 * Implement RFC5746. Re-enable renegotiation but require the extension
9527 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9528 turns out to be a bad idea. It has been replaced by
9529 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9530 SSL_CTX_set_options(). This is really not recommended unless you
9531 know what you are doing.
9533 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9535 * Fixes to stateless session resumption handling. Use initial_ctx when
9536 issuing and attempting to decrypt tickets in case it has changed during
9537 servername handling. Use a non-zero length session ID when attempting
9538 stateless session resumption: this makes it possible to determine if
9539 a resumption has occurred immediately after receiving server hello
9540 (several places in OpenSSL subtly assume this) instead of later in
9545 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9546 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9547 fixes for a few places where the return code is not checked
9550 *Julia Lawall <julia@diku.dk>*
9552 * Add --strict-warnings option to Configure script to include devteam
9553 warnings in other configurations.
9557 * Add support for --libdir option and LIBDIR variable in makefiles. This
9558 makes it possible to install openssl libraries in locations which
9559 have names other than "lib", for example "/usr/lib64" which some
9562 *Steve Henson, based on patch from Jeremy Utley*
9564 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9565 X690 8.9.12 and can produce some misleading textual output of OIDs.
9567 *Steve Henson, reported by Dan Kaminsky*
9569 * Delete MD2 from algorithm tables. This follows the recommendation in
9570 several standards that it is not used in new applications due to
9571 several cryptographic weaknesses. For binary compatibility reasons
9572 the MD2 API is still compiled in by default.
9576 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9581 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9582 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9585 *Guenter <lists@gknw.net>*
9587 * Fix the server certificate chain building code to use X509_verify_cert(),
9588 it used to have an ad-hoc builder which was unable to cope with anything
9589 other than a simple chain.
9591 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9593 * Don't check self signed certificate signatures in X509_verify_cert()
9594 by default (a flag can override this): it just wastes time without
9595 adding any security. As a useful side effect self signed root CAs
9596 with non-FIPS digests are now usable in FIPS mode.
9600 * In dtls1_process_out_of_seq_message() the check if the current message
9601 is already buffered was missing. For every new message was memory
9602 allocated, allowing an attacker to perform an denial of service attack
9603 with sending out of seq handshake messages until there is no memory
9604 left. Additionally every future message was buffered, even if the
9605 sequence number made no sense and would be part of another handshake.
9606 So only messages with sequence numbers less than 10 in advance will be
9607 buffered. ([CVE-2009-1378])
9609 *Robin Seggelmann, discovered by Daniel Mentz*
9611 * Records are buffered if they arrive with a future epoch to be
9612 processed after finishing the corresponding handshake. There is
9613 currently no limitation to this buffer allowing an attacker to perform
9614 a DOS attack with sending records with future epochs until there is no
9615 memory left. This patch adds the pqueue_size() function to determine
9616 the size of a buffer and limits the record buffer to 100 entries.
9619 *Robin Seggelmann, discovered by Daniel Mentz*
9621 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9622 parent structure is freed. ([CVE-2009-1379])
9626 * Handle non-blocking I/O properly in SSL_shutdown() call.
9628 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9630 * Add `2.5.4.*` OIDs
9632 *Ilya O. <vrghost@gmail.com>*
9634 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9636 * Disable renegotiation completely - this fixes a severe security
9637 problem ([CVE-2009-3555]) at the cost of breaking all
9638 renegotiation. Renegotiation can be re-enabled by setting
9639 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9640 run-time. This is really not recommended unless you know what
9645 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9647 * Don't set val to NULL when freeing up structures, it is freed up by
9648 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9649 zeroing past the valid field. ([CVE-2009-0789])
9651 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9653 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9654 checked correctly. This would allow some invalid signed attributes to
9655 appear to verify correctly. ([CVE-2009-0591])
9657 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9659 * Reject UniversalString and BMPString types with invalid lengths. This
9660 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9661 a legal length. ([CVE-2009-0590])
9665 * Set S/MIME signing as the default purpose rather than setting it
9666 unconditionally. This allows applications to override it at the store
9671 * Permit restricted recursion of ASN1 strings. This is needed in practice
9672 to handle some structures.
9676 * Improve efficiency of mem_gets: don't search whole buffer each time
9679 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9681 * New -hex option for openssl rand.
9685 * Print out UTF8String and NumericString when parsing ASN1.
9689 * Support NumericString type for name components.
9693 * Allow CC in the environment to override the automatically chosen
9694 compiler. Note that nothing is done to ensure flags work with the
9699 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9701 * Properly check EVP_VerifyFinal() and similar return values
9704 *Ben Laurie, Bodo Moeller, Google Security Team*
9706 * Enable TLS extensions by default.
9710 * Allow the CHIL engine to be loaded, whether the application is
9711 multithreaded or not. (This does not release the developer from the
9712 obligation to set up the dynamic locking callbacks.)
9714 *Sander Temme <sander@temme.net>*
9716 * Use correct exit code if there is an error in dgst command.
9718 *Steve Henson; problem pointed out by Roland Dirlewanger*
9720 * Tweak Configure so that you need to say "experimental-jpake" to enable
9721 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9725 * Add experimental JPAKE support, including demo authentication in
9726 s_client and s_server.
9730 * Set the comparison function in v3_addr_canonize().
9732 *Rob Austein <sra@hactrn.net>*
9734 * Add support for XMPP STARTTLS in s_client.
9736 *Philip Paeps <philip@freebsd.org>*
9738 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9739 to ensure that even with this option, only ciphersuites in the
9740 server's preference list will be accepted. (Note that the option
9741 applies only when resuming a session, so the earlier behavior was
9742 just about the algorithm choice for symmetric cryptography.)
9746 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9748 * Fix NULL pointer dereference if a DTLS server received
9749 ChangeCipherSpec as first record ([CVE-2009-1386]).
9753 * Fix a state transition in s3_srvr.c and d1_srvr.c
9754 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9758 * The fix in 0.9.8c that supposedly got rid of unsafe
9759 double-checked locking was incomplete for RSA blinding,
9760 addressing just one layer of what turns out to have been
9761 doubly unsafe triple-checked locking.
9763 So now fix this for real by retiring the MONT_HELPER macro
9764 in crypto/rsa/rsa_eay.c.
9766 *Bodo Moeller; problem pointed out by Marius Schilder*
9768 * Various precautionary measures:
9770 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9772 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9773 (NB: This would require knowledge of the secret session ticket key
9774 to exploit, in which case you'd be SOL either way.)
9776 - Change bn_nist.c so that it will properly handle input BIGNUMs
9777 outside the expected range.
9779 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9782 *Neel Mehta, Bodo Moeller*
9784 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9785 the load fails. Useful for distros.
9787 *Ben Laurie and the FreeBSD team*
9789 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9793 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9797 * Expand ENGINE to support engine supplied SSL client certificate functions.
9799 This work was sponsored by Logica.
9803 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9804 keystores. Support for SSL/TLS client authentication too.
9805 Not compiled unless enable-capieng specified to Configure.
9807 This work was sponsored by Logica.
9811 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9812 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9813 attribute creation routines such as certificate requests and PKCS#12
9818 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9820 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9821 handshake which could lead to a client crash as found using the
9822 Codenomicon TLS test suite ([CVE-2008-1672])
9824 *Steve Henson, Mark Cox*
9826 * Fix double free in TLS server name extensions which could lead to
9827 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9831 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9833 Clear the error queue to ensure that error entries left from
9834 older function calls do not interfere with the correct operation.
9836 *Lutz Jaenicke, Erik de Castro Lopo*
9838 * Remove root CA certificates of commercial CAs:
9840 The OpenSSL project does not recommend any specific CA and does not
9841 have any policy with respect to including or excluding any CA.
9842 Therefore it does not make any sense to ship an arbitrary selection
9843 of root CA certificates with the OpenSSL software.
9847 * RSA OAEP patches to fix two separate invalid memory reads.
9848 The first one involves inputs when 'lzero' is greater than
9849 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9850 before the beginning of from). The second one involves inputs where
9851 the 'db' section contains nothing but zeroes (there is a one-byte
9852 invalid read after the end of 'db').
9854 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9856 * Partial backport from 0.9.9-dev:
9858 Introduce bn_mul_mont (dedicated Montgomery multiplication
9859 procedure) as a candidate for BIGNUM assembler implementation.
9860 While 0.9.9-dev uses assembler for various architectures, only
9861 x86_64 is available by default here in the 0.9.8 branch, and
9862 32-bit x86 is available through a compile-time setting.
9864 To try the 32-bit x86 assembler implementation, use Configure
9865 option "enable-montasm" (which exists only for this backport).
9867 As "enable-montasm" for 32-bit x86 disclaims code stability
9868 anyway, in this constellation we activate additional code
9869 backported from 0.9.9-dev for further performance improvements,
9870 namely BN_from_montgomery_word. (To enable this otherwise,
9871 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9873 *Andy Polyakov (backport partially by Bodo Moeller)*
9875 * Add TLS session ticket callback. This allows an application to set
9876 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9877 values. This is useful for key rollover for example where several key
9878 sets may exist with different names.
9882 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9883 This was broken until now in 0.9.8 releases, such that the only way
9884 a registered ENGINE could be used (assuming it initialises
9885 successfully on the host) was to explicitly set it as the default
9886 for the relevant algorithms. This is in contradiction with 0.9.7
9887 behaviour and the documentation. With this fix, when an ENGINE is
9888 registered into a given algorithm's table of implementations, the
9889 'uptodate' flag is reset so that auto-discovery will be used next
9890 time a new context for that algorithm attempts to select an
9893 *Ian Lister (tweaked by Geoff Thorpe)*
9895 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9896 implementation in the following ways:
9898 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9901 Lack of BER streaming support means one pass streaming processing is
9902 only supported if data is detached: setting the streaming flag is
9903 ignored for embedded content.
9905 CMS support is disabled by default and must be explicitly enabled
9906 with the enable-cms configuration option.
9910 * Update the GMP engine glue to do direct copies between BIGNUM and
9911 mpz_t when openssl and GMP use the same limb size. Otherwise the
9912 existing "conversion via a text string export" trick is still used.
9914 *Paul Sheer <paulsheer@gmail.com>*
9916 * Zlib compression BIO. This is a filter BIO which compressed and
9917 uncompresses any data passed through it.
9921 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9922 RFC3394 compatible AES key wrapping.
9926 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9927 sets string data without copying. X509_ALGOR_set0() and
9928 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9929 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9930 from an X509_ATTRIBUTE structure optionally checking it occurs only
9931 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9936 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9937 to get the expected BN_FLG_CONSTTIME behavior.
9939 *Bodo Moeller (Google)*
9943 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9944 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9945 - added some more tests to do_tests.pl
9946 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9947 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9948 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9949 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9950 - various changes to netware.pl to enable gcc-cross builds on Win32
9952 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9953 - various changes to fix missing prototype warnings
9954 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9955 - added AES, WHIRLPOOL and CPUID assembler code to build files
9956 - added missing AES assembler make rules to mk1mf.pl
9957 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9959 *Guenter Knauf <eflash@gmx.net>*
9961 * Implement certificate status request TLS extension defined in RFC3546.
9962 A client can set the appropriate parameters and receive the encoded
9963 OCSP response via a callback. A server can query the supplied parameters
9964 and set the encoded OCSP response in the callback. Add simplified examples
9965 to s_client and s_server.
9969 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9972 + Binary incompatibility of ssl_ctx_st structure
9973 + DTLS interoperation with non-compliant servers
9974 + Don't call get_session_cb() without proposed session
9975 + Fix ia64 assembler code
9977 *Andy Polyakov, Steve Henson*
9979 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9981 * DTLS Handshake overhaul. There were longstanding issues with
9982 OpenSSL DTLS implementation, which were making it impossible for
9983 RFC 4347 compliant client to communicate with OpenSSL server.
9984 Unfortunately just fixing these incompatibilities would "cut off"
9985 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9986 server keeps tolerating non RFC compliant syntax. The opposite is
9987 not true, 0.9.8f client can not communicate with earlier server.
9988 This update even addresses CVE-2007-4995.
9992 * Changes to avoid need for function casts in OpenSSL: some compilers
9993 (gcc 4.2 and later) reject their use.
9994 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9997 * Add RFC4507 support to OpenSSL. This includes the corrections in
9998 RFC4507bis. The encrypted ticket format is an encrypted encoded
9999 SSL_SESSION structure, that way new session features are automatically
10002 If a client application caches session in an SSL_SESSION structure
10003 support is transparent because tickets are now stored in the encoded
10006 The SSL_CTX structure automatically generates keys for ticket
10007 protection in servers so again support should be possible
10008 with no application modification.
10010 If a client or server wishes to disable RFC4507 support then the option
10011 SSL_OP_NO_TICKET can be set.
10013 Add a TLS extension debugging callback to allow the contents of any client
10014 or server extensions to be examined.
10016 This work was sponsored by Google.
10020 * Add initial support for TLS extensions, specifically for the server_name
10021 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10022 have new members for a host name. The SSL data structure has an
10023 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10024 stored in that context to allow for session resumption, even after the
10025 SSL has been switched to a new SSL_CTX in reaction to a client's
10026 server_name extension.
10028 New functions (subject to change):
10030 SSL_get_servername()
10031 SSL_get_servername_type()
10034 New CTRL codes and macros (subject to change):
10036 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10037 - SSL_CTX_set_tlsext_servername_callback()
10038 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10039 - SSL_CTX_set_tlsext_servername_arg()
10040 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10042 openssl s_client has a new '-servername ...' option.
10044 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10045 '-key2 ...', '-servername_fatal' (subject to change). This allows
10046 testing the HostName extension for a specific single host name ('-cert'
10047 and '-key' remain fallbacks for handshakes without HostName
10048 negotiation). If the unrecognized_name alert has to be sent, this by
10049 default is a warning; it becomes fatal with the '-servername_fatal'
10052 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10054 * Add AES and SSE2 assembly language support to VC++ build.
10058 * Mitigate attack on final subtraction in Montgomery reduction.
10062 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10063 (which previously caused an internal error).
10067 * Squeeze another 10% out of IGE mode when in != out.
10071 * AES IGE mode speedup.
10073 *Dean Gaudet (Google)*
10075 * Add the Korean symmetric 128-bit cipher SEED (see
10076 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10077 add SEED ciphersuites from RFC 4162:
10079 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10080 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10081 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10082 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10084 To minimize changes between patchlevels in the OpenSSL 0.9.8
10085 series, SEED remains excluded from compilation unless OpenSSL
10086 is configured with 'enable-seed'.
10088 *KISA, Bodo Moeller*
10090 * Mitigate branch prediction attacks, which can be practical if a
10091 single processor is shared, allowing a spy process to extract
10092 information. For detailed background information, see
10093 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10094 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10095 and Necessary Software Countermeasures"). The core of the change
10096 are new versions BN_div_no_branch() and
10097 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10098 respectively, which are slower, but avoid the security-relevant
10099 conditional branches. These are automatically called by BN_div()
10100 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10101 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10102 remove a conditional branch.
10104 BN_FLG_CONSTTIME is the new name for the previous
10105 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10106 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10107 in the exponent causes BN_mod_exp_mont() to use the alternative
10108 implementation in BN_mod_exp_mont_consttime().) The old name
10109 remains as a deprecated alias.
10111 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10112 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10113 constant-time implementations for more than just exponentiation.
10114 Here too the old name is kept as a deprecated alias.
10116 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10117 the BN_BLINDING structure gets an independent copy of the
10118 modulus. This means that the previous `BIGNUM *m` argument to
10119 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10120 essentially becomes `const BIGNUM *m`, although we can't actually
10121 change this in the header file before 0.9.9. It allows
10122 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10123 enable BN_FLG_CONSTTIME.
10125 *Matthew D Wood (Intel Corp)*
10127 * In the SSL/TLS server implementation, be strict about session ID
10128 context matching (which matters if an application uses a single
10129 external cache for different purposes). Previously,
10130 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10131 set. This did ensure strict client verification, but meant that,
10132 with applications using a single external cache for quite
10133 different requirements, clients could circumvent ciphersuite
10134 restrictions for a given session ID context by starting a session
10135 in a different context.
10139 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10140 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10141 authentication-only ciphersuites.
10145 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10146 not complete and could lead to a possible single byte overflow
10147 ([CVE-2007-5135]) [Ben Laurie]
10149 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10151 * Since AES128 and AES256 (and similarly Camellia128 and
10152 Camellia256) share a single mask bit in the logic of
10153 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10154 kludge to work properly if AES128 is available and AES256 isn't
10155 (or if Camellia128 is available and Camellia256 isn't).
10159 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10160 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10161 When a point or a seed is encoded in a BIT STRING, we need to
10162 prevent the removal of trailing zero bits to get the proper DER
10163 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10164 of a NamedBitList, for which trailing 0 bits need to be removed.)
10168 * Have SSL/TLS server implementation tolerate "mismatched" record
10169 protocol version while receiving ClientHello even if the
10170 ClientHello is fragmented. (The server can't insist on the
10171 particular protocol version it has chosen before the ServerHello
10172 message has informed the client about his choice.)
10176 * Add RFC 3779 support.
10178 *Rob Austein for ARIN, Ben Laurie*
10180 * Load error codes if they are not already present instead of using a
10181 static variable. This allows them to be cleanly unloaded and reloaded.
10182 Improve header file function name parsing.
10186 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10187 or CAPABILITY handshake as required by RFCs.
10189 *Goetz Babin-Ebell*
10191 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10193 * Introduce limits to prevent malicious keys being able to
10194 cause a denial of service. ([CVE-2006-2940])
10196 *Steve Henson, Bodo Moeller*
10198 * Fix ASN.1 parsing of certain invalid structures that can result
10199 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10201 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10202 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10204 * Fix SSL client code which could crash if connecting to a
10205 malicious SSLv2 server. ([CVE-2006-4343])
10207 *Tavis Ormandy and Will Drewry, Google Security Team*
10209 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10210 match only those. Before that, "AES256-SHA" would be interpreted
10211 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10212 the same strength classification in 0.9.7h) as we currently only
10213 have a single AES bit in the ciphersuite description bitmap.
10214 That change, however, also applied to ciphersuite strings such as
10215 "RC4-MD5" that intentionally matched multiple ciphersuites --
10216 namely, SSL 2.0 ciphersuites in addition to the more common ones
10217 from SSL 3.0/TLS 1.0.
10219 So we change the selection algorithm again: Naming an explicit
10220 ciphersuite selects this one ciphersuite, and any other similar
10221 ciphersuite (same bitmap) from *other* protocol versions.
10222 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10223 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10225 Since SSL 2.0 does not have any ciphersuites for which the
10226 128/256 bit distinction would be relevant, this works for now.
10227 The proper fix will be to use different bits for AES128 and
10228 AES256, which would have avoided the problems from the beginning;
10229 however, bits are scarce, so we can only do this in a new release
10230 (not just a patchlevel) when we can change the SSL_CIPHER
10231 definition to split the single 'unsigned long mask' bitmap into
10232 multiple values to extend the available space.
10236 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10238 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10239 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10241 * Add AES IGE and biIGE modes.
10245 * Change the Unix randomness entropy gathering to use poll() when
10246 possible instead of select(), since the latter has some
10247 undesirable limitations.
10249 *Darryl Miles via Richard Levitte and Bodo Moeller*
10251 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10252 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10253 cannot be implicitly activated as part of, e.g., the "AES" alias.
10254 However, please upgrade to OpenSSL 0.9.9[-dev] for
10255 non-experimental use of the ECC ciphersuites to get TLS extension
10256 support, which is required for curve and point format negotiation
10257 to avoid potential handshake problems.
10261 * Disable rogue ciphersuites:
10263 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10264 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10265 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10267 The latter two were purportedly from
10268 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10271 Also deactivate the remaining ciphersuites from
10272 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10273 unofficial, and the ID has long expired.
10277 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10278 dual-core machines) and other potential thread-safety issues.
10282 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10283 versions), which is now available for royalty-free use
10284 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10285 Also, add Camellia TLS ciphersuites from RFC 4132.
10287 To minimize changes between patchlevels in the OpenSSL 0.9.8
10288 series, Camellia remains excluded from compilation unless OpenSSL
10289 is configured with 'enable-camellia'.
10293 * Disable the padding bug check when compression is in use. The padding
10294 bug check assumes the first packet is of even length, this is not
10295 necessarily true if compression is enabled and can result in false
10296 positives causing handshake failure. The actual bug test is ancient
10297 code so it is hoped that implementations will either have fixed it by
10298 now or any which still have the bug do not support compression.
10302 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10304 * When applying a cipher rule check to see if string match is an explicit
10305 cipher suite and only match that one cipher suite if it is.
10309 * Link in manifests for VC++ if needed.
10311 *Austin Ziegler <halostatue@gmail.com>*
10313 * Update support for ECC-based TLS ciphersuites according to
10314 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10315 TLS extensions, which are supported starting with the 0.9.9
10316 branch, not in the OpenSSL 0.9.8 branch).
10320 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10321 opaque EVP_CIPHER_CTX handling.
10325 * Fixes and enhancements to zlib compression code. We now only use
10326 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10327 to conform with the standards mentioned here:
10328 <http://www.zlib.net/DLL_FAQ.txt>
10329 Static zlib linking now works on Windows and the new --with-zlib-include
10330 --with-zlib-lib options to Configure can be used to supply the location
10331 of the headers and library. Gracefully handle case where zlib library
10336 * Several fixes and enhancements to the OID generation code. The old code
10337 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10338 handle numbers larger than ULONG_MAX, truncated printing and had a
10339 non standard OBJ_obj2txt() behaviour.
10343 * Add support for building of engines under engine/ as shared libraries
10344 under VC++ build system.
10348 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10349 Hopefully, we will not see any false combination of paths any more.
10353 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10355 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10356 (part of SSL_OP_ALL). This option used to disable the
10357 countermeasure against man-in-the-middle protocol-version
10358 rollback in the SSL 2.0 server implementation, which is a bad
10359 idea. ([CVE-2005-2969])
10361 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10362 for Information Security, National Institute of Advanced Industrial
10363 Science and Technology [AIST], Japan)*
10365 * Add two function to clear and return the verify parameter flags.
10369 * Keep cipherlists sorted in the source instead of sorting them at
10370 runtime, thus removing the need for a lock.
10374 * Avoid some small subgroup attacks in Diffie-Hellman.
10376 *Nick Mathewson and Ben Laurie*
10378 * Add functions for well-known primes.
10382 * Extended Windows CE support.
10384 *Satoshi Nakamura and Andy Polyakov*
10386 * Initialize SSL_METHOD structures at compile time instead of during
10387 runtime, thus removing the need for a lock.
10391 * Make PKCS7_decrypt() work even if no certificate is supplied by
10392 attempting to decrypt each encrypted key in turn. Add support to
10397 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10399 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10402 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10406 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10407 key into the same file any more.
10411 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10415 * Add -utf8 command line and config file option to 'ca'.
10417 *Stefan <stf@udoma.org*
10419 * Removed the macro des_crypt(), as it seems to conflict with some
10420 libraries. Use DES_crypt().
10424 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10425 involves renaming the source and generated shared-libs for
10426 both. The engines will accept the corrected or legacy ids
10427 ('ncipher' and '4758_cca' respectively) when binding. NB,
10428 this only applies when building 'shared'.
10430 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10432 * Add attribute functions to EVP_PKEY structure. Modify
10433 PKCS12_create() to recognize a CSP name attribute and
10434 use it. Make -CSP option work again in pkcs12 utility.
10438 * Add new functionality to the bn blinding code:
10439 - automatic re-creation of the BN_BLINDING parameters after
10440 a fixed number of uses (currently 32)
10441 - add new function for parameter creation
10442 - introduce flags to control the update behaviour of the
10443 BN_BLINDING parameters
10444 - hide BN_BLINDING structure
10445 Add a second BN_BLINDING slot to the RSA structure to improve
10446 performance when a single RSA object is shared among several
10451 * Add support for DTLS.
10453 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10455 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10456 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10460 * Remove buggy and incomplete DH cert support from
10461 ssl/ssl_rsa.c and ssl/s3_both.c
10465 * Use SHA-1 instead of MD5 as the default digest algorithm for
10466 the `apps/openssl` commands.
10470 * Compile clean with "-Wall -Wmissing-prototypes
10471 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10472 DEBUG_SAFESTACK must also be set.
10476 * Change ./Configure so that certain algorithms can be disabled by default.
10477 The new counterpiece to "no-xxx" is "enable-xxx".
10479 The patented RC5 and MDC2 algorithms will now be disabled unless
10480 "enable-rc5" and "enable-mdc2", respectively, are specified.
10482 (IDEA remains enabled despite being patented. This is because IDEA
10483 is frequently required for interoperability, and there is no license
10484 fee for non-commercial use. As before, "no-idea" can be used to
10485 avoid this algorithm.)
10489 * Add processing of proxy certificates (see RFC 3820). This work was
10490 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10491 EGEE (Enabling Grids for E-science in Europe).
10495 * RC4 performance overhaul on modern architectures/implementations, such
10496 as Intel P4, IA-64 and AMD64.
10500 * New utility extract-section.pl. This can be used specify an alternative
10501 section number in a pod file instead of having to treat each file as
10502 a separate case in Makefile. This can be done by adding two lines to the
10505 =for comment openssl_section:XXX
10507 The blank line is mandatory.
10511 * New arguments -certform, -keyform and -pass for s_client and s_server
10512 to allow alternative format key and certificate files and passphrase
10517 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10518 update associated structures and add various utility functions.
10520 Add new policy related verify parameters, include policy checking in
10521 standard verify code. Enhance 'smime' application with extra parameters
10522 to support policy checking and print out.
10526 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10527 Nehemiah processors. These extensions support AES encryption in hardware
10528 as well as RNG (though RNG support is currently disabled).
10530 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10532 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10536 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10538 *Andy Polyakov and a number of other people*
10540 * Improved PowerPC platform support. Most notably BIGNUM assembler
10541 implementation contributed by IBM.
10543 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10545 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10546 exponent rather than 'unsigned long'. There is a corresponding change to
10547 the new 'rsa_keygen' element of the RSA_METHOD structure.
10549 *Jelte Jansen, Geoff Thorpe*
10551 * Functionality for creating the initial serial number file is now
10552 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10554 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10555 number file to 1, which is bound to cause problems. To avoid
10556 the problems while respecting compatibility between different 0.9.7
10557 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10558 CA.pl for serial number initialization. With the new release 0.9.8,
10559 we can fix the problem directly in the 'ca' utility.)
10563 * Reduced header interdependencies by declaring more opaque objects in
10564 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10565 give fewer recursive includes, which could break lazy source code - so
10566 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10567 developers should define this symbol when building and using openssl to
10568 ensure they track the recommended behaviour, interfaces, [etc], but
10569 backwards-compatible behaviour prevails when this isn't defined.
10573 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10577 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10578 This will generate a random key of the appropriate length based on the
10579 cipher context. The EVP_CIPHER can provide its own random key generation
10580 routine to support keys of a specific form. This is used in the des and
10581 3des routines to generate a key of the correct parity. Update S/MIME
10582 code to use new functions and hence generate correct parity DES keys.
10583 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10584 valid (weak or incorrect parity).
10588 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10589 as looking them up. This is useful when the verified structure may contain
10590 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10591 present unless the new PKCS7_NO_CRL flag is asserted.
10595 * Extend ASN1 oid configuration module. It now additionally accepts the
10598 shortName = some long name, 1.2.3.4
10602 * Reimplemented the BN_CTX implementation. There is now no more static
10603 limitation on the number of variables it can handle nor the depth of the
10604 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10605 information can now expand as required, and rather than having a single
10606 static array of bignums, BN_CTX now uses a linked-list of such arrays
10607 allowing it to expand on demand whilst maintaining the usefulness of
10608 BN_CTX's "bundling".
10612 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10613 to allow all RSA operations to function using a single BN_CTX.
10617 * Preliminary support for certificate policy evaluation and checking. This
10618 is initially intended to pass the tests outlined in "Conformance Testing
10619 of Relying Party Client Certificate Path Processing Logic" v1.07.
10623 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10624 remained unused and not that useful. A variety of other little bignum
10625 tweaks and fixes have also been made continuing on from the audit (see
10630 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10631 associated ASN1, EVP and SSL functions and old ASN1 macros.
10635 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10636 and this should never fail. So the return value from the use of
10637 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10638 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10642 * BN_CTX_get() should return zero-valued bignums, providing the same
10643 initialised value as BN_new().
10645 *Geoff Thorpe, suggested by Ulf Möller*
10647 * Support for inhibitAnyPolicy certificate extension.
10651 * An audit of the BIGNUM code is underway, for which debugging code is
10652 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10653 is considered valid when processing BIGNUMs, and causes execution to
10654 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10655 further steps are taken to deliberately pollute unused data in BIGNUM
10656 structures to try and expose faulty code further on. For now, openssl will
10657 (in its default mode of operation) continue to tolerate the inconsistent
10658 forms that it has tolerated in the past, but authors and packagers should
10659 consider trying openssl and their own applications when compiled with
10660 these debugging symbols defined. It will help highlight potential bugs in
10661 their own code, and will improve the test coverage for OpenSSL itself. At
10662 some point, these tighter rules will become openssl's default to improve
10663 maintainability, though the assert()s and other overheads will remain only
10664 in debugging configurations. See bn.h for more details.
10666 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10668 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10669 that can only be obtained through BN_CTX_new() (which implicitly
10670 initialises it). The presence of this function only made it possible
10671 to overwrite an existing structure (and cause memory leaks).
10675 * Because of the callback-based approach for implementing LHASH as a
10676 template type, lh_insert() adds opaque objects to hash-tables and
10677 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10678 to clean up those corresponding objects before destroying the hash table
10679 (and losing the object pointers). So some over-zealous constifications in
10680 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10681 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10682 prototyped to have "const" restrictions on the object pointers they are
10683 given (and so aren't required to cast them away any more).
10687 * The tmdiff.h API was so ugly and minimal that our own timing utility
10688 (speed) prefers to use its own implementation. The two implementations
10689 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10690 its object type properly exposed (MS_TM) instead of casting to/from
10691 `char *`. This may still change yet if someone realises MS_TM and
10693 aren't necessarily the greatest nomenclatures - but this is what was used
10694 internally to the implementation so I've used that for now.
10698 * Ensure that deprecated functions do not get compiled when
10699 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10700 the self-tests were still using deprecated key-generation functions so
10701 these have been updated also.
10705 * Reorganise PKCS#7 code to separate the digest location functionality
10706 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10707 New function PKCS7_set_digest() to set the digest type for PKCS#7
10708 digestedData type. Add additional code to correctly generate the
10709 digestedData type and add support for this type in PKCS7 initialization
10714 * New function PKCS7_set0_type_other() this initializes a PKCS7
10715 structure of type "other".
10719 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10720 sure the loop does correctly stop and breaking ("division by zero")
10721 modulus operations are not performed. The (pre-generated) prime
10722 table crypto/bn/bn_prime.h was already correct, but it could not be
10723 re-generated on some platforms because of the "division by zero"
10724 situation in the script.
10726 *Ralf S. Engelschall*
10728 * Update support for ECC-based TLS ciphersuites according to
10729 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10730 SHA-1 now is only used for "small" curves (where the
10731 representation of a field element takes up to 24 bytes); for
10732 larger curves, the field element resulting from ECDH is directly
10733 used as premaster secret.
10735 *Douglas Stebila (Sun Microsystems Laboratories)*
10737 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10738 curve secp160r1 to the tests.
10740 *Douglas Stebila (Sun Microsystems Laboratories)*
10742 * Add the possibility to load symbols globally with DSO.
10744 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10746 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10747 control of the error stack.
10751 * Add support for STORE in ENGINE.
10755 * Add the STORE type. The intention is to provide a common interface
10756 to certificate and key stores, be they simple file-based stores, or
10757 HSM-type store, or LDAP stores, or...
10758 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10762 * Add a generic structure called OPENSSL_ITEM. This can be used to
10763 pass a list of arguments to any function as well as provide a way
10764 for a function to pass data back to the caller.
10768 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10769 works like BUF_strdup() but can be used to duplicate a portion of
10770 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10775 * Add the function sk_find_ex() which works like sk_find(), but will
10776 return an index to an element even if an exact match couldn't be
10777 found. The index is guaranteed to point at the element where the
10778 searched-for key would be inserted to preserve sorting order.
10782 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10783 takes an extra flags argument for optional functionality. Currently,
10784 the following flags are defined:
10786 OBJ_BSEARCH_VALUE_ON_NOMATCH
10787 This one gets OBJ_bsearch_ex() to return a pointer to the first
10788 element where the comparing function returns a negative or zero
10791 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10792 This one gets OBJ_bsearch_ex() to return a pointer to the first
10793 element where the comparing function returns zero. This is useful
10794 if there are more than one element where the comparing function
10799 * Make it possible to create self-signed certificates with 'openssl ca'
10800 in such a way that the self-signed certificate becomes part of the
10801 CA database and uses the same mechanisms for serial number generation
10802 as all other certificate signing. The new flag '-selfsign' enables
10803 this functionality. Adapt CA.sh and CA.pl.in.
10807 * Add functionality to check the public key of a certificate request
10808 against a given private. This is useful to check that a certificate
10809 request can be signed by that key (self-signing).
10813 * Make it possible to have multiple active certificates with the same
10814 subject in the CA index file. This is done only if the keyword
10815 'unique_subject' is set to 'no' in the main CA section (default
10816 if 'CA_default') of the configuration file. The value is saved
10817 with the database itself in a separate index attribute file,
10818 named like the index file with '.attr' appended to the name.
10822 * Generate multi-valued AVAs using '+' notation in config files for
10827 * Support for nameConstraints certificate extension.
10831 * Support for policyConstraints certificate extension.
10835 * Support for policyMappings certificate extension.
10839 * Make sure the default DSA_METHOD implementation only uses its
10840 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10841 and change its own handlers to be NULL so as to remove unnecessary
10842 indirection. This lets alternative implementations fallback to the
10843 default implementation more easily.
10847 * Support for directoryName in GeneralName related extensions
10852 * Make it possible to link applications using Makefile.shared.
10853 Make that possible even when linking against static libraries!
10857 * Support for single pass processing for S/MIME signing. This now
10858 means that S/MIME signing can be done from a pipe, in addition
10859 cleartext signing (multipart/signed type) is effectively streaming
10860 and the signed data does not need to be all held in memory.
10862 This is done with a new flag PKCS7_STREAM. When this flag is set
10863 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10864 is done after the data is output (and digests calculated) in
10865 SMIME_write_PKCS7().
10869 * Add full support for -rpath/-R, both in shared libraries and
10870 applications, at least on the platforms where it's known how
10875 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10876 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10877 will now compute a table of multiples of the generator that
10878 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10879 faster (notably in the case of a single point multiplication,
10880 scalar * generator).
10882 *Nils Larsch, Bodo Moeller*
10884 * IPv6 support for certificate extensions. The various extensions
10885 which use the IP:a.b.c.d can now take IPv6 addresses using the
10886 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10891 * Added an ENGINE that implements RSA by performing private key
10892 exponentiations with the GMP library. The conversions to and from
10893 GMP's mpz_t format aren't optimised nor are any montgomery forms
10894 cached, and on x86 it appears OpenSSL's own performance has caught up.
10895 However there are likely to be other architectures where GMP could
10896 provide a boost. This ENGINE is not built in by default, but it can be
10897 specified at Configure time and should be accompanied by the necessary
10898 linker additions, eg;
10899 ./config -DOPENSSL_USE_GMP -lgmp
10903 * "openssl engine" will not display ENGINE/DSO load failure errors when
10904 testing availability of engines with "-t" - the old behaviour is
10905 produced by increasing the feature's verbosity with "-tt".
10909 * ECDSA routines: under certain error conditions uninitialized BN objects
10910 could be freed. Solution: make sure initialization is performed early
10911 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10916 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10917 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10918 software implementations. For DSA and DH, parameter generation can
10919 also be overridden by providing the appropriate method callbacks.
10923 * Change the "progress" mechanism used in key-generation and
10924 primality testing to functions that take a new BN_GENCB pointer in
10925 place of callback/argument pairs. The new API functions have `_ex`
10926 postfixes and the older functions are reimplemented as wrappers for
10927 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10928 declarations of the old functions to help (graceful) attempts to
10929 migrate to the new functions. Also, the new key-generation API
10930 functions operate on a caller-supplied key-structure and return
10931 success/failure rather than returning a key or NULL - this is to
10932 help make "keygen" another member function of RSA_METHOD etc.
10934 Example for using the new callback interface:
10936 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10937 void *my_arg = ...;
10940 BN_GENCB_set(&my_cb, my_callback, my_arg);
10942 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10943 /* For the meaning of a, b in calls to my_callback(), see the
10944 * documentation of the function that calls the callback.
10945 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10946 * my_callback should return 1 if it wants BN_is_prime_ex()
10947 * to continue, or 0 to stop.
10952 * Change the ZLIB compression method to be stateful, and make it
10953 available to TLS with the number defined in
10954 draft-ietf-tls-compression-04.txt.
10958 * Add the ASN.1 structures and functions for CertificatePair, which
10959 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10961 CertificatePair ::= SEQUENCE {
10962 forward [0] Certificate OPTIONAL,
10963 reverse [1] Certificate OPTIONAL,
10964 -- at least one of the pair shall be present -- }
10966 Also implement the PEM functions to read and write certificate
10967 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10969 This needed to be defined, mostly for the sake of the LDAP
10970 attribute crossCertificatePair, but may prove useful elsewhere as
10975 * Make it possible to inhibit symlinking of shared libraries in
10976 Makefile.shared, for Cygwin's sake.
10980 * Extend the BIGNUM API by creating a function
10981 void BN_set_negative(BIGNUM *a, int neg);
10982 and a macro that behave like
10983 int BN_is_negative(const BIGNUM *a);
10985 to avoid the need to access 'a->neg' directly in applications.
10989 * Implement fast modular reduction for pseudo-Mersenne primes
10990 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10991 EC_GROUP_new_curve_GFp() will now automatically use this
10994 *Nils Larsch <nla@trustcenter.de>*
10996 * Add new lock type (CRYPTO_LOCK_BN).
11000 * Change the ENGINE framework to automatically load engines
11001 dynamically from specific directories unless they could be
11002 found to already be built in or loaded. Move all the
11003 current engines except for the cryptodev one to a new
11004 directory engines/.
11005 The engines in engines/ are built as shared libraries if
11006 the "shared" options was given to ./Configure or ./config.
11007 Otherwise, they are inserted in libcrypto.a.
11008 /usr/local/ssl/engines is the default directory for dynamic
11009 engines, but that can be overridden at configure time through
11010 the usual use of --prefix and/or --openssldir, and at run
11011 time with the environment variable OPENSSL_ENGINES.
11013 *Geoff Thorpe and Richard Levitte*
11015 * Add Makefile.shared, a helper makefile to build shared
11016 libraries. Adapt Makefile.org.
11020 * Add version info to Win32 DLLs.
11022 *Peter 'Luna' Runestig" <peter@runestig.com>*
11024 * Add new 'medium level' PKCS#12 API. Certificates and keys
11025 can be added using this API to created arbitrary PKCS#12
11026 files while avoiding the low level API.
11028 New options to PKCS12_create(), key or cert can be NULL and
11029 will then be omitted from the output file. The encryption
11030 algorithm NIDs can be set to -1 for no encryption, the mac
11031 iteration count can be set to 0 to omit the mac.
11033 Enhance pkcs12 utility by making the -nokeys and -nocerts
11034 options work when creating a PKCS#12 file. New option -nomac
11035 to omit the mac, NONE can be set for an encryption algorithm.
11036 New code is modified to use the enhanced PKCS12_create()
11037 instead of the low level API.
11041 * Extend ASN1 encoder to support indefinite length constructed
11042 encoding. This can output sequences tags and octet strings in
11043 this form. Modify pk7_asn1.c to support indefinite length
11044 encoding. This is experimental and needs additional code to
11045 be useful, such as an ASN1 bio and some enhanced streaming
11048 Extend template encode functionality so that tagging is passed
11049 down to the template encoder.
11053 * Let 'openssl req' fail if an argument to '-newkey' is not
11054 recognized instead of using RSA as a default.
11058 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11059 As these are not official, they are not included in "ALL";
11060 the "ECCdraft" ciphersuite group alias can be used to select them.
11062 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11064 * Add ECDH engine support.
11066 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11068 * Add ECDH in new directory crypto/ecdh/.
11070 *Douglas Stebila (Sun Microsystems Laboratories)*
11072 * Let BN_rand_range() abort with an error after 100 iterations
11073 without success (which indicates a broken PRNG).
11077 * Change BN_mod_sqrt() so that it verifies that the input value
11078 is really the square of the return value. (Previously,
11079 BN_mod_sqrt would show GIGO behaviour.)
11083 * Add named elliptic curves over binary fields from X9.62, SECG,
11084 and WAP/WTLS; add OIDs that were still missing.
11086 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11088 * Extend the EC library for elliptic curves over binary fields
11089 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11092 EC_GF2m_simple_method
11096 EC_GROUP_new_curve_GF2m
11097 EC_GROUP_set_curve_GF2m
11098 EC_GROUP_get_curve_GF2m
11099 EC_POINT_set_affine_coordinates_GF2m
11100 EC_POINT_get_affine_coordinates_GF2m
11101 EC_POINT_set_compressed_coordinates_GF2m
11103 Point compression for binary fields is disabled by default for
11104 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11107 As binary polynomials are represented as BIGNUMs, various members
11108 of the EC_GROUP and EC_POINT data structures can be shared
11109 between the implementations for prime fields and binary fields;
11110 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11111 are essentially identical to their `..._GFp` counterparts.
11112 (For simplicity, the `..._GFp` prefix has been dropped from
11113 various internal method names.)
11115 An internal 'field_div' method (similar to 'field_mul' and
11116 'field_sqr') has been added; this is used only for binary fields.
11118 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11120 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11121 through methods ('mul', 'precompute_mult').
11123 The generic implementations (now internally called 'ec_wNAF_mul'
11124 and 'ec_wNAF_precomputed_mult') remain the default if these
11125 methods are undefined.
11127 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11129 * New function EC_GROUP_get_degree, which is defined through
11130 EC_METHOD. For curves over prime fields, this returns the bit
11131 length of the modulus.
11133 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11135 * New functions EC_GROUP_dup, EC_POINT_dup.
11136 (These simply call ..._new and ..._copy).
11138 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11140 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11141 Polynomials are represented as BIGNUMs (where the sign bit is not
11142 used) in the following functions [macros]:
11145 BN_GF2m_sub [= BN_GF2m_add]
11146 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11147 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11148 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11150 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11151 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11152 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11153 BN_GF2m_cmp [= BN_ucmp]
11155 (Note that only the 'mod' functions are actually for fields GF(2^m).
11156 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11158 For some functions, an the irreducible polynomial defining a
11159 field can be given as an 'unsigned int[]' with strictly
11160 decreasing elements giving the indices of those bits that are set;
11161 i.e., p[] represents the polynomial
11162 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11164 p[0] > p[1] > ... > p[k] = 0.
11165 This applies to the following functions:
11168 BN_GF2m_mod_mul_arr
11169 BN_GF2m_mod_sqr_arr
11170 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11171 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11172 BN_GF2m_mod_exp_arr
11173 BN_GF2m_mod_sqrt_arr
11174 BN_GF2m_mod_solve_quad_arr
11178 Conversion can be performed by the following functions:
11183 bntest.c has additional tests for binary polynomial arithmetic.
11185 Two implementations for BN_GF2m_mod_div() are available.
11186 The default algorithm simply uses BN_GF2m_mod_inv() and
11187 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11188 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11189 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11191 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11193 * Add new error code 'ERR_R_DISABLED' that can be used when some
11194 functionality is disabled at compile-time.
11196 *Douglas Stebila <douglas.stebila@sun.com>*
11198 * Change default behaviour of 'openssl asn1parse' so that more
11199 information is visible when viewing, e.g., a certificate:
11201 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11202 mode the content of non-printable OCTET STRINGs is output in a
11203 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11204 avoid the appearance of a printable string.
11206 *Nils Larsch <nla@trustcenter.de>*
11208 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11210 EC_GROUP_set_asn1_flag()
11211 EC_GROUP_get_asn1_flag()
11212 EC_GROUP_set_point_conversion_form()
11213 EC_GROUP_get_point_conversion_form()
11214 These control ASN1 encoding details:
11215 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11216 has been set to OPENSSL_EC_NAMED_CURVE.
11217 - Points are encoded in uncompressed form by default; options for
11218 asn1_for are as for point2oct, namely
11219 POINT_CONVERSION_COMPRESSED
11220 POINT_CONVERSION_UNCOMPRESSED
11221 POINT_CONVERSION_HYBRID
11223 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11225 EC_GROUP_set_seed()
11226 EC_GROUP_get0_seed()
11227 EC_GROUP_get_seed_len()
11228 This is used only for ASN1 purposes (so far).
11230 *Nils Larsch <nla@trustcenter.de>*
11232 * Add 'field_type' member to EC_METHOD, which holds the NID
11233 of the appropriate field type OID. The new function
11234 EC_METHOD_get_field_type() returns this value.
11236 *Nils Larsch <nla@trustcenter.de>*
11239 EC_POINT_point2bn()
11240 EC_POINT_bn2point()
11241 EC_POINT_point2hex()
11242 EC_POINT_hex2point()
11243 providing useful interfaces to EC_POINT_point2oct() and
11244 EC_POINT_oct2point().
11246 *Nils Larsch <nla@trustcenter.de>*
11248 * Change internals of the EC library so that the functions
11249 EC_GROUP_set_generator()
11250 EC_GROUP_get_generator()
11251 EC_GROUP_get_order()
11252 EC_GROUP_get_cofactor()
11253 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11254 to methods, which would lead to unnecessary code duplication when
11255 adding different types of curves.
11257 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11259 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11260 arithmetic, and such that modified wNAFs are generated
11261 (which avoid length expansion in many cases).
11265 * Add a function EC_GROUP_check_discriminant() (defined via
11266 EC_METHOD) that verifies that the curve discriminant is non-zero.
11268 Add a function EC_GROUP_check() that makes some sanity tests
11269 on a EC_GROUP, its generator and order. This includes
11270 EC_GROUP_check_discriminant().
11272 *Nils Larsch <nla@trustcenter.de>*
11274 * Add ECDSA in new directory crypto/ecdsa/.
11276 Add applications 'openssl ecparam' and 'openssl ecdsa'
11277 (these are based on 'openssl dsaparam' and 'openssl dsa').
11279 ECDSA support is also included in various other files across the
11280 library. Most notably,
11281 - 'openssl req' now has a '-newkey ecdsa:file' option;
11282 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11283 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11284 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11285 them suitable for ECDSA where domain parameters must be
11286 extracted before the specific public key;
11287 - ECDSA engine support has been added.
11289 *Nils Larsch <nla@trustcenter.de>*
11291 * Include some named elliptic curves, and add OIDs from X9.62,
11292 SECG, and WAP/WTLS. Each curve can be obtained from the new
11294 EC_GROUP_new_by_curve_name(),
11295 and the list of available named curves can be obtained with
11296 EC_get_builtin_curves().
11297 Also add a 'curve_name' member to EC_GROUP objects, which can be
11299 EC_GROUP_set_curve_name()
11300 EC_GROUP_get_curve_name()
11302 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11304 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11305 was actually never needed) and in BN_mul(). The removal in BN_mul()
11306 required a small change in bn_mul_part_recursive() and the addition
11307 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11308 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11309 bn_sub_words() and bn_add_words() except they take arrays with
11314 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11316 * Cleanse PEM buffers before freeing them since they may contain
11319 *Benjamin Bennett <ben@psc.edu>*
11321 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11322 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11323 authentication-only ciphersuites.
11327 * Since AES128 and AES256 share a single mask bit in the logic of
11328 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11329 kludge to work properly if AES128 is available and AES256 isn't.
11333 * Expand security boundary to match 1.1.1 module.
11337 * Remove redundant features: hash file source, editing of test vectors
11338 modify fipsld to use external fips_premain.c signature.
11342 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11343 run algorithm test programs.
11347 * Make algorithm test programs more tolerant of whitespace.
11351 * Have SSL/TLS server implementation tolerate "mismatched" record
11352 protocol version while receiving ClientHello even if the
11353 ClientHello is fragmented. (The server can't insist on the
11354 particular protocol version it has chosen before the ServerHello
11355 message has informed the client about his choice.)
11359 * Load error codes if they are not already present instead of using a
11360 static variable. This allows them to be cleanly unloaded and reloaded.
11364 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11366 * Introduce limits to prevent malicious keys being able to
11367 cause a denial of service. ([CVE-2006-2940])
11369 *Steve Henson, Bodo Moeller*
11371 * Fix ASN.1 parsing of certain invalid structures that can result
11372 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11374 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11375 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11377 * Fix SSL client code which could crash if connecting to a
11378 malicious SSLv2 server. ([CVE-2006-4343])
11380 *Tavis Ormandy and Will Drewry, Google Security Team*
11382 * Change ciphersuite string processing so that an explicit
11383 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11384 will no longer include "AES128-SHA"), and any other similar
11385 ciphersuite (same bitmap) from *other* protocol versions (so that
11386 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11387 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11388 changes from 0.9.8b and 0.9.8d.
11392 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11394 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11395 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11397 * Change the Unix randomness entropy gathering to use poll() when
11398 possible instead of select(), since the latter has some
11399 undesirable limitations.
11401 *Darryl Miles via Richard Levitte and Bodo Moeller*
11403 * Disable rogue ciphersuites:
11405 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11406 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11407 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11409 The latter two were purportedly from
11410 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11413 Also deactivate the remaining ciphersuites from
11414 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11415 unofficial, and the ID has long expired.
11419 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11420 dual-core machines) and other potential thread-safety issues.
11424 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11426 * Adapt fipsld and the build system to link against the validated FIPS
11427 module in FIPS mode.
11431 * Fixes for VC++ 2005 build under Windows.
11435 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11436 from a Windows bash shell such as MSYS. It is autodetected from the
11437 "config" script when run from a VC++ environment. Modify standard VC++
11438 build to use fipscanister.o from the GNU make build.
11442 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11444 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11445 The value now differs depending on if you build for FIPS or not.
11446 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11447 safely run with a non-FIPSed libcrypto, as it may crash because of
11448 the difference induced by this change.
11452 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11454 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11455 (part of SSL_OP_ALL). This option used to disable the
11456 countermeasure against man-in-the-middle protocol-version
11457 rollback in the SSL 2.0 server implementation, which is a bad
11458 idea. ([CVE-2005-2969])
11460 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11461 for Information Security, National Institute of Advanced Industrial
11462 Science and Technology [AIST, Japan)]*
11464 * Minimal support for X9.31 signatures and PSS padding modes. This is
11465 mainly for FIPS compliance and not fully integrated at this stage.
11469 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11470 the exponentiation using a fixed-length exponent. (Otherwise,
11471 the information leaked through timing could expose the secret key
11472 after many signatures; cf. Bleichenbacher's attack on DSA with
11477 * Make a new fixed-window mod_exp implementation the default for
11478 RSA, DSA, and DH private-key operations so that the sequence of
11479 squares and multiplies and the memory access pattern are
11480 independent of the particular secret key. This will mitigate
11481 cache-timing and potential related attacks.
11483 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11484 and this is automatically used by BN_mod_exp_mont() if the new flag
11485 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11486 will use this BN flag for private exponents unless the flag
11487 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11488 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11490 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11492 * Change the client implementation for SSLv23_method() and
11493 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11494 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11495 (Previously, the SSL 2.0 backwards compatible Client Hello
11496 message format would be used even with SSL_OP_NO_SSLv2.)
11500 * Add support for smime-type MIME parameter in S/MIME messages which some
11505 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11506 a threadsafe manner. Modify rsa code to use new function and add calls
11507 to dsa and dh code (which had race conditions before).
11511 * Include the fixed error library code in the C error file definitions
11512 instead of fixing them up at runtime. This keeps the error code
11513 structures constant.
11517 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11519 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11522 * Fixes for newer kerberos headers. NB: the casts are needed because
11523 the 'length' field is signed on one version and unsigned on another
11524 with no (?) obvious way to tell the difference, without these VC++
11525 complains. Also the "definition" of FAR (blank) is no longer included
11526 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11527 some needed definitions.
11531 * Undo Cygwin change.
11535 * Added support for proxy certificates according to RFC 3820.
11536 Because they may be a security thread to unaware applications,
11537 they must be explicitly allowed in run-time. See
11538 docs/HOWTO/proxy_certificates.txt for further information.
11542 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11544 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11545 server and client random values. Previously
11546 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11547 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11549 This change has negligible security impact because:
11551 1. Server and client random values still have 24 bytes of pseudo random
11554 2. Server and client random values are sent in the clear in the initial
11557 3. The master secret is derived using the premaster secret (48 bytes in
11558 size for static RSA ciphersuites) as well as client server and random
11561 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11564 *Stephen Henson, reported by UK NISCC*
11566 * Use Windows randomness collection on Cygwin.
11570 * Fix hang in EGD/PRNGD query when communication socket is closed
11571 prematurely by EGD/PRNGD.
11573 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11575 * Prompt for pass phrases when appropriate for PKCS12 input format.
11579 * Back-port of selected performance improvements from development
11580 branch, as well as improved support for PowerPC platforms.
11584 * Add lots of checks for memory allocation failure, error codes to indicate
11585 failure and freeing up memory if a failure occurs.
11587 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11589 * Add new -passin argument to dgst.
11593 * Perform some character comparisons of different types in X509_NAME_cmp:
11594 this is needed for some certificates that re-encode DNs into UTF8Strings
11595 (in violation of RFC3280) and can't or won't issue name rollover
11600 * Make an explicit check during certificate validation to see that
11601 the CA setting in each certificate on the chain is correct. As a
11602 side effect always do the following basic checks on extensions,
11603 not just when there's an associated purpose to the check:
11605 - if there is an unhandled critical extension (unless the user
11606 has chosen to ignore this fault)
11607 - if the path length has been exceeded (if one is set at all)
11608 - that certain extensions fit the associated purpose (if one has
11613 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11615 * Avoid a race condition when CRLs are checked in a multi threaded
11616 environment. This would happen due to the reordering of the revoked
11617 entries during signature checking and serial number lookup. Now the
11618 encoding is cached and the serial number sort performed under a lock.
11619 Add new STACK function sk_is_sorted().
11623 * Add Delta CRL to the extension code.
11627 * Various fixes to s3_pkt.c so alerts are sent properly.
11629 *David Holmes <d.holmes@f5.com>*
11631 * Reduce the chances of duplicate issuer name and serial numbers (in
11632 violation of RFC3280) using the OpenSSL certificate creation utilities.
11633 This is done by creating a random 64 bit value for the initial serial
11634 number when a serial number file is created or when a self signed
11635 certificate is created using 'openssl req -x509'. The initial serial
11636 number file is created using 'openssl x509 -next_serial' in CA.pl
11637 rather than being initialized to 1.
11641 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11643 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11644 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11646 *Joe Orton, Steve Henson*
11648 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11651 *Joe Orton, Steve Henson*
11653 * Make it possible to have multiple active certificates with the same
11654 subject in the CA index file. This is done only if the keyword
11655 'unique_subject' is set to 'no' in the main CA section (default
11656 if 'CA_default') of the configuration file. The value is saved
11657 with the database itself in a separate index attribute file,
11658 named like the index file with '.attr' appended to the name.
11662 * X509 verify fixes. Disable broken certificate workarounds when
11663 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11664 keyUsage extension present. Don't accept CRLs with unhandled critical
11665 extensions: since verify currently doesn't process CRL extensions this
11666 rejects a CRL with *any* critical extensions. Add new verify error codes
11671 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11672 A clarification of RFC2560 will require the use of OCTET STRINGs and
11673 some implementations cannot handle the current raw format. Since OpenSSL
11674 copies and compares OCSP nonces as opaque blobs without any attempt at
11675 parsing them this should not create any compatibility issues.
11679 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11680 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11681 this HMAC (and other) operations are several times slower than OpenSSL
11686 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11688 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11690 * Use the correct content when signing type "other".
11694 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11696 * Fix various bugs revealed by running the NISCC test suite:
11698 Stop out of bounds reads in the ASN1 code when presented with
11699 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11701 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11703 If verify callback ignores invalid public key errors don't try to check
11704 certificate signature with the NULL public key.
11708 * New -ignore_err option in ocsp application to stop the server
11709 exiting on the first error in a request.
11713 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11714 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11719 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11720 extra data after the compression methods not only for TLS 1.0
11721 but also for SSL 3.0 (as required by the specification).
11723 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11725 * Change X509_certificate_type() to mark the key as exported/exportable
11726 when it's 512 *bits* long, not 512 bytes.
11730 * Change AES_cbc_encrypt() so it outputs exact multiple of
11731 blocks during encryption.
11735 * Various fixes to base64 BIO and non blocking I/O. On write
11736 flushes were not handled properly if the BIO retried. On read
11737 data was not being buffered properly and had various logic bugs.
11738 This also affects blocking I/O when the data being decoded is a
11743 * Various S/MIME bugfixes and compatibility changes:
11744 output correct application/pkcs7 MIME type if
11745 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11746 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11747 of files as .eml work). Correctly handle very long lines in MIME
11752 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11754 * Countermeasure against the Klima-Pokorny-Rosa extension of
11755 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11756 a protocol version number mismatch like a decryption error
11757 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11761 * Turn on RSA blinding by default in the default implementation
11762 to avoid a timing attack. Applications that don't want it can call
11763 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11764 They would be ill-advised to do so in most cases.
11766 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11768 * Change RSA blinding code so that it works when the PRNG is not
11769 seeded (in this case, the secret RSA exponent is abused as
11770 an unpredictable seed -- if it is not unpredictable, there
11771 is no point in blinding anyway). Make RSA blinding thread-safe
11772 by remembering the creator's thread ID in rsa->blinding and
11773 having all other threads use local one-time blinding factors
11774 (this requires more computation than sharing rsa->blinding, but
11775 avoids excessive locking; and if an RSA object is not shared
11776 between threads, blinding will still be very fast).
11780 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11781 ENGINE as defaults for all supported algorithms irrespective of
11782 the 'flags' parameter. 'flags' is now honoured, so applications
11783 should make sure they are passing it correctly.
11787 * Target "mingw" now allows native Windows code to be generated in
11788 the Cygwin environment as well as with the MinGW compiler.
11792 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11794 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11795 via timing by performing a MAC computation even if incorrect
11796 block cipher padding has been found. This is a countermeasure
11797 against active attacks where the attacker has to distinguish
11798 between bad padding and a MAC verification error. ([CVE-2003-0078])
11800 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11801 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11802 Martin Vuagnoux (EPFL, Ilion)*
11804 * Make the no-err option work as intended. The intention with no-err
11805 is not to have the whole error stack handling routines removed from
11806 libcrypto, it's only intended to remove all the function name and
11807 reason texts, thereby removing some of the footprint that may not
11808 be interesting if those errors aren't displayed anyway.
11810 NOTE: it's still possible for any application or module to have its
11811 own set of error texts inserted. The routines are there, just not
11812 used by default when no-err is given.
11816 * Add support for FreeBSD on IA64.
11818 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11820 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11821 Kerberos function mit_des_cbc_cksum(). Before this change,
11822 the value returned by DES_cbc_cksum() was like the one from
11823 mit_des_cbc_cksum(), except the bytes were swapped.
11825 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11827 * Allow an application to disable the automatic SSL chain building.
11828 Before this a rather primitive chain build was always performed in
11829 ssl3_output_cert_chain(): an application had no way to send the
11830 correct chain if the automatic operation produced an incorrect result.
11832 Now the chain builder is disabled if either:
11834 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11836 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11838 The reasoning behind this is that an application would not want the
11839 auto chain building to take place if extra chain certificates are
11840 present and it might also want a means of sending no additional
11841 certificates (for example the chain has two certificates and the
11846 * Add the possibility to build without the ENGINE framework.
11848 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11850 * Under Win32 gmtime() can return NULL: check return value in
11851 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11855 * DSA routines: under certain error conditions uninitialized BN objects
11856 could be freed. Solution: make sure initialization is performed early
11857 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11858 Nils Larsch <nla@trustcenter.de> via PR#459)
11862 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11863 checked on reconnect on the client side, therefore session resumption
11864 could still fail with a "ssl session id is different" error. This
11865 behaviour is masked when SSL_OP_ALL is used due to
11866 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11867 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11868 followup to PR #377.
11872 * IA-32 assembler support enhancements: unified ELF targets, support
11873 for SCO/Caldera platforms, fix for Cygwin shared build.
11877 * Add support for FreeBSD on sparc64. As a consequence, support for
11878 FreeBSD on non-x86 processors is separate from x86 processors on
11879 the config script, much like the NetBSD support.
11881 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11883 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11885 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11888 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11889 code (06) was taken as the first octet of the session ID and the last
11890 octet was ignored consequently. As a result SSLv2 client side session
11891 caching could not have worked due to the session ID mismatch between
11893 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11898 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11899 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11904 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11905 seems that in spite of existing for more than a year, many application
11906 author have done nothing to provide the necessary callbacks, which
11907 means that this particular engine will not work properly anywhere.
11908 This is a very unfortunate situation which forces us, in the name
11909 of usability, to give the hw_ncipher.c a static lock, which is part
11911 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11912 appear in 0.9.8 or later. We EXPECT application authors to have
11913 dealt properly with this when 0.9.8 is released (unless we actually
11914 make such changes in the libcrypto locking code that changes will
11915 have to be made anyway).
11919 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11920 octets have been read, EOF or an error occurs. Without this change
11921 some truncated ASN1 structures will not produce an error.
11925 * Disable Heimdal support, since it hasn't been fully implemented.
11926 Still give the possibility to force the use of Heimdal, but with
11927 warnings and a request that patches get sent to openssl-dev.
11931 * Add the VC-CE target, introduce the WINCE sysname, and add
11932 INSTALL.WCE and appropriate conditionals to make it build.
11934 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11936 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11937 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11938 edit numbers of the version.
11940 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11942 * Introduce safe string copy and catenation functions
11943 (BUF_strlcpy() and BUF_strlcat()).
11945 *Ben Laurie (CHATS) and Richard Levitte*
11947 * Avoid using fixed-size buffers for one-line DNs.
11949 *Ben Laurie (CHATS)*
11951 * Add BUF_MEM_grow_clean() to avoid information leakage when
11952 resizing buffers containing secrets, and use where appropriate.
11954 *Ben Laurie (CHATS)*
11956 * Avoid using fixed size buffers for configuration file location.
11958 *Ben Laurie (CHATS)*
11960 * Avoid filename truncation for various CA files.
11962 *Ben Laurie (CHATS)*
11964 * Use sizeof in preference to magic numbers.
11966 *Ben Laurie (CHATS)*
11968 * Avoid filename truncation in cert requests.
11970 *Ben Laurie (CHATS)*
11972 * Add assertions to check for (supposedly impossible) buffer
11975 *Ben Laurie (CHATS)*
11977 * Don't cache truncated DNS entries in the local cache (this could
11978 potentially lead to a spoofing attack).
11980 *Ben Laurie (CHATS)*
11982 * Fix various buffers to be large enough for hex/decimal
11983 representations in a platform independent manner.
11985 *Ben Laurie (CHATS)*
11987 * Add CRYPTO_realloc_clean() to avoid information leakage when
11988 resizing buffers containing secrets, and use where appropriate.
11990 *Ben Laurie (CHATS)*
11992 * Add BIO_indent() to avoid much slightly worrying code to do
11995 *Ben Laurie (CHATS)*
11997 * Convert sprintf()/BIO_puts() to BIO_printf().
11999 *Ben Laurie (CHATS)*
12001 * buffer_gets() could terminate with the buffer only half
12004 *Ben Laurie (CHATS)*
12006 * Add assertions to prevent user-supplied crypto functions from
12007 overflowing internal buffers by having large block sizes, etc.
12009 *Ben Laurie (CHATS)*
12011 * New OPENSSL_assert() macro (similar to assert(), but enabled
12014 *Ben Laurie (CHATS)*
12016 * Eliminate unused copy of key in RC4.
12018 *Ben Laurie (CHATS)*
12020 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12022 *Ben Laurie (CHATS)*
12024 * Fix off-by-one error in EGD path.
12026 *Ben Laurie (CHATS)*
12028 * If RANDFILE path is too long, ignore instead of truncating.
12030 *Ben Laurie (CHATS)*
12032 * Eliminate unused and incorrectly sized X.509 structure
12035 *Ben Laurie (CHATS)*
12037 * Eliminate unused and dangerous function knumber().
12039 *Ben Laurie (CHATS)*
12041 * Eliminate unused and dangerous structure, KSSL_ERR.
12043 *Ben Laurie (CHATS)*
12045 * Protect against overlong session ID context length in an encoded
12046 session object. Since these are local, this does not appear to be
12049 *Ben Laurie (CHATS)*
12051 * Change from security patch (see 0.9.6e below) that did not affect
12052 the 0.9.6 release series:
12054 Remote buffer overflow in SSL3 protocol - an attacker could
12055 supply an oversized master key in Kerberos-enabled versions.
12058 *Ben Laurie (CHATS)*
12060 * Change the SSL kerb5 codes to match RFC 2712.
12064 * Make -nameopt work fully for req and add -reqopt switch.
12066 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12068 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12070 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12072 * Make sure tests can be performed even if the corresponding algorithms
12073 have been removed entirely. This was also the last step to make
12074 OpenSSL compilable with DJGPP under all reasonable conditions.
12076 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12078 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12079 to allow version independent disabling of normally unselected ciphers,
12080 which may be activated as a side-effect of selecting a single cipher.
12082 (E.g., cipher list string "RSA" enables ciphersuites that are left
12083 out of "ALL" because they do not provide symmetric encryption.
12084 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12086 *Lutz Jaenicke, Bodo Moeller*
12088 * Add appropriate support for separate platform-dependent build
12089 directories. The recommended way to make a platform-dependent
12090 build directory is the following (tested on Linux), maybe with
12093 # Place yourself outside of the OpenSSL source tree. In
12094 # this example, the environment variable OPENSSL_SOURCE
12095 # is assumed to contain the absolute OpenSSL source directory.
12096 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12097 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12098 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12099 mkdir -p `dirname $F`
12100 ln -s $OPENSSL_SOURCE/$F $F
12103 To be absolutely sure not to disturb the source tree, a "make clean"
12104 is a good thing. If it isn't successful, don't worry about it,
12105 it probably means the source directory is very clean.
12109 * Make sure any ENGINE control commands make local copies of string
12110 pointers passed to them whenever necessary. Otherwise it is possible
12111 the caller may have overwritten (or deallocated) the original string
12112 data when a later ENGINE operation tries to use the stored values.
12114 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12116 * Improve diagnostics in file reading and command-line digests.
12118 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12120 * Add AES modes CFB and OFB to the object database. Correct an
12121 error in AES-CFB decryption.
12125 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12126 allows existing EVP_CIPHER_CTX structures to be reused after
12127 calling `EVP_*Final()`. This behaviour is used by encryption
12128 BIOs and some applications. This has the side effect that
12129 applications must explicitly clean up cipher contexts with
12130 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12134 * Check the values of dna and dnb in bn_mul_recursive before calling
12135 bn_mul_comba (a non zero value means the a or b arrays do not contain
12136 n2 elements) and fallback to bn_mul_normal if either is not zero.
12140 * Fix escaping of non-ASCII characters when using the -subj option
12141 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12145 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12146 form for "surname", serialNumber has no short form.
12147 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12148 therefore remove "mail" short name for "internet 7".
12149 The OID for unique identifiers in X509 certificates is
12150 x500UniqueIdentifier, not uniqueIdentifier.
12151 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12155 * Add an "init" command to the ENGINE config module and auto initialize
12156 ENGINEs. Without any "init" command the ENGINE will be initialized
12157 after all ctrl commands have been executed on it. If init=1 the
12158 ENGINE is initialized at that point (ctrls before that point are run
12159 on the uninitialized ENGINE and after on the initialized one). If
12160 init=0 then the ENGINE will not be initialized at all.
12164 * Fix the 'app_verify_callback' interface so that the user-defined
12165 argument is actually passed to the callback: In the
12166 SSL_CTX_set_cert_verify_callback() prototype, the callback
12167 declaration has been changed from
12170 int (*cb)(X509_STORE_CTX *,void *);
12171 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12172 i=s->ctx->app_verify_callback(&ctx)
12173 has been changed into
12174 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12176 To update applications using SSL_CTX_set_cert_verify_callback(),
12177 a dummy argument can be added to their callback functions.
12179 *D. K. Smetters <smetters@parc.xerox.com>*
12181 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12183 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12185 * Add and OPENSSL_LOAD_CONF define which will cause
12186 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12187 This allows older applications to transparently support certain
12188 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12189 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12190 load the config file and OPENSSL_add_all_algorithms_conf() which will
12191 always load it have also been added.
12195 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12196 Adjust NIDs and EVP layer.
12198 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12200 * Config modules support in openssl utility.
12202 Most commands now load modules from the config file,
12203 though in a few (such as version) this isn't done
12204 because it couldn't be used for anything.
12206 In the case of ca and req the config file used is
12207 the same as the utility itself: that is the -config
12208 command line option can be used to specify an
12213 * Move default behaviour from OPENSSL_config(). If appname is NULL
12214 use "openssl_conf" if filename is NULL use default openssl config file.
12218 * Add an argument to OPENSSL_config() to allow the use of an alternative
12219 config section name. Add a new flag to tolerate a missing config file
12220 and move code to CONF_modules_load_file().
12224 * Support for crypto accelerator cards from Accelerated Encryption
12225 Processing, www.aep.ie. (Use engine 'aep')
12226 The support was copied from 0.9.6c [engine] and adapted/corrected
12227 to work with the new engine framework.
12229 *AEP Inc. and Richard Levitte*
12231 * Support for SureWare crypto accelerator cards from Baltimore
12232 Technologies. (Use engine 'sureware')
12233 The support was copied from 0.9.6c [engine] and adapted
12234 to work with the new engine framework.
12238 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12239 make the newer ENGINE framework commands for the CHIL engine work.
12241 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12243 * Make it possible to produce shared libraries on ReliantUNIX.
12245 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12247 * Add the configuration target debug-linux-ppro.
12248 Make 'openssl rsa' use the general key loading routines
12249 implemented in `apps.c`, and make those routines able to
12250 handle the key format FORMAT_NETSCAPE and the variant
12253 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12255 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12257 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12259 * Add -keyform to rsautl, and document -engine.
12261 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12263 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12264 BIO_R_NO_SUCH_FILE error code rather than the generic
12265 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12269 * Add new functions
12270 ERR_peek_last_error
12271 ERR_peek_last_error_line
12272 ERR_peek_last_error_line_data.
12273 These are similar to
12275 ERR_peek_error_line
12276 ERR_peek_error_line_data,
12277 but report on the latest error recorded rather than the first one
12278 still in the error queue.
12280 *Ben Laurie, Bodo Moeller*
12282 * default_algorithms option in ENGINE config module. This allows things
12284 default_algorithms = ALL
12285 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12289 * Preliminary ENGINE config module.
12293 * New experimental application configuration code.
12297 * Change the AES code to follow the same name structure as all other
12298 symmetric ciphers, and behave the same way. Move everything to
12299 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12301 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12303 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12305 *Ben Laurie and Theo de Raadt*
12307 * Add option to output public keys in req command.
12309 *Massimiliano Pala madwolf@openca.org*
12311 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12312 (up to about 10% better than before for P-192 and P-224).
12316 * New functions/macros
12318 SSL_CTX_set_msg_callback(ctx, cb)
12319 SSL_CTX_set_msg_callback_arg(ctx, arg)
12320 SSL_set_msg_callback(ssl, cb)
12321 SSL_set_msg_callback_arg(ssl, arg)
12323 to request calling a callback function
12325 void cb(int write_p, int version, int content_type,
12326 const void *buf, size_t len, SSL *ssl, void *arg)
12328 whenever a protocol message has been completely received
12329 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12330 protocol version according to which the SSL library interprets
12331 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12332 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12333 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12334 specification (change_cipher_spec(20), alert(21), handshake(22)).
12335 'buf' and 'len' point to the actual message, 'ssl' to the
12336 SSL object, and 'arg' is the application-defined value set by
12337 SSL[_CTX]_set_msg_callback_arg().
12339 'openssl s_client' and 'openssl s_server' have new '-msg' options
12340 to enable a callback that displays all protocol messages.
12344 * Change the shared library support so shared libraries are built as
12345 soon as the corresponding static library is finished, and thereby get
12346 openssl and the test programs linked against the shared library.
12347 This still only happens when the keyword "shard" has been given to
12348 the configuration scripts.
12350 NOTE: shared library support is still an experimental thing, and
12351 backward binary compatibility is still not guaranteed.
12353 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12355 * Add support for Subject Information Access extension.
12357 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12359 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12360 additional bytes when new memory had to be allocated, not just
12361 when reusing an existing buffer.
12365 * New command line and configuration option 'utf8' for the req command.
12366 This allows field values to be specified as UTF8 strings.
12370 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12371 runs for the former and machine-readable output for the latter.
12375 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12376 of the e-mail address in the DN (i.e., it will go into a certificate
12377 extension only). The new configuration file option 'email_in_dn = no'
12378 has the same effect.
12380 *Massimiliano Pala madwolf@openca.org*
12382 * Change all functions with names starting with `des_` to be starting
12383 with `DES_` instead. Add wrappers that are compatible with libdes,
12384 but are named `_ossl_old_des_*`. Finally, add macros that map the
12385 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12386 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12387 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12390 Since we provide two compatibility mappings, the user needs to
12391 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12392 compatibility is desired. The default (i.e., when that macro
12393 isn't defined) is OpenSSL 0.9.6c compatibility.
12395 There are also macros that enable and disable the support of old
12396 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12397 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12398 are defined, the default will apply: to support the old des routines.
12400 In either case, one must include openssl/des.h to get the correct
12401 definitions. Do not try to just include openssl/des_old.h, that
12404 NOTE: This is a major break of an old API into a new one. Software
12405 authors are encouraged to switch to the `DES_` style functions. Some
12406 time in the future, des_old.h and the libdes compatibility functions
12407 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12408 default), and then completely removed.
12412 * Test for certificates which contain unsupported critical extensions.
12413 If such a certificate is found during a verify operation it is
12414 rejected by default: this behaviour can be overridden by either
12415 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12416 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12417 X509_supported_extension() has also been added which returns 1 if a
12418 particular extension is supported.
12422 * Modify the behaviour of EVP cipher functions in similar way to digests
12423 to retain compatibility with existing code.
12427 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12428 compatibility with existing code. In particular the 'ctx' parameter does
12429 not have to be to be initialized before the call to EVP_DigestInit() and
12430 it is tidied up after a call to EVP_DigestFinal(). New function
12431 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12432 EVP_MD_CTX_copy() changed to not require the destination to be
12433 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12434 requires the destination to be valid.
12436 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12437 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12441 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12442 so that complete 'Handshake' protocol structures are kept in memory
12443 instead of overwriting 'msg_type' and 'length' with 'body' data.
12447 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12449 *Massimo Santin via Richard Levitte*
12451 * Major restructuring to the underlying ENGINE code. This includes
12452 reduction of linker bloat, separation of pure "ENGINE" manipulation
12453 (initialisation, etc) from functionality dealing with implementations
12454 of specific crypto interfaces. This change also introduces integrated
12455 support for symmetric ciphers and digest implementations - so ENGINEs
12456 can now accelerate these by providing EVP_CIPHER and EVP_MD
12457 implementations of their own. This is detailed in
12458 [crypto/engine/README.md](crypto/engine/README.md)
12459 as it couldn't be adequately described here. However, there are a few
12460 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12461 were changed in the original introduction of ENGINE code have now
12462 reverted back - the hooking from this code to ENGINE is now a good
12463 deal more passive and at run-time, operations deal directly with
12464 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12465 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12466 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12467 they were not being used by the framework as there is no concept of a
12468 BIGNUM_METHOD and they could not be generalised to the new
12469 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12470 ENGINE_cpy() has been removed as it cannot be consistently defined in
12475 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12479 * Change mkdef.pl to sort symbols that get the same entry number,
12480 and make sure the automatically generated functions `ERR_load_*`
12481 become part of libeay.num as well.
12485 * New function SSL_renegotiate_pending(). This returns true once
12486 renegotiation has been requested (either SSL_renegotiate() call
12487 or HelloRequest/ClientHello received from the peer) and becomes
12488 false once a handshake has been completed.
12489 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12490 sends a HelloRequest, but does not ensure that a handshake takes
12491 place. SSL_renegotiate_pending() is useful for checking if the
12492 client has followed the request.)
12496 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12497 By default, clients may request session resumption even during
12498 renegotiation (if session ID contexts permit); with this option,
12499 session resumption is possible only in the first handshake.
12501 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12502 more bits available for options that should not be part of
12503 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12507 * Add some demos for certificate and certificate request creation.
12511 * Make maximum certificate chain size accepted from the peer application
12512 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12513 "Douglas E. Engert" <deengert@anl.gov>.
12517 * Add support for shared libraries for Unixware-7
12518 (Boyd Lynn Gerber <gerberb@zenez.com>).
12522 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12523 be done prior to destruction. Use this to unload error strings from
12524 ENGINEs that load their own error strings. NB: This adds two new API
12525 functions to "get" and "set" this destroy handler in an ENGINE.
12529 * Alter all existing ENGINE implementations (except "openssl" and
12530 "openbsd") to dynamically instantiate their own error strings. This
12531 makes them more flexible to be built both as statically-linked ENGINEs
12532 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12533 Also, add stub code to each that makes building them as self-contained
12534 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12538 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12539 implementations into applications that are completely implemented in
12540 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12541 commands that can be used to configure what shared-library to load and
12542 to control aspects of the way it is handled. Also, made an update to
12543 the [README-Engine.md](README-Engine.md) file
12544 that brings its information up-to-date and
12545 provides some information and instructions on the "dynamic" ENGINE
12546 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12550 * Make it possible to unload ranges of ERR strings with a new
12551 "ERR_unload_strings" function.
12555 * Add a copy() function to EVP_MD.
12559 * Make EVP_MD routines take a context pointer instead of just the
12560 md_data void pointer.
12564 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12565 that the digest can only process a single chunk of data
12566 (typically because it is provided by a piece of
12567 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12568 is only going to provide a single chunk of data, and hence the
12569 framework needn't accumulate the data for oneshot drivers.
12573 * As with "ERR", make it possible to replace the underlying "ex_data"
12574 functions. This change also alters the storage and management of global
12575 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12576 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12577 index counters. The API functions that use this state have been changed
12578 to take a "class_index" rather than pointers to the class's local STACK
12579 and counter, and there is now an API function to dynamically create new
12580 classes. This centralisation allows us to (a) plug a lot of the
12581 thread-safety problems that existed, and (b) makes it possible to clean
12582 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12583 such data would previously have always leaked in application code and
12584 workarounds were in place to make the memory debugging turn a blind eye
12585 to it. Application code that doesn't use this new function will still
12586 leak as before, but their memory debugging output will announce it now
12587 rather than letting it slide.
12589 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12590 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12591 has a return value to indicate success or failure.
12595 * Make it possible to replace the underlying "ERR" functions such that the
12596 global state (2 LHASH tables and 2 locks) is only used by the "default"
12597 implementation. This change also adds two functions to "get" and "set"
12598 the implementation prior to it being automatically set the first time
12599 any other ERR function takes place. Ie. an application can call "get",
12600 pass the return value to a module it has just loaded, and that module
12601 can call its own "set" function using that value. This means the
12602 module's "ERR" operations will use (and modify) the error state in the
12603 application and not in its own statically linked copy of OpenSSL code.
12607 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12608 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12609 the operation, and provides a more encapsulated way for external code
12610 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12611 to use these functions rather than manually incrementing the counts.
12613 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12617 * Add EVP test program.
12621 * Add symmetric cipher support to ENGINE. Expect the API to change!
12625 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12626 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12627 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12628 These allow a CRL to be built without having to access X509_CRL fields
12629 directly. Modify 'ca' application to use new functions.
12633 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12634 bug workarounds. Rollback attack detection is a security feature.
12635 The problem will only arise on OpenSSL servers when TLSv1 is not
12636 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12637 Software authors not wanting to support TLSv1 will have special reasons
12638 for their choice and can explicitly enable this option.
12640 *Bodo Moeller, Lutz Jaenicke*
12642 * Rationalise EVP so it can be extended: don't include a union of
12643 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12644 (similar to those existing for EVP_CIPHER_CTX).
12649 EVP_MD_CTX_init(&md); /* new function call */
12650 EVP_DigestInit(&md, EVP_sha1());
12651 EVP_DigestUpdate(&md, in, len);
12652 EVP_DigestFinal(&md, out, NULL);
12653 EVP_MD_CTX_cleanup(&md); /* new function call */
12657 * Make DES key schedule conform to the usual scheme, as well as
12658 correcting its structure. This means that calls to DES functions
12659 now have to pass a pointer to a des_key_schedule instead of a
12660 plain des_key_schedule (which was actually always a pointer
12663 des_key_schedule ks;
12665 des_set_key_checked(..., &ks);
12666 des_ncbc_encrypt(..., &ks, ...);
12668 (Note that a later change renames 'des_...' into 'DES_...'.)
12672 * Initial reduction of linker bloat: the use of some functions, such as
12673 PEM causes large amounts of unused functions to be linked in due to
12674 poor organisation. For example pem_all.c contains every PEM function
12675 which has a knock on effect of linking in large amounts of (unused)
12676 ASN1 code. Grouping together similar functions and splitting unrelated
12677 functions prevents this.
12681 * Cleanup of EVP macros.
12685 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12686 correct `_ecb suffix`.
12690 * Add initial OCSP responder support to ocsp application. The
12691 revocation information is handled using the text based index
12692 use by the ca application. The responder can either handle
12693 requests generated internally, supplied in files (for example
12694 via a CGI script) or using an internal minimal server.
12698 * Add configuration choices to get zlib compression for TLS.
12702 * Changes to Kerberos SSL for RFC 2712 compliance:
12703 1. Implemented real KerberosWrapper, instead of just using
12704 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12705 2. Implemented optional authenticator field of KerberosWrapper.
12707 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12708 and authenticator structs; see crypto/krb5/.
12710 Generalized Kerberos calls to support multiple Kerberos libraries.
12711 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12712 via Richard Levitte*
12714 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12715 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12716 values for each of the key sizes rather than having just
12717 parameters (and 'speed' generating keys each time).
12721 * Speed up EVP routines.
12724 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12725 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12726 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12727 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12729 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12730 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12731 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12734 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12736 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12740 * Added the OS2-EMX target.
12742 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12744 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12745 New functions to support `NCONF `routines in extension code.
12746 New function `CONF_set_nconf()`
12747 to allow functions which take an `NCONF` to also handle the old `LHASH`
12748 structure: this means that the old `CONF` compatible routines can be
12749 retained (in particular w.rt. extensions) without having to duplicate the
12750 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12754 * Enhance the general user interface with mechanisms for inner control
12755 and with possibilities to have yes/no kind of prompts.
12759 * Change all calls to low level digest routines in the library and
12760 applications to use EVP. Add missing calls to HMAC_cleanup() and
12761 don't assume HMAC_CTX can be copied using memcpy().
12763 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12765 * Add the possibility to control engines through control names but with
12766 arbitrary arguments instead of just a string.
12767 Change the key loaders to take a UI_METHOD instead of a callback
12768 function pointer. NOTE: this breaks binary compatibility with earlier
12769 versions of OpenSSL [engine].
12770 Adapt the nCipher code for these new conditions and add a card insertion
12775 * Enhance the general user interface with mechanisms to better support
12776 dialog box interfaces, application-defined prompts, the possibility
12777 to use defaults (for example default passwords from somewhere else)
12778 and interrupts/cancellations.
12782 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12783 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12787 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12788 tidy up some unnecessarily weird code in 'sk_new()').
12790 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12792 * Change the key loading routines for ENGINEs to use the same kind
12793 callback (pem_password_cb) as all other routines that need this
12798 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12799 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12800 than this minimum value is recommended.
12804 * New random seeder for OpenVMS, using the system process statistics
12805 that are easily reachable.
12809 * Windows apparently can't transparently handle global
12810 variables defined in DLLs. Initialisations such as:
12812 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12814 won't compile. This is used by the any applications that need to
12815 declare their own ASN1 modules. This was fixed by adding the option
12816 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12817 needed for static libraries under Win32.
12821 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12822 setting of purpose and trust fields. New X509_STORE trust and
12823 purpose functions and tidy up setting in other SSL functions.
12827 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12828 structure. These are inherited by X509_STORE_CTX when it is
12829 initialised. This allows various defaults to be set in the
12830 X509_STORE structure (such as flags for CRL checking and custom
12831 purpose or trust settings) for functions which only use X509_STORE_CTX
12832 internally such as S/MIME.
12834 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12835 trust settings if they are not set in X509_STORE. This allows X509_STORE
12836 purposes and trust (in S/MIME for example) to override any set by default.
12838 Add command line options for CRL checking to smime, s_client and s_server
12843 * Initial CRL based revocation checking. If the CRL checking flag(s)
12844 are set then the CRL is looked up in the X509_STORE structure and
12845 its validity and signature checked, then if the certificate is found
12846 in the CRL the verify fails with a revoked error.
12848 Various new CRL related callbacks added to X509_STORE_CTX structure.
12850 Command line options added to 'verify' application to support this.
12852 This needs some additional work, such as being able to handle multiple
12853 CRLs with different times, extension based lookup (rather than just
12854 by subject name) and ultimately more complete V2 CRL extension
12859 * Add a general user interface API (crypto/ui/). This is designed
12860 to replace things like des_read_password and friends (backward
12861 compatibility functions using this new API are provided).
12862 The purpose is to remove prompting functions from the DES code
12863 section as well as provide for prompting through dialog boxes in
12864 a window system and the like.
12868 * Add "ex_data" support to ENGINE so implementations can add state at a
12869 per-structure level rather than having to store it globally.
12873 * Make it possible for ENGINE structures to be copied when retrieved by
12874 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12875 This causes the "original" ENGINE structure to act like a template,
12876 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12877 operational state can be localised to each ENGINE structure, despite the
12878 fact they all share the same "methods". New ENGINE structures returned in
12879 this case have no functional references and the return value is the single
12880 structural reference. This matches the single structural reference returned
12881 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12886 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12887 needs to match any other type at all we need to manually clear the
12892 * Changes to the "openssl engine" utility to include;
12893 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12894 about an ENGINE's available control commands.
12895 - executing control commands from command line arguments using the
12896 '-pre' and '-post' switches. '-post' is only used if '-t' is
12897 specified and the ENGINE is successfully initialised. The syntax for
12898 the individual commands are colon-separated, for example;
12899 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12903 * New dynamic control command support for ENGINEs. ENGINEs can now
12904 declare their own commands (numbers), names (strings), descriptions,
12905 and input types for run-time discovery by calling applications. A
12906 subset of these commands are implicitly classed as "executable"
12907 depending on their input type, and only these can be invoked through
12908 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12909 can be based on user input, config files, etc). The distinction is
12910 that "executable" commands cannot return anything other than a boolean
12911 result and can only support numeric or string input, whereas some
12912 discoverable commands may only be for direct use through
12913 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12914 pointers, or other custom uses. The "executable" commands are to
12915 support parameterisations of ENGINE behaviour that can be
12916 unambiguously defined by ENGINEs and used consistently across any
12917 OpenSSL-based application. Commands have been added to all the
12918 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12919 control over shared-library paths without source code alterations.
12923 * Changed all ENGINE implementations to dynamically allocate their
12924 ENGINEs rather than declaring them statically. Apart from this being
12925 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12926 this also allows the implementations to compile without using the
12927 internal engine_int.h header.
12931 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12932 'const' value. Any code that should be able to modify a RAND_METHOD
12933 should already have non-const pointers to it (ie. they should only
12934 modify their own ones).
12938 * Made a variety of little tweaks to the ENGINE code.
12939 - "atalla" and "ubsec" string definitions were moved from header files
12940 to C code. "nuron" string definitions were placed in variables
12941 rather than hard-coded - allowing parameterisation of these values
12942 later on via ctrl() commands.
12943 - Removed unused "#if 0"'d code.
12944 - Fixed engine list iteration code so it uses ENGINE_free() to release
12945 structural references.
12946 - Constified the RAND_METHOD element of ENGINE structures.
12947 - Constified various get/set functions as appropriate and added
12948 missing functions (including a catch-all ENGINE_cpy that duplicates
12949 all ENGINE values onto a new ENGINE except reference counts/state).
12950 - Removed NULL parameter checks in get/set functions. Setting a method
12951 or function to NULL is a way of cancelling out a previously set
12952 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12953 and doesn't justify the extra error symbols and code.
12954 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12955 flags from engine_int.h to engine.h.
12956 - Changed prototypes for ENGINE handler functions (init(), finish(),
12957 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12961 * Implement binary inversion algorithm for BN_mod_inverse in addition
12962 to the algorithm using long division. The binary algorithm can be
12963 used only if the modulus is odd. On 32-bit systems, it is faster
12964 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12965 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12966 up to 450 bits. In 64-bit environments, the binary algorithm
12967 appears to be advantageous for much longer moduli; here we use it
12968 for moduli up to 2048 bits.
12972 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12973 could not support the combine flag in choice fields.
12977 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12978 extensions from a certificate request to the certificate.
12982 * Allow multiple 'certopt' and 'nameopt' options to be separated
12983 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12984 file: this allows the display of the certificate about to be
12985 signed to be customised, to allow certain fields to be included
12986 or excluded and extension details. The old system didn't display
12987 multicharacter strings properly, omitted fields not in the policy
12988 and couldn't display additional details such as extensions.
12992 * Function EC_POINTs_mul for multiple scalar multiplication
12993 of an arbitrary number of elliptic curve points
12994 \sum scalars[i]*points[i],
12995 optionally including the generator defined for the EC_GROUP:
12996 scalar*generator + \sum scalars[i]*points[i].
12998 EC_POINT_mul is a simple wrapper function for the typical case
12999 that the point list has just one item (besides the optional
13004 * First EC_METHODs for curves over GF(p):
13006 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13007 operations and provides various method functions that can also
13008 operate with faster implementations of modular arithmetic.
13010 EC_GFp_mont_method() reuses most functions that are part of
13011 EC_GFp_simple_method, but uses Montgomery arithmetic.
13013 *Bodo Moeller; point addition and point doubling
13014 implementation directly derived from source code provided by
13015 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13017 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13018 crypto/ec/ec_lib.c):
13020 Curves are EC_GROUP objects (with an optional group generator)
13021 based on EC_METHODs that are built into the library.
13023 Points are EC_POINT objects based on EC_GROUP objects.
13025 Most of the framework would be able to handle curves over arbitrary
13026 finite fields, but as there are no obvious types for fields other
13027 than GF(p), some functions are limited to that for now.
13031 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13032 that the file contains a complete HTTP response.
13036 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13037 change the def and num file printf format specifier from "%-40sXXX"
13038 to "%-39s XXX". The latter will always guarantee a space after the
13039 field while the former will cause them to run together if the field
13040 is 40 of more characters long.
13044 * Constify the cipher and digest 'method' functions and structures
13045 and modify related functions to take constant EVP_MD and EVP_CIPHER
13050 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13051 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13055 * Modify `EVP_Digest*()` routines so they now return values. Although the
13056 internal software routines can never fail additional hardware versions
13061 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13063 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13064 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13067 ERR_R_NESTED_ASN1_ERROR
13069 ERR_R_MISSING_ASN1_EOS
13070 were 4 .. 9, conflicting with
13071 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13073 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13074 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13076 Add new error code 'ERR_R_INTERNAL_ERROR'.
13080 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13085 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13086 sets the subject name for a new request or supersedes the
13087 subject name in a given request. Formats that can be parsed are
13088 'CN=Some Name, OU=myOU, C=IT'
13090 'CN=Some Name/OU=myOU/C=IT'.
13092 Add options '-batch' and '-verbose' to 'openssl req'.
13094 *Massimiliano Pala <madwolf@hackmasters.net>*
13096 * Introduce the possibility to access global variables through
13097 functions on platform were that's the best way to handle exporting
13098 global variables in shared libraries. To enable this functionality,
13099 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13100 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13101 is normally done by Configure or something similar).
13103 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13104 in the source file (foo.c) like this:
13106 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13107 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13109 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13110 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13112 OPENSSL_DECLARE_GLOBAL(int,foo);
13113 #define foo OPENSSL_GLOBAL_REF(foo)
13114 OPENSSL_DECLARE_GLOBAL(double,bar);
13115 #define bar OPENSSL_GLOBAL_REF(bar)
13117 The #defines are very important, and therefore so is including the
13118 header file everywhere where the defined globals are used.
13120 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13121 of ASN.1 items, but that structure is a bit different.
13123 The largest change is in util/mkdef.pl which has been enhanced with
13124 better and easier to understand logic to choose which symbols should
13125 go into the Windows .def files as well as a number of fixes and code
13126 cleanup (among others, algorithm keywords are now sorted
13127 lexicographically to avoid constant rewrites).
13131 * In BN_div() keep a copy of the sign of 'num' before writing the
13132 result to 'rm' because if rm==num the value will be overwritten
13133 and produce the wrong result if 'num' is negative: this caused
13134 problems with BN_mod() and BN_nnmod().
13138 * Function OCSP_request_verify(). This checks the signature on an
13139 OCSP request and verifies the signer certificate. The signer
13140 certificate is just checked for a generic purpose and OCSP request
13145 * Add OCSP_check_validity() function to check the validity of OCSP
13146 responses. OCSP responses are prepared in real time and may only
13147 be a few seconds old. Simply checking that the current time lies
13148 between thisUpdate and nextUpdate max reject otherwise valid responses
13149 caused by either OCSP responder or client clock inaccuracy. Instead
13150 we allow thisUpdate and nextUpdate to fall within a certain period of
13151 the current time. The age of the response can also optionally be
13152 checked. Two new options -validity_period and -status_age added to
13157 * If signature or public key algorithm is unrecognized print out its
13158 OID rather that just UNKNOWN.
13162 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13163 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13164 ID to be generated from the issuer certificate alone which can then be
13165 passed to OCSP_id_issuer_cmp().
13169 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13170 ASN1 modules to export functions returning ASN1_ITEM pointers
13171 instead of the ASN1_ITEM structures themselves. This adds several
13172 new macros which allow the underlying ASN1 function/structure to
13173 be accessed transparently. As a result code should not use ASN1_ITEM
13174 references directly (such as &X509_it) but instead use the relevant
13175 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13176 use of the new ASN1 code on platforms where exporting structures
13177 is problematical (for example in shared libraries) but exporting
13178 functions returning pointers to structures is not.
13182 * Add support for overriding the generation of SSL/TLS session IDs.
13183 These callbacks can be registered either in an SSL_CTX or per SSL.
13184 The purpose of this is to allow applications to control, if they wish,
13185 the arbitrary values chosen for use as session IDs, particularly as it
13186 can be useful for session caching in multiple-server environments. A
13187 command-line switch for testing this (and any client code that wishes
13188 to use such a feature) has been added to "s_server".
13190 *Geoff Thorpe, Lutz Jaenicke*
13192 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13193 of the form `#if defined(...) || defined(...) || ...` and
13194 `#if !defined(...) && !defined(...) && ...`. This also avoids
13195 the growing number of special cases it was previously handling.
13199 * Make all configuration macros available for application by making
13200 sure they are available in opensslconf.h, by giving them names starting
13201 with `OPENSSL_` to avoid conflicts with other packages and by making
13202 sure e_os2.h will cover all platform-specific cases together with
13204 Additionally, it is now possible to define configuration/platform-
13205 specific names (called "system identities"). In the C code, these
13206 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13207 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13208 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13213 * New option -set_serial to 'req' and 'x509' this allows the serial
13214 number to use to be specified on the command line. Previously self
13215 signed certificates were hard coded with serial number 0 and the
13216 CA options of 'x509' had to use a serial number in a file which was
13221 * New options to 'ca' utility to support V2 CRL entry extensions.
13222 Currently CRL reason, invalidity date and hold instruction are
13223 supported. Add new CRL extensions to V3 code and some new objects.
13227 * New function EVP_CIPHER_CTX_set_padding() this is used to
13228 disable standard block padding (aka PKCS#5 padding) in the EVP
13229 API, which was previously mandatory. This means that the data is
13230 not padded in any way and so the total length much be a multiple
13231 of the block size, otherwise an error occurs.
13235 * Initial (incomplete) OCSP SSL support.
13239 * New function OCSP_parse_url(). This splits up a URL into its host,
13240 port and path components: primarily to parse OCSP URLs. New -url
13241 option to ocsp utility.
13245 * New nonce behavior. The return value of OCSP_check_nonce() now
13246 reflects the various checks performed. Applications can decide
13247 whether to tolerate certain situations such as an absent nonce
13248 in a response when one was present in a request: the ocsp application
13249 just prints out a warning. New function OCSP_add1_basic_nonce()
13250 this is to allow responders to include a nonce in a response even if
13251 the request is nonce-less.
13255 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13256 skipped when using openssl x509 multiple times on a single input file,
13257 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13261 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13262 set string type: to handle setting ASN1_TIME structures. Fix ca
13263 utility to correctly initialize revocation date of CRLs.
13267 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13268 the clients preferred ciphersuites and rather use its own preferences.
13269 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13270 Internet Explorer by ensuring unchanged hash method during stepup.
13271 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13275 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13276 to aes and add a new 'exist' option to print out symbols that don't
13281 * Additional options to ocsp utility to allow flags to be set and
13282 additional certificates supplied.
13286 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13287 OCSP client a number of certificate to only verify the response
13292 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13293 handle the new API. Currently only ECB, CBC modes supported. Add new
13296 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13297 Encryption Standard (AES) Ciphersuites for Transport Layer
13298 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13299 not enabled by default and were not part of the "ALL" ciphersuite
13300 alias because they were not yet official; they could be
13301 explicitly requested by specifying the "AESdraft" ciphersuite
13302 group alias. In the final release of OpenSSL 0.9.7, the group
13303 alias is called "AES" and is part of "ALL".)
13305 *Ben Laurie, Steve Henson, Bodo Moeller*
13307 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13308 request to response.
13312 * Functions for OCSP responders. OCSP_request_onereq_count(),
13313 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13314 extract information from a certificate request. OCSP_response_create()
13315 creates a response and optionally adds a basic response structure.
13316 OCSP_basic_add1_status() adds a complete single response to a basic
13317 response and returns the OCSP_SINGLERESP structure just added (to allow
13318 extensions to be included for example). OCSP_basic_add1_cert() adds a
13319 certificate to a basic response and OCSP_basic_sign() signs a basic
13320 response with various flags. New helper functions ASN1_TIME_check()
13321 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13322 (converts ASN1_TIME to GeneralizedTime).
13326 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13327 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13328 structure from a certificate. X509_pubkey_digest() digests the public_key
13329 contents: this is used in various key identifiers.
13333 * Make sk_sort() tolerate a NULL argument.
13335 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13337 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13338 passed by the function are trusted implicitly. If any of them signed the
13339 response then it is assumed to be valid and is not verified.
13343 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13344 to data. This was previously part of the PKCS7 ASN1 code. This
13345 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13346 *Steve Henson, reported by Kenneth R. Robinette
13347 <support@securenetterm.com>*
13349 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13350 routines: without these tracing memory leaks is very painful.
13351 Fix leaks in PKCS12 and PKCS7 routines.
13355 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13356 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13357 effectively meant GeneralizedTime would never be used. Now it
13358 is initialised to -1 but X509_time_adj() now has to check the value
13359 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13360 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13361 *Steve Henson, reported by Kenneth R. Robinette
13362 <support@securenetterm.com>*
13364 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13365 result in a zero length in the ASN1_INTEGER structure which was
13366 not consistent with the structure when d2i_ASN1_INTEGER() was used
13367 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13368 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13369 where it did not print out a minus for negative ASN1_INTEGER.
13373 * Add summary printout to ocsp utility. The various functions which
13374 convert status values to strings have been renamed to:
13375 OCSP_response_status_str(), OCSP_cert_status_str() and
13376 OCSP_crl_reason_str() and are no longer static. New options
13377 to verify nonce values and to disable verification. OCSP response
13378 printout format cleaned up.
13382 * Add additional OCSP certificate checks. These are those specified
13383 in RFC2560. This consists of two separate checks: the CA of the
13384 certificate being checked must either be the OCSP signer certificate
13385 or the issuer of the OCSP signer certificate. In the latter case the
13386 OCSP signer certificate must contain the OCSP signing extended key
13387 usage. This check is performed by attempting to match the OCSP
13388 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13389 in the OCSP_CERTID structures of the response.
13393 * Initial OCSP certificate verification added to OCSP_basic_verify()
13394 and related routines. This uses the standard OpenSSL certificate
13395 verify routines to perform initial checks (just CA validity) and
13396 to obtain the certificate chain. Then additional checks will be
13397 performed on the chain. Currently the root CA is checked to see
13398 if it is explicitly trusted for OCSP signing. This is used to set
13399 a root CA as a global signing root: that is any certificate that
13400 chains to that CA is an acceptable OCSP signing certificate.
13404 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13405 extensions from a separate configuration file.
13406 As when reading extensions from the main configuration file,
13407 the '-extensions ...' option may be used for specifying the
13410 *Massimiliano Pala <madwolf@comune.modena.it>*
13412 * New OCSP utility. Allows OCSP requests to be generated or
13413 read. The request can be sent to a responder and the output
13414 parsed, outputted or printed in text form. Not complete yet:
13415 still needs to check the OCSP response validity.
13419 * New subcommands for 'openssl ca':
13420 `openssl ca -status <serial>` prints the status of the cert with
13421 the given serial number (according to the index file).
13422 `openssl ca -updatedb` updates the expiry status of certificates
13425 *Massimiliano Pala <madwolf@comune.modena.it>*
13427 * New '-newreq-nodes' command option to CA.pl. This is like
13428 '-newreq', but calls 'openssl req' with the '-nodes' option
13429 so that the resulting key is not encrypted.
13431 *Damien Miller <djm@mindrot.org>*
13433 * New configuration for the GNU Hurd.
13435 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13437 * Initial code to implement OCSP basic response verify. This
13438 is currently incomplete. Currently just finds the signer's
13439 certificate and verifies the signature on the response.
13443 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13444 value of OPENSSLDIR. This is available via the new '-d' option
13445 to 'openssl version', and is also included in 'openssl version -a'.
13449 * Allowing defining memory allocation callbacks that will be given
13450 file name and line number information in additional arguments
13451 (a `const char*` and an int). The basic functionality remains, as
13452 well as the original possibility to just replace malloc(),
13453 realloc() and free() by functions that do not know about these
13454 additional arguments. To register and find out the current
13455 settings for extended allocation functions, the following
13456 functions are provided:
13458 CRYPTO_set_mem_ex_functions
13459 CRYPTO_set_locked_mem_ex_functions
13460 CRYPTO_get_mem_ex_functions
13461 CRYPTO_get_locked_mem_ex_functions
13463 These work the same way as CRYPTO_set_mem_functions and friends.
13464 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13465 extended allocation function is enabled.
13466 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13467 a conventional allocation function is enabled.
13469 *Richard Levitte, Bodo Moeller*
13471 * Finish off removing the remaining LHASH function pointer casts.
13472 There should no longer be any prototype-casting required when using
13473 the LHASH abstraction, and any casts that remain are "bugs". See
13474 the callback types and macros at the head of lhash.h for details
13475 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13479 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13480 If /dev/[u]random devices are not available or do not return enough
13481 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13483 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13484 /etc/entropy will be queried once each in this sequence, querying stops
13485 when enough entropy was collected without querying more sockets.
13489 * Change the Unix RAND_poll() variant to be able to poll several
13490 random devices, as specified by DEVRANDOM, until a sufficient amount
13491 of data has been collected. We spend at most 10 ms on each file
13492 (select timeout) and read in non-blocking mode. DEVRANDOM now
13493 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13494 (previously it was just the string "/dev/urandom"), so on typical
13495 platforms the 10 ms delay will never occur.
13496 Also separate out the Unix variant to its own file, rand_unix.c.
13497 For VMS, there's a currently-empty rand_vms.c.
13501 * Move OCSP client related routines to ocsp_cl.c. These
13502 provide utility functions which an application needing
13503 to issue a request to an OCSP responder and analyse the
13504 response will typically need: as opposed to those which an
13505 OCSP responder itself would need which will be added later.
13507 OCSP_request_sign() signs an OCSP request with an API similar
13508 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13509 response. OCSP_response_get1_basic() extracts basic response
13510 from response. OCSP_resp_find_status(): finds and extracts status
13511 information from an OCSP_CERTID structure (which will be created
13512 when the request structure is built). These are built from lower
13513 level functions which work on OCSP_SINGLERESP structures but
13514 won't normally be used unless the application wishes to examine
13515 extensions in the OCSP response for example.
13517 Replace nonce routines with a pair of functions.
13518 OCSP_request_add1_nonce() adds a nonce value and optionally
13519 generates a random value. OCSP_check_nonce() checks the
13520 validity of the nonce in an OCSP response.
13524 * Change function OCSP_request_add() to OCSP_request_add0_id().
13525 This doesn't copy the supplied OCSP_CERTID and avoids the
13526 need to free up the newly created id. Change return type
13527 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13528 This can then be used to add extensions to the request.
13529 Deleted OCSP_request_new(), since most of its functionality
13530 is now in OCSP_REQUEST_new() (and the case insensitive name
13531 clash) apart from the ability to set the request name which
13532 will be added elsewhere.
13536 * Update OCSP API. Remove obsolete extensions argument from
13537 various functions. Extensions are now handled using the new
13538 OCSP extension code. New simple OCSP HTTP function which
13539 can be used to send requests and parse the response.
13543 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13544 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13545 uses the special reorder version of SET OF to sort the attributes
13546 and reorder them to match the encoded order. This resolves a long
13547 standing problem: a verify on a PKCS7 structure just after signing
13548 it used to fail because the attribute order did not match the
13549 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13550 it uses the received order. This is necessary to tolerate some broken
13551 software that does not order SET OF. This is handled by encoding
13552 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13553 to produce the required SET OF.
13557 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13558 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13559 files to get correct declarations of the ASN.1 item variables.
13563 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13564 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13565 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13566 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13567 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13568 ASN1_ITEM and no wrapper functions.
13572 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13573 replace the old function pointer based I/O routines. Change most of
13574 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13578 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13579 lines, recognize more "algorithms" that can be deselected, and make
13580 it complain about algorithm deselection that isn't recognised.
13584 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13585 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13586 to use new functions. Add NO_ASN1_OLD which can be set to remove
13587 some old style ASN1 functions: this can be used to determine if old
13588 code will still work when these eventually go away.
13592 * New extension functions for OCSP structures, these follow the
13593 same conventions as certificates and CRLs.
13597 * New function X509V3_add1_i2d(). This automatically encodes and
13598 adds an extension. Its behaviour can be customised with various
13599 flags to append, replace or delete. Various wrappers added for
13600 certificates and CRLs.
13604 * Fix to avoid calling the underlying ASN1 print routine when
13605 an extension cannot be parsed. Correct a typo in the
13606 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13610 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13611 entries for variables.
13615 * Add functionality to `apps/openssl.c` for detecting locking
13616 problems: As the program is single-threaded, all we have
13617 to do is register a locking callback using an array for
13618 storing which locks are currently held by the program.
13622 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13623 SSL_get_ex_data_X509_STORE_idx(), which is used in
13624 ssl_verify_cert_chain() and thus can be called at any time
13625 during TLS/SSL handshakes so that thread-safety is essential.
13626 Unfortunately, the ex_data design is not at all suited
13627 for multi-threaded use, so it probably should be abolished.
13631 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13633 *Broadcom, tweaked and integrated by Geoff Thorpe*
13635 * Move common extension printing code to new function
13636 X509V3_print_extensions(). Reorganise OCSP print routines and
13637 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13641 * New function X509_signature_print() to remove duplication in some
13646 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13647 set (this was treated exactly the same as SET OF previously). This
13648 is used to reorder the STACK representing the structure to match the
13649 encoding. This will be used to get round a problem where a PKCS7
13650 structure which was signed could not be verified because the STACK
13651 order did not reflect the encoded order.
13655 * Reimplement the OCSP ASN1 module using the new code.
13659 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13660 for its ASN1 operations. The old style function pointers still exist
13661 for now but they will eventually go away.
13665 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13666 completely replaces the old ASN1 functionality with a table driven
13667 encoder and decoder which interprets an ASN1_ITEM structure describing
13668 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13669 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13670 has also been converted to the new form.
13674 * Change BN_mod_exp_recp so that negative moduli are tolerated
13675 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13676 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13677 for negative moduli.
13681 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13682 of not touching the result's sign bit.
13686 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13691 * Changed the LHASH code to use prototypes for callbacks, and created
13692 macros to declare and implement thin (optionally static) functions
13693 that provide type-safety and avoid function pointer casting for the
13694 type-specific callbacks.
13698 * Added Kerberos Cipher Suites to be used with TLS, as written in
13700 *Veers Staats <staatsvr@asc.hpc.mil>,
13701 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13703 * Reformat the FAQ so the different questions and answers can be divided
13704 in sections depending on the subject.
13708 * Have the zlib compression code load ZLIB.DLL dynamically under
13713 * New function BN_mod_sqrt for computing square roots modulo a prime
13714 (using the probabilistic Tonelli-Shanks algorithm unless
13715 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13716 be handled deterministically).
13718 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13720 * Make BN_mod_inverse faster by explicitly handling small quotients
13721 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13722 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13726 * New function BN_kronecker.
13730 * Fix BN_gcd so that it works on negative inputs; the result is
13731 positive unless both parameters are zero.
13732 Previously something reasonably close to an infinite loop was
13733 possible because numbers could be growing instead of shrinking
13734 in the implementation of Euclid's algorithm.
13738 * Fix BN_is_word() and BN_is_one() macros to take into account the
13739 sign of the number in question.
13741 Fix BN_is_word(a,w) to work correctly for w == 0.
13743 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13744 because its test if the absolute value of 'a' equals 'w'.
13745 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13746 it exists mostly for use in the implementations of BN_is_zero(),
13747 BN_is_one(), and BN_is_word().
13751 * New function BN_swap.
13755 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13756 the exponentiation functions are more likely to produce reasonable
13757 results on negative inputs.
13761 * Change BN_mod_mul so that the result is always non-negative.
13762 Previously, it could be negative if one of the factors was negative;
13763 I don't think anyone really wanted that behaviour.
13767 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13768 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13769 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13770 and add new functions:
13779 BN_mod_lshift1_quick
13781 BN_mod_lshift_quick
13783 These functions always generate non-negative results.
13785 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13786 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13788 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13789 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13790 be reduced modulo `m`.
13792 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13795 The following entry accidentally appeared in the CHANGES file
13796 distributed with OpenSSL 0.9.7. The modifications described in
13797 it do *not* apply to OpenSSL 0.9.7.
13799 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13800 was actually never needed) and in BN_mul(). The removal in BN_mul()
13801 required a small change in bn_mul_part_recursive() and the addition
13802 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13803 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13804 bn_sub_words() and bn_add_words() except they take arrays with
13810 * In 'openssl passwd', verify passwords read from the terminal
13811 unless the '-salt' option is used (which usually means that
13812 verification would just waste user's time since the resulting
13813 hash is going to be compared with some given password hash)
13814 or the new '-noverify' option is used.
13816 This is an incompatible change, but it does not affect
13817 non-interactive use of 'openssl passwd' (passwords on the command
13818 line, '-stdin' option, '-in ...' option) and thus should not
13819 cause any problems.
13823 * Remove all references to RSAref, since there's no more need for it.
13827 * Make DSO load along a path given through an environment variable
13828 (SHLIB_PATH) with shl_load().
13832 * Constify the ENGINE code as a result of BIGNUM constification.
13833 Also constify the RSA code and most things related to it. In a
13834 few places, most notable in the depth of the ASN.1 code, ugly
13835 casts back to non-const were required (to be solved at a later
13840 * Make it so the openssl application has all engines loaded by default.
13844 * Constify the BIGNUM routines a little more.
13848 * Add the following functions:
13850 ENGINE_load_cswift()
13852 ENGINE_load_atalla()
13853 ENGINE_load_nuron()
13854 ENGINE_load_builtin_engines()
13856 That way, an application can itself choose if external engines that
13857 are built-in in OpenSSL shall ever be used or not. The benefit is
13858 that applications won't have to be linked with libdl or other dso
13859 libraries unless it's really needed.
13861 Changed 'openssl engine' to load all engines on demand.
13862 Changed the engine header files to avoid the duplication of some
13863 declarations (they differed!).
13867 * 'openssl engine' can now list capabilities.
13871 * Better error reporting in 'openssl engine'.
13875 * Never call load_dh_param(NULL) in s_server.
13879 * Add engine application. It can currently list engines by name and
13880 identity, and test if they are actually available.
13884 * Improve RPM specification file by forcing symbolic linking and making
13885 sure the installed documentation is also owned by root.root.
13887 *Damien Miller <djm@mindrot.org>*
13889 * Give the OpenSSL applications more possibilities to make use of
13890 keys (public as well as private) handled by engines.
13894 * Add OCSP code that comes from CertCo.
13898 * Add VMS support for the Rijndael code.
13902 * Added untested support for Nuron crypto accelerator.
13906 * Add support for external cryptographic devices. This code was
13907 previously distributed separately as the "engine" branch.
13909 *Geoff Thorpe, Richard Levitte*
13911 * Rework the filename-translation in the DSO code. It is now possible to
13912 have far greater control over how a "name" is turned into a filename
13913 depending on the operating environment and any oddities about the
13914 different shared library filenames on each system.
13918 * Support threads on FreeBSD-elf in Configure.
13922 * Fix for SHA1 assembly problem with MASM: it produces
13923 warnings about corrupt line number information when assembling
13924 with debugging information. This is caused by the overlapping
13927 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13930 NCONF_get_number() has no error checking at all. As a replacement,
13931 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13932 promoted strongly. The old NCONF_get_number is kept around for
13933 binary backward compatibility.
13934 Make it possible for methods to load from something other than a BIO,
13935 by providing a function pointer that is given a name instead of a BIO.
13936 For example, this could be used to load configuration data from an
13941 * Fix for non blocking accept BIOs. Added new I/O special reason
13942 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13943 with non blocking I/O was not possible because no retry code was
13944 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13949 * Added the beginnings of Rijndael support.
13953 * Fix for bug in DirectoryString mask setting. Add support for
13954 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13955 to allow certificate printing to more controllable, additional
13956 'certopt' option to 'x509' to allow new printing options to be
13961 * Clean old EAY MD5 hack from e_os.h.
13965 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13967 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13968 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13970 *Joe Orton, Steve Henson*
13972 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13974 * Fix additional bug revealed by the NISCC test suite:
13976 Stop bug triggering large recursion when presented with
13977 certain ASN.1 tags ([CVE-2003-0851])
13981 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13983 * Fix various bugs revealed by running the NISCC test suite:
13985 Stop out of bounds reads in the ASN1 code when presented with
13986 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13988 If verify callback ignores invalid public key errors don't try to check
13989 certificate signature with the NULL public key.
13993 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13994 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13999 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14000 extra data after the compression methods not only for TLS 1.0
14001 but also for SSL 3.0 (as required by the specification).
14003 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14005 * Change X509_certificate_type() to mark the key as exported/exportable
14006 when it's 512 *bits* long, not 512 bytes.
14010 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14012 * Countermeasure against the Klima-Pokorny-Rosa extension of
14013 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14014 a protocol version number mismatch like a decryption error
14015 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14019 * Turn on RSA blinding by default in the default implementation
14020 to avoid a timing attack. Applications that don't want it can call
14021 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14022 They would be ill-advised to do so in most cases.
14024 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14026 * Change RSA blinding code so that it works when the PRNG is not
14027 seeded (in this case, the secret RSA exponent is abused as
14028 an unpredictable seed -- if it is not unpredictable, there
14029 is no point in blinding anyway). Make RSA blinding thread-safe
14030 by remembering the creator's thread ID in rsa->blinding and
14031 having all other threads use local one-time blinding factors
14032 (this requires more computation than sharing rsa->blinding, but
14033 avoids excessive locking; and if an RSA object is not shared
14034 between threads, blinding will still be very fast).
14038 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14040 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14041 via timing by performing a MAC computation even if incorrect
14042 block cipher padding has been found. This is a countermeasure
14043 against active attacks where the attacker has to distinguish
14044 between bad padding and a MAC verification error. ([CVE-2003-0078])
14046 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14047 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14048 Martin Vuagnoux (EPFL, Ilion)*
14050 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14052 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14053 memory from its contents. This is done with a counter that will
14054 place alternating values in each byte. This can be used to solve
14055 two issues: 1) the removal of calls to memset() by highly optimizing
14056 compilers, and 2) cleansing with other values than 0, since those can
14057 be read through on certain media, for example a swap space on disk.
14061 * Bugfix: client side session caching did not work with external caching,
14062 because the session->cipher setting was not restored when reloading
14063 from the external cache. This problem was masked, when
14064 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14065 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14069 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14070 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14072 *Zeev Lieber <zeev-l@yahoo.com>*
14074 * Undo an undocumented change introduced in 0.9.6e which caused
14075 repeated calls to OpenSSL_add_all_ciphers() and
14076 OpenSSL_add_all_digests() to be ignored, even after calling
14081 * Change the default configuration reader to deal with last line not
14082 being properly terminated.
14086 * Change X509_NAME_cmp() so it applies the special rules on handling
14087 DN values that are of type PrintableString, as well as RDNs of type
14088 emailAddress where the value has the type ia5String.
14090 *stefank@valicert.com via Richard Levitte*
14092 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14093 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14094 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14095 the bitwise-OR of the two for use by the majority of applications
14096 wanting this behaviour, and update the docs. The documented
14097 behaviour and actual behaviour were inconsistent and had been
14098 changing anyway, so this is more a bug-fix than a behavioural
14101 *Geoff Thorpe, diagnosed by Nadav Har'El*
14103 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14104 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14108 * Fix initialization code race conditions in
14109 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14110 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14111 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14112 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14113 ssl2_get_cipher_by_char(),
14114 ssl3_get_cipher_by_char().
14116 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14118 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14119 the cached sessions are flushed, as the remove_cb() might use ex_data
14120 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14121 (see [openssl.org #212]).
14123 *Geoff Thorpe, Lutz Jaenicke*
14125 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14126 length, instead of the encoding length to d2i_ASN1_OBJECT.
14130 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14132 * [In 0.9.6g-engine release:]
14133 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14135 *Lynn Gazis <lgazis@rainbow.com>*
14137 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14139 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14140 and get fix the header length calculation.
14141 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14142 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14144 * Use proper error handling instead of 'assertions' in buffer
14145 overflow checks added in 0.9.6e. This prevents DoS (the
14146 assertions could call abort()).
14148 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14150 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14152 * Add various sanity checks to asn1_get_length() to reject
14153 the ASN1 length bytes if they exceed sizeof(long), will appear
14154 negative or the content length exceeds the length of the
14157 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14159 * Fix cipher selection routines: ciphers without encryption had no flags
14160 for the cipher strength set and where therefore not handled correctly
14161 by the selection routines (PR #130).
14165 * Fix EVP_dsa_sha macro.
14170 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14171 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14172 that was added in OpenSSL 0.9.6d.
14174 As the countermeasure turned out to be incompatible with some
14175 broken SSL implementations, the new option is part of SSL_OP_ALL.
14176 SSL_OP_ALL is usually employed when compatibility with weird SSL
14177 implementations is desired (e.g. '-bugs' option to 's_client' and
14178 's_server'), so the new option is automatically set in many
14183 * Changes in security patch:
14185 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14186 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14187 Air Force Materiel Command, USAF, under agreement number
14190 * Add various sanity checks to asn1_get_length() to reject
14191 the ASN1 length bytes if they exceed sizeof(long), will appear
14192 negative or the content length exceeds the length of the
14193 supplied buffer. ([CVE-2002-0659])
14195 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14197 * Assertions for various potential buffer overflows, not known to
14198 happen in practice.
14200 *Ben Laurie (CHATS)*
14202 * Various temporary buffers to hold ASCII versions of integers were
14203 too small for 64 bit platforms. ([CVE-2002-0655])
14204 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14206 * Remote buffer overflow in SSL3 protocol - an attacker could
14207 supply an oversized session ID to a client. ([CVE-2002-0656])
14209 *Ben Laurie (CHATS)*
14211 * Remote buffer overflow in SSL2 protocol - an attacker could
14212 supply an oversized client master key. ([CVE-2002-0656])
14214 *Ben Laurie (CHATS)*
14216 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14218 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14219 encoded as NULL) with id-dsa-with-sha1.
14221 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14223 * Check various `X509_...()` return values in `apps/req.c`.
14225 *Nils Larsch <nla@trustcenter.de>*
14227 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14228 an end-of-file condition would erroneously be flagged, when the CRLF
14229 was just at the end of a processed block. The bug was discovered when
14230 processing data through a buffering memory BIO handing the data to a
14231 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14232 <ptsekov@syntrex.com> and Nedelcho Stanev.
14236 * Implement a countermeasure against a vulnerability recently found
14237 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14238 before application data chunks to avoid the use of known IVs
14239 with data potentially chosen by the attacker.
14243 * Fix length checks in ssl3_get_client_hello().
14247 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14248 to prevent ssl3_read_internal() from incorrectly assuming that
14249 ssl3_read_bytes() found application data while handshake
14250 processing was enabled when in fact s->s3->in_read_app_data was
14251 merely automatically cleared during the initial handshake.
14253 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14255 * Fix object definitions for Private and Enterprise: they were not
14256 recognized in their shortname (=lowercase) representation. Extend
14257 obj_dat.pl to issue an error when using undefined keywords instead
14258 of silently ignoring the problem (Svenning Sorensen
14259 <sss@sss.dnsalias.net>).
14263 * Fix DH_generate_parameters() so that it works for 'non-standard'
14264 generators, i.e. generators other than 2 and 5. (Previously, the
14265 code did not properly initialise the 'add' and 'rem' values to
14266 BN_generate_prime().)
14268 In the new general case, we do not insist that 'generator' is
14269 actually a primitive root: This requirement is rather pointless;
14270 a generator of the order-q subgroup is just as good, if not
14275 * Map new X509 verification errors to alerts. Discovered and submitted by
14276 Tom Wu <tom@arcot.com>.
14280 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14281 returning non-zero before the data has been completely received
14282 when using non-blocking I/O.
14284 *Bodo Moeller; problem pointed out by John Hughes*
14286 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14288 *Ben Laurie, Lutz Jaenicke*
14290 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14291 Yoram Zahavi <YoramZ@gilian.com>).
14295 * Add information about CygWin 1.3 and on, and preserve proper
14296 configuration for the versions before that.
14298 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14300 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14301 check whether we deal with a copy of a session and do not delete from
14302 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14303 <izhar@checkpoint.com>.
14307 * Do not store session data into the internal session cache, if it
14308 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14309 flag is set). Proposed by Aslam <aslam@funk.com>.
14313 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14318 * [In 0.9.6d-engine release:]
14319 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14321 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14323 * Add the configuration target linux-s390x.
14325 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14327 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14328 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14329 variable as an indication that a ClientHello message has been
14330 received. As the flag value will be lost between multiple
14331 invocations of ssl3_accept when using non-blocking I/O, the
14332 function may not be aware that a handshake has actually taken
14333 place, thus preventing a new session from being added to the
14336 To avoid this problem, we now set s->new_session to 2 instead of
14337 using a local variable.
14339 *Lutz Jaenicke, Bodo Moeller*
14341 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14342 if the SSL_R_LENGTH_MISMATCH error is detected.
14344 *Geoff Thorpe, Bodo Moeller*
14346 * New 'shared_ldflag' column in Configure platform table.
14350 * Fix EVP_CIPHER_mode macro.
14352 *"Dan S. Camper" <dan@bti.net>*
14354 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14355 type, we must throw them away by setting rr->length to 0.
14357 *D P Chang <dpc@qualys.com>*
14359 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14361 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14362 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14363 worked incorrectly for those cases where range = `10..._2` and
14364 `3*range` is two bits longer than range.)
14368 * Only add signing time to PKCS7 structures if it is not already
14373 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14374 OBJ_ld_ce should be OBJ_id_ce.
14375 Also some ip-pda OIDs in crypto/objects/objects.txt were
14376 incorrect (cf. RFC 3039).
14378 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14380 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14381 returns early because it has nothing to do.
14383 *Andy Schneider <andy.schneider@bjss.co.uk>*
14385 * [In 0.9.6c-engine release:]
14386 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14388 *Andy Schneider <andy.schneider@bjss.co.uk>*
14390 * [In 0.9.6c-engine release:]
14391 Add support for Cryptographic Appliance's keyserver technology.
14392 (Use engine 'keyclient')
14394 *Cryptographic Appliances and Geoff Thorpe*
14396 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14397 is called via tools/c89.sh because arguments have to be
14398 rearranged (all '-L' options must appear before the first object
14401 *Richard Shapiro <rshapiro@abinitio.com>*
14403 * [In 0.9.6c-engine release:]
14404 Add support for Broadcom crypto accelerator cards, backported
14407 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14409 * [In 0.9.6c-engine release:]
14410 Add support for SureWare crypto accelerator cards from
14411 Baltimore Technologies. (Use engine 'sureware')
14413 *Baltimore Technologies and Mark Cox*
14415 * [In 0.9.6c-engine release:]
14416 Add support for crypto accelerator cards from Accelerated
14417 Encryption Processing, www.aep.ie. (Use engine 'aep')
14419 *AEP Inc. and Mark Cox*
14421 * Add a configuration entry for gcc on UnixWare.
14423 *Gary Benson <gbenson@redhat.com>*
14425 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14426 messages are stored in a single piece (fixed-length part and
14427 variable-length part combined) and fix various bugs found on the way.
14431 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14432 instead. BIO_gethostbyname() does not know what timeouts are
14433 appropriate, so entries would stay in cache even when they have
14435 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14437 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14438 faced with a pathologically small ClientHello fragment that does
14439 not contain client_version: Instead of aborting with an error,
14440 simply choose the highest available protocol version (i.e.,
14441 TLS 1.0 unless it is disabled). In practice, ClientHello
14442 messages are never sent like this, but this change gives us
14443 strictly correct behaviour at least for TLS.
14447 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14448 never resets s->method to s->ctx->method when called from within
14449 one of the SSL handshake functions.
14451 *Bodo Moeller; problem pointed out by Niko Baric*
14453 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14454 (sent using the client's version number) if client_version is
14455 smaller than the protocol version in use. Also change
14456 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14457 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14458 the client will at least see that alert.
14462 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14467 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14468 client receives HelloRequest while in a handshake.
14470 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14472 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14473 should end in 'break', not 'goto end' which circumvents various
14474 cleanups done in state SSL_ST_OK. But session related stuff
14475 must be disabled for SSL_ST_OK in the case that we just sent a
14478 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14479 before just sending a HelloRequest.
14481 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14483 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14484 reveal whether illegal block cipher padding was found or a MAC
14485 verification error occurred. (Neither SSLerr() codes nor alerts
14486 are directly visible to potential attackers, but the information
14487 may leak via logfiles.)
14489 Similar changes are not required for the SSL 2.0 implementation
14490 because the number of padding bytes is sent in clear for SSL 2.0,
14491 and the extra bytes are just ignored. However ssl/s2_pkt.c
14492 failed to verify that the purported number of padding bytes is in
14497 * Add OpenUNIX-8 support including shared libraries
14498 (Boyd Lynn Gerber <gerberb@zenez.com>).
14502 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14503 'wristwatch attack' using huge encoding parameters (cf.
14504 James H. Manger's CRYPTO 2001 paper). Note that the
14505 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14506 encoding parameters and hence was not vulnerable.
14510 * BN_sqr() bug fix.
14512 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14514 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14515 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14516 followed by modular reduction.
14518 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14520 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14521 equivalent based on BN_pseudo_rand() instead of BN_rand().
14525 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14526 This function was broken, as the check for a new client hello message
14527 to handle SGC did not allow these large messages.
14528 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14532 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14536 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14537 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14541 * Rework the configuration and shared library support for Tru64 Unix.
14542 The configuration part makes use of modern compiler features and
14543 still retains old compiler behavior for those that run older versions
14544 of the OS. The shared library support part includes a variant that
14545 uses the RPATH feature, and is available through the special
14546 configuration target "alpha-cc-rpath", which will never be selected
14549 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14551 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14552 with the same message size as in ssl3_get_certificate_request().
14553 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14554 messages might inadvertently be reject as too long.
14556 *Petr Lampa <lampa@fee.vutbr.cz>*
14558 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14562 * Modified SSL library such that the verify_callback that has been set
14563 specifically for an SSL object with SSL_set_verify() is actually being
14564 used. Before the change, a verify_callback set with this function was
14565 ignored and the verify_callback() set in the SSL_CTX at the time of
14566 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14567 to allow the necessary settings.
14571 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14572 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14573 done automatically (in contradiction to the requirements of the C
14574 standard). This made problems when used from OpenSSH.
14578 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14579 dh->length and always used
14581 BN_rand_range(priv_key, dh->p).
14583 BN_rand_range() is not necessary for Diffie-Hellman, and this
14584 specific range makes Diffie-Hellman unnecessarily inefficient if
14585 dh->length (recommended exponent length) is much smaller than the
14586 length of dh->p. We could use BN_rand_range() if the order of
14587 the subgroup was stored in the DH structure, but we only have
14592 BN_rand(priv_key, l, ...)
14594 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14601 RSA_eay_public_encrypt
14602 RSA_eay_private_decrypt
14603 RSA_eay_private_encrypt (signing)
14604 RSA_eay_public_decrypt (signature verification)
14606 (default implementations for RSA_public_encrypt,
14607 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14608 always reject numbers >= n.
14612 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14613 to synchronize access to 'locking_thread'. This is necessary on
14614 systems where access to 'locking_thread' (an 'unsigned long'
14615 variable) is not atomic.
14619 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14620 *before* setting the 'crypto_lock_rand' flag. The previous code had
14621 a race condition if 0 is a valid thread ID.
14623 *Travis Vitek <vitek@roguewave.com>*
14625 * Add support for shared libraries under Irix.
14627 *Albert Chin-A-Young <china@thewrittenword.com>*
14629 * Add configuration option to build on Linux on both big-endian and
14630 little-endian MIPS.
14632 *Ralf Baechle <ralf@uni-koblenz.de>*
14634 * Add the possibility to create shared libraries on HP-UX.
14638 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14640 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14641 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14642 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14643 PRNG state recovery was possible based on the output of
14644 one PRNG request appropriately sized to gain knowledge on
14645 'md' followed by enough consecutive 1-byte PRNG requests
14646 to traverse all of 'state'.
14648 1. When updating 'md_local' (the current thread's copy of 'md')
14649 during PRNG output generation, hash all of the previous
14650 'md_local' value, not just the half used for PRNG output.
14652 2. Make the number of bytes from 'state' included into the hash
14653 independent from the number of PRNG bytes requested.
14655 The first measure alone would be sufficient to avoid
14656 Markku-Juhani's attack. (Actually it had never occurred
14657 to me that the half of 'md_local' used for chaining was the
14658 half from which PRNG output bytes were taken -- I had always
14659 assumed that the secret half would be used.) The second
14660 measure makes sure that additional data from 'state' is never
14661 mixed into 'md_local' in small portions; this heuristically
14662 further strengthens the PRNG.
14666 * Fix crypto/bn/asm/mips3.s.
14670 * When only the key is given to "enc", the IV is undefined. Print out
14671 an error message in this case.
14675 * Handle special case when X509_NAME is empty in X509 printing routines.
14679 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14680 positive and less than q.
14684 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14685 used: it isn't thread safe and the add_lock_callback should handle
14688 *Paul Rose <Paul.Rose@bridge.com>*
14690 * Verify that incoming data obeys the block size in
14691 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14697 *Ulf Möller, Bodo Möller*
14699 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14700 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14701 when fixing the server behaviour for backwards-compatible 'client
14702 hello' messages. (Note that the attack is impractical against
14703 SSL 3.0 and TLS 1.0 anyway because length and version checking
14704 means that the probability of guessing a valid ciphertext is
14705 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14708 Before 0.9.5, the countermeasure (hide the error by generating a
14709 random 'decryption result') did not work properly because
14710 ERR_clear_error() was missing, meaning that SSL_get_error() would
14711 detect the supposedly ignored error.
14713 Both problems are now fixed.
14717 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14718 (previously it was 1024).
14722 * Fix for compatibility mode trust settings: ignore trust settings
14723 unless some valid trust or reject settings are present.
14727 * Fix for blowfish EVP: its a variable length cipher.
14731 * Fix various bugs related to DSA S/MIME verification. Handle missing
14732 parameters in DSA public key structures and return an error in the
14733 DSA routines if parameters are absent.
14737 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14738 in the current directory if neither $RANDFILE nor $HOME was set.
14739 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14740 caused some confusion to Windows users who haven't defined $HOME.
14741 Thus RAND_file_name() is changed again: e_os.h can define a
14742 DEFAULT_HOME, which will be used if $HOME is not set.
14743 For Windows, we use "C:"; on other platforms, we still require
14744 environment variables.
14746 * Move 'if (!initialized) RAND_poll()' into regions protected by
14747 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14748 having multiple threads call RAND_poll() concurrently.
14752 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14753 combination of a flag and a thread ID variable.
14754 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14755 flag), *other* threads can enter ssleay_add_bytes without obeying
14756 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14757 that they do not hold after the first thread unsets add_do_not_lock).
14761 * Change bctest again: '-x' expressions are not available in all
14762 versions of 'test'.
14766 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14768 * Fix a couple of memory leaks in PKCS7_dataDecode()
14770 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14772 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14773 the default extension for executables, if any. Also, make the perl
14774 scripts that use symlink() to test if it really exists and use "cp"
14775 if it doesn't. All this made OpenSSL compilable and installable in
14780 * Fix for asn1_GetSequence() for indefinite length constructed data.
14781 If SEQUENCE is length is indefinite just set c->slen to the total
14782 amount of data available.
14784 *Steve Henson, reported by shige@FreeBSD.org*
14786 *This change does not apply to 0.9.7.*
14788 * Change bctest to avoid here-documents inside command substitution
14789 (workaround for FreeBSD /bin/sh bug).
14790 For compatibility with Ultrix, avoid shell functions (introduced
14791 in the bctest version that searches along $PATH).
14795 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14796 with des_encrypt() defined on some operating systems, like Solaris
14801 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14802 On the Importance of Eliminating Errors in Cryptographic
14803 Computations, J. Cryptology 14 (2001) 2, 101-119,
14804 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14808 * MIPS assembler BIGNUM division bug fix.
14812 * Disabled incorrect Alpha assembler code.
14816 * Fix PKCS#7 decode routines so they correctly update the length
14817 after reading an EOC for the EXPLICIT tag.
14821 *This change does not apply to 0.9.7.*
14823 * Fix bug in PKCS#12 key generation routines. This was triggered
14824 if a 3DES key was generated with a 0 initial byte. Include
14825 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14826 (but broken) behaviour.
14830 * Enhance bctest to search for a working bc along $PATH and print
14833 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14835 * Fix memory leaks in err.c: free err_data string if necessary;
14836 don't write to the wrong index in ERR_set_error_data.
14840 * Implement ssl23_peek (analogous to ssl23_read), which previously
14845 * Replace rdtsc with `_emit` statements for VC++ version 5.
14847 *Jeremy Cooper <jeremy@baymoo.org>*
14849 * Make it possible to reuse SSLv2 sessions.
14853 * In copy_email() check for >= 0 as a return value for
14854 X509_NAME_get_index_by_NID() since 0 is a valid index.
14856 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14858 * Avoid coredump with unsupported or invalid public keys by checking if
14859 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14860 PKCS7_verify() fails with non detached data.
14864 * Don't use getenv in library functions when run as setuid/setgid.
14865 New function OPENSSL_issetugid().
14869 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14870 due to incorrect handling of multi-threading:
14872 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14874 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14876 3. Count how many times MemCheck_off() has been called so that
14877 nested use can be treated correctly. This also avoids
14878 inband-signalling in the previous code (which relied on the
14879 assumption that thread ID 0 is impossible).
14883 * Add "-rand" option also to s_client and s_server.
14887 * Fix CPU detection on Irix 6.x.
14888 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14889 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14891 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14896 *This change does not apply to 0.9.7.*
14898 * Use the cached encoding of an X509_NAME structure rather than
14899 copying it. This is apparently the reason for the libsafe "errors"
14900 but the code is actually correct.
14904 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14905 Bleichenbacher's DSA attack.
14906 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14907 to be set and top=0 forces the highest bit to be set; top=-1 is new
14908 and leaves the highest bit random.
14910 *Ulf Moeller, Bodo Moeller*
14912 * In the `NCONF_...`-based implementations for `CONF_...` queries
14913 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14914 a temporary CONF structure with the data component set to NULL
14915 (which gives segmentation faults in lh_retrieve).
14916 Instead, use NULL for the CONF pointer in CONF_get_string and
14917 CONF_get_number (which may use environment variables) and directly
14918 return NULL from CONF_get_section.
14922 * Fix potential buffer overrun for EBCDIC.
14926 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14927 keyUsage if basicConstraints absent for a CA.
14931 * Make SMIME_write_PKCS7() write mail header values with a format that
14932 is more generally accepted (no spaces before the semicolon), since
14933 some programs can't parse those values properly otherwise. Also make
14934 sure BIO's that break lines after each write do not create invalid
14939 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14940 macros previously used would not encode an empty SEQUENCE OF
14941 and break the signature.
14945 *This change does not apply to 0.9.7.*
14947 * Zero the premaster secret after deriving the master secret in
14952 * Add some EVP_add_digest_alias registrations (as found in
14953 OpenSSL_add_all_digests()) to SSL_library_init()
14954 aka OpenSSL_add_ssl_algorithms(). This provides improved
14955 compatibility with peers using X.509 certificates
14956 with unconventional AlgorithmIdentifier OIDs.
14960 * Fix for Irix with NO_ASM.
14962 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14964 * ./config script fixes.
14966 *Ulf Moeller, Richard Levitte*
14968 * Fix 'openssl passwd -1'.
14972 * Change PKCS12_key_gen_asc() so it can cope with non null
14973 terminated strings whose length is passed in the passlen
14974 parameter, for example from PEM callbacks. This was done
14975 by adding an extra length parameter to asc2uni().
14977 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14979 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14980 call failed, free the DSA structure.
14984 * Fix to uni2asc() to cope with zero length Unicode strings.
14985 These are present in some PKCS#12 files.
14989 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14990 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14991 when writing a 32767 byte record.
14993 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14995 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14996 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14998 (RSA objects have a reference count access to which is protected
14999 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15000 so they are meant to be shared between threads.)
15001 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15002 "Reddie, Steven" <Steven.Reddie@ca.com>*
15004 * Fix a deadlock in CRYPTO_mem_leaks().
15008 * Use better test patterns in bntest.
15012 * rand_win.c fix for Borland C.
15016 * BN_rshift bugfix for n == 0.
15020 * Add a 'bctest' script that checks for some known 'bc' bugs
15021 so that 'make test' does not abort just because 'bc' is broken.
15025 * Store verify_result within SSL_SESSION also for client side to
15026 avoid potential security hole. (Re-used sessions on the client side
15027 always resulted in verify_result==X509_V_OK, not using the original
15028 result of the server certificate verification.)
15032 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15033 SSL3_RT_APPLICATION_DATA, return 0.
15034 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15039 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15040 releases, have been re-implemented by renaming the previous
15041 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15042 and ssl3_read_internal, respectively, and adding 'peek' parameters
15043 to them. The new ssl[23]_{read,peek} functions are calls to
15044 ssl[23]_read_internal with the 'peek' flag set appropriately.
15045 A 'peek' parameter has also been added to ssl3_read_bytes, which
15046 does the actual work for ssl3_read_internal.
15050 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15051 the method-specific "init()" handler. Also clean up ex_data after
15052 calling the method-specific "finish()" handler. Previously, this was
15053 happening the other way round.
15057 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15058 The previous value, 12, was not always sufficient for BN_mod_exp().
15062 * Make sure that shared libraries get the internal name engine with
15063 the full version number and not just 0. This should mark the
15064 shared libraries as not backward compatible. Of course, this should
15065 be changed again when we can guarantee backward binary compatibility.
15069 * Fix typo in get_cert_by_subject() in by_dir.c
15071 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15073 * Rework the system to generate shared libraries:
15075 - Make note of the expected extension for the shared libraries and
15076 if there is a need for symbolic links from for example libcrypto.so.0
15077 to libcrypto.so.0.9.7. There is extended info in Configure for
15080 - Make as few rebuilds of the shared libraries as possible.
15082 - Still avoid linking the OpenSSL programs with the shared libraries.
15084 - When installing, install the shared libraries separately from the
15089 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15091 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15092 and not in SSL_clear because the latter is also used by the
15093 accept/connect functions; previously, the settings made by
15094 SSL_set_read_ahead would be lost during the handshake.
15096 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15098 * Correct util/mkdef.pl to be selective about disabled algorithms.
15099 Previously, it would create entries for disabled algorithms no
15104 * Added several new manual pages for SSL_* function.
15108 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15110 * In ssl23_get_client_hello, generate an error message when faced
15111 with an initial SSL 3.0/TLS record that is too small to contain the
15112 first two bytes of the ClientHello message, i.e. client_version.
15113 (Note that this is a pathologic case that probably has never happened
15114 in real life.) The previous approach was to use the version number
15115 from the record header as a substitute; but our protocol choice
15116 should not depend on that one because it is not authenticated
15117 by the Finished messages.
15121 * More robust randomness gathering functions for Windows.
15123 *Jeffrey Altman <jaltman@columbia.edu>*
15125 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15126 not set then we don't setup the error code for issuer check errors
15127 to avoid possibly overwriting other errors which the callback does
15128 handle. If an application does set the flag then we assume it knows
15129 what it is doing and can handle the new informational codes
15134 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15135 a general "ANY" type, as such it should be able to decode anything
15136 including tagged types. However it didn't check the class so it would
15137 wrongly interpret tagged types in the same way as their universal
15138 counterpart and unknown types were just rejected. Changed so that the
15139 tagged and unknown types are handled in the same way as a SEQUENCE:
15140 that is the encoding is stored intact. There is also a new type
15141 "V_ASN1_OTHER" which is used when the class is not universal, in this
15142 case we have no idea what the actual type is so we just lump them all
15147 * On VMS, stdout may very well lead to a file that is written to
15148 in a record-oriented fashion. That means that every write() will
15149 write a separate record, which will be read separately by the
15150 programs trying to read from it. This can be very confusing.
15152 The solution is to put a BIO filter in the way that will buffer
15153 text until a linefeed is reached, and then write everything a
15154 line at a time, so every record written will be an actual line,
15155 not chunks of lines and not (usually doesn't happen, but I've
15156 seen it once) several lines in one record. BIO_f_linebuffer() is
15159 Currently, it's a VMS-only method, because that's where it has
15160 been tested well enough.
15164 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15165 it can return incorrect results.
15166 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15167 but it was in 0.9.6-beta[12].)
15171 * Disable the check for content being present when verifying detached
15172 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15173 include zero length content when signing messages.
15177 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15178 BIO_ctrl (for BIO pairs).
15182 * Add DSO method for VMS.
15186 * Bug fix: Montgomery multiplication could produce results with the
15191 * Add RPM specification openssl.spec and modify it to build three
15192 packages. The default package contains applications, application
15193 documentation and run-time libraries. The devel package contains
15194 include files, static libraries and function documentation. The
15195 doc package contains the contents of the doc directory. The original
15196 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15200 * Add a large number of documentation files for many SSL routines.
15202 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15204 * Add a configuration entry for Sony News 4.
15206 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15208 * Don't set the two most significant bits to one when generating a
15209 random number < q in the DSA library.
15213 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15214 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15215 the underlying transport is blocking) if a handshake took place.
15216 (The default behaviour is needed by applications such as s_client
15217 and s_server that use select() to determine when to use SSL_read;
15218 but for applications that know in advance when to expect data, it
15219 just makes things more complicated.)
15223 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15228 * Add a few more EBCDIC conditionals that make `req` and `x509`
15229 work better on such systems.
15231 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15233 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15234 Update PKCS12_parse() so it copies the friendlyName and the
15235 keyid to the certificates aux info.
15239 * Fix bug in PKCS7_verify() which caused an infinite loop
15240 if there was more than one signature.
15242 *Sven Uszpelkat <su@celocom.de>*
15244 * Major change in util/mkdef.pl to include extra information
15245 about each symbol, as well as presenting variables as well
15246 as functions. This change means that there's n more need
15247 to rebuild the .num files when some algorithms are excluded.
15251 * Allow the verify time to be set by an application,
15252 rather than always using the current time.
15256 * Phase 2 verify code reorganisation. The certificate
15257 verify code now looks up an issuer certificate by a
15258 number of criteria: subject name, authority key id
15259 and key usage. It also verifies self signed certificates
15260 by the same criteria. The main comparison function is
15261 X509_check_issued() which performs these checks.
15263 Lot of changes were necessary in order to support this
15264 without completely rewriting the lookup code.
15266 Authority and subject key identifier are now cached.
15268 The LHASH 'certs' is X509_STORE has now been replaced
15269 by a STACK_OF(X509_OBJECT). This is mainly because an
15270 LHASH can't store or retrieve multiple objects with
15271 the same hash value.
15273 As a result various functions (which were all internal
15274 use only) have changed to handle the new X509_STORE
15275 structure. This will break anything that messed round
15276 with X509_STORE internally.
15278 The functions X509_STORE_add_cert() now checks for an
15279 exact match, rather than just subject name.
15281 The X509_STORE API doesn't directly support the retrieval
15282 of multiple certificates matching a given criteria, however
15283 this can be worked round by performing a lookup first
15284 (which will fill the cache with candidate certificates)
15285 and then examining the cache for matches. This is probably
15286 the best we can do without throwing out X509_LOOKUP
15287 entirely (maybe later...).
15289 The X509_VERIFY_CTX structure has been enhanced considerably.
15291 All certificate lookup operations now go via a get_issuer()
15292 callback. Although this currently uses an X509_STORE it
15293 can be replaced by custom lookups. This is a simple way
15294 to bypass the X509_STORE hackery necessary to make this
15295 work and makes it possible to use more efficient techniques
15296 in future. A very simple version which uses a simple
15297 STACK for its trusted certificate store is also provided
15298 using X509_STORE_CTX_trusted_stack().
15300 The verify_cb() and verify() callbacks now have equivalents
15301 in the X509_STORE_CTX structure.
15303 X509_STORE_CTX also has a 'flags' field which can be used
15304 to customise the verify behaviour.
15308 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15309 excludes S/MIME capabilities.
15313 * When a certificate request is read in keep a copy of the
15314 original encoding of the signed data and use it when outputting
15315 again. Signatures then use the original encoding rather than
15316 a decoded, encoded version which may cause problems if the
15317 request is improperly encoded.
15321 * For consistency with other BIO_puts implementations, call
15322 buffer_write(b, ...) directly in buffer_puts instead of calling
15325 In BIO_puts, increment b->num_write as in BIO_write.
15327 *Peter.Sylvester@EdelWeb.fr*
15329 * Fix BN_mul_word for the case where the word is 0. (We have to use
15330 BN_zero, we may not return a BIGNUM with an array consisting of
15331 words set to zero.)
15335 * Avoid calling abort() from within the library when problems are
15336 detected, except if preprocessor symbols have been defined
15337 (such as REF_CHECK, BN_DEBUG etc.).
15341 * New openssl application 'rsautl'. This utility can be
15342 used for low level RSA operations. DER public key
15343 BIO/fp routines also added.
15347 * New Configure entry and patches for compiling on QNX 4.
15349 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15351 * A demo state-machine implementation was sponsored by
15352 Nuron (<http://www.nuron.com/>) and is now available in
15353 demos/state_machine.
15357 * New options added to the 'dgst' utility for signature
15358 generation and verification.
15362 * Unrecognized PKCS#7 content types are now handled via a
15363 catch all ASN1_TYPE structure. This allows unsupported
15364 types to be stored as a "blob" and an application can
15365 encode and decode it manually.
15369 * Fix various signed/unsigned issues to make a_strex.c
15370 compile under VC++.
15372 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15374 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15375 length if passed a buffer. ASN1_INTEGER_to_BN failed
15376 if passed a NULL BN and its argument was negative.
15378 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15380 * Modification to PKCS#7 encoding routines to output definite
15381 length encoding. Since currently the whole structures are in
15382 memory there's not real point in using indefinite length
15383 constructed encoding. However if OpenSSL is compiled with
15384 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15388 * Added BIO_vprintf() and BIO_vsnprintf().
15392 * Added more prefixes to parse for in the strings written
15393 through a logging bio, to cover all the levels that are available
15394 through syslog. The prefixes are now:
15396 PANIC, EMERG, EMR => LOG_EMERG
15397 ALERT, ALR => LOG_ALERT
15398 CRIT, CRI => LOG_CRIT
15399 ERROR, ERR => LOG_ERR
15400 WARNING, WARN, WAR => LOG_WARNING
15401 NOTICE, NOTE, NOT => LOG_NOTICE
15402 INFO, INF => LOG_INFO
15403 DEBUG, DBG => LOG_DEBUG
15405 and as before, if none of those prefixes are present at the
15406 beginning of the string, LOG_ERR is chosen.
15408 On Win32, the `LOG_*` levels are mapped according to this:
15410 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15411 LOG_WARNING => EVENTLOG_WARNING_TYPE
15412 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15416 * Made it possible to reconfigure with just the configuration
15417 argument "reconf" or "reconfigure". The command line arguments
15418 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15419 and are retrieved from there when reconfiguring.
15425 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15427 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15431 * The obj_dat.pl script was messing up the sorting of object
15432 names. The reason was that it compared the quoted version
15433 of strings as a result "OCSP" > "OCSP Signing" because
15434 " > SPACE. Changed script to store unquoted versions of
15435 names and add quotes on output. It was also omitting some
15436 names from the lookup table if they were given a default
15437 value (that is if SN is missing it is given the same
15438 value as LN and vice versa), these are now added on the
15439 grounds that if an object has a name we should be able to
15440 look it up. Finally added warning output when duplicate
15441 short or long names are found.
15445 * Changes needed for Tandem NSK.
15447 *Scott Uroff <scott@xypro.com>*
15449 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15450 RSA_padding_check_SSLv23(), special padding was never detected
15451 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15452 version rollback attacks was not effective.
15454 In s23_clnt.c, don't use special rollback-attack detection padding
15455 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15456 client; similarly, in s23_srvr.c, don't do the rollback check if
15457 SSL 2.0 is the only protocol enabled in the server.
15461 * Make it possible to get hexdumps of unprintable data with 'openssl
15462 asn1parse'. By implication, the functions ASN1_parse_dump() and
15463 BIO_dump_indent() are added.
15467 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15468 these print out strings and name structures based on various
15469 flags including RFC2253 support and proper handling of
15470 multibyte characters. Added options to the 'x509' utility
15471 to allow the various flags to be set.
15475 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15476 Also change the functions X509_cmp_current_time() and
15477 X509_gmtime_adj() work with an ASN1_TIME structure,
15478 this will enable certificates using GeneralizedTime in validity
15479 dates to be checked.
15483 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15484 negative public key encodings) on by default,
15485 NO_NEG_PUBKEY_BUG can be set to disable it.
15489 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15490 content octets. An i2c_ASN1_OBJECT is unnecessary because
15491 the encoding can be trivially obtained from the structure.
15495 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15496 not read locks (`CRYPTO_r_[un]lock`).
15500 * A first attempt at creating official support for shared
15501 libraries through configuration. I've kept it so the
15502 default is static libraries only, and the OpenSSL programs
15503 are always statically linked for now, but there are
15504 preparations for dynamic linking in place.
15505 This has been tested on Linux and Tru64.
15509 * Randomness polling function for Win9x, as described in:
15510 Peter Gutmann, Software Generation of Practically Strong
15515 * Fix so PRNG is seeded in req if using an already existing
15520 * New options to smime application. -inform and -outform
15521 allow alternative formats for the S/MIME message including
15522 PEM and DER. The -content option allows the content to be
15523 specified separately. This should allow things like Netscape
15524 form signing output easier to verify.
15528 * Fix the ASN1 encoding of tags using the 'long form'.
15532 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15533 STRING types. These convert content octets to and from the
15534 underlying type. The actual tag and length octets are
15535 already assumed to have been read in and checked. These
15536 are needed because all other string types have virtually
15537 identical handling apart from the tag. By having versions
15538 of the ASN1 functions that just operate on content octets
15539 IMPLICIT tagging can be handled properly. It also allows
15540 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15541 and ASN1_INTEGER are identical apart from the tag.
15545 * Change the handling of OID objects as follows:
15547 - New object identifiers are inserted in objects.txt, following
15548 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15549 - objects.pl is used to process obj_mac.num and create a new
15551 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15554 This is currently kind of a hack, and the perl code in objects.pl
15555 isn't very elegant, but it works as I intended. The simplest way
15556 to check that it worked correctly is to look in obj_dat.h and
15557 check the array nid_objs and make sure the objects haven't moved
15558 around (this is important!). Additions are OK, as well as
15559 consistent name changes.
15563 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15567 * Addition of the command line parameter '-rand file' to 'openssl req'.
15568 The given file adds to whatever has already been seeded into the
15569 random pool through the RANDFILE configuration file option or
15570 environment variable, or the default random state file.
15574 * mkstack.pl now sorts each macro group into lexical order.
15575 Previously the output order depended on the order the files
15576 appeared in the directory, resulting in needless rewriting
15581 * Patches to make OpenSSL compile under Win32 again. Mostly
15582 work arounds for the VC++ problem that it treats func() as
15583 func(void). Also stripped out the parts of mkdef.pl that
15584 added extra typesafe functions: these no longer exist.
15588 * Reorganisation of the stack code. The macros are now all
15589 collected in safestack.h . Each macro is defined in terms of
15590 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15591 DEBUG_SAFESTACK is now handled in terms of function casts,
15592 this has the advantage of retaining type safety without the
15593 use of additional functions. If DEBUG_SAFESTACK is not defined
15594 then the non typesafe macros are used instead. Also modified the
15595 mkstack.pl script to handle the new form. Needs testing to see
15596 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15597 the default if no major problems. Similar behaviour for ASN1_SET_OF
15598 and PKCS12_STACK_OF.
15602 * When some versions of IIS use the 'NET' form of private key the
15603 key derivation algorithm is different. Normally MD5(password) is
15604 used as a 128 bit RC4 key. In the modified case
15605 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15606 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15607 as the old Netscape_RSA functions except they have an additional
15608 'sgckey' parameter which uses the modified algorithm. Also added
15609 an -sgckey command line option to the rsa utility. Thanks to
15610 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15611 algorithm to openssl-dev.
15615 * The evp_local.h macros were using 'c.##kname' which resulted in
15616 invalid expansion on some systems (SCO 5.0.5 for example).
15617 Corrected to 'c.kname'.
15619 *Phillip Porch <root@theporch.com>*
15621 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15622 a STACK of email addresses from a certificate or request, these look
15623 in the subject name and the subject alternative name extensions and
15624 omit any duplicate addresses.
15628 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15629 This makes DSA verification about 2 % faster.
15633 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15634 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15635 plus overhead for 1024 bit moduli).
15636 This makes exponentiations about 0.5 % faster for 1024 bit
15637 exponents (as measured by "openssl speed rsa2048").
15641 * Rename memory handling macros to avoid conflicts with other
15643 Malloc => OPENSSL_malloc
15644 Malloc_locked => OPENSSL_malloc_locked
15645 Realloc => OPENSSL_realloc
15646 Free => OPENSSL_free
15650 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15651 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15655 * CygWin32 support.
15657 *John Jarvie <jjarvie@newsguy.com>*
15659 * The type-safe stack code has been rejigged. It is now only compiled
15660 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15661 by default all type-specific stack functions are "#define"d back to
15662 standard stack functions. This results in more streamlined output
15663 but retains the type-safety checking possibilities of the original
15668 * The STACK code has been cleaned up, and certain type declarations
15669 that didn't make a lot of sense have been brought in line. This has
15670 also involved a cleanup of sorts in safestack.h to more correctly
15671 map type-safe stack functions onto their plain stack counterparts.
15672 This work has also resulted in a variety of "const"ifications of
15673 lots of the code, especially `_cmp` operations which should normally
15674 be prototyped with "const" parameters anyway.
15678 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15679 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15680 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15681 where all of 'md' is used each time the PRNG is used, but 'state'
15682 is used only indexed by a cyclic counter. As entropy may not be
15683 well distributed from the beginning, 'md' is important as a
15684 chaining variable. However, the output function chains only half
15685 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15686 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15687 in all of 'state' being rewritten, with the new values depending
15688 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15692 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15693 the handshake is continued after ssl_verify_cert_chain();
15694 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15695 can lead to 'unexplainable' connection aborts later.
15697 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15699 * Major EVP API cipher revision.
15700 Add hooks for extra EVP features. This allows various cipher
15701 parameters to be set in the EVP interface. Support added for variable
15702 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15703 setting of RC2 and RC5 parameters.
15705 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15708 Remove lots of duplicated code from the EVP library. For example *every*
15709 cipher init() function handles the 'iv' in the same way according to the
15710 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15711 for CFB and OFB modes they zero ctx->num.
15713 New functionality allows removal of S/MIME code RC2 hack.
15715 Most of the routines have the same form and so can be declared in terms
15718 By shifting this to the top level EVP_CipherInit() it can be removed from
15719 all individual ciphers. If the cipher wants to handle IVs or keys
15720 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15723 Change lots of functions like EVP_EncryptUpdate() to now return a
15724 value: although software versions of the algorithms cannot fail
15725 any installed hardware versions can.
15729 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15730 this option is set, tolerate broken clients that send the negotiated
15731 protocol version number instead of the requested protocol version
15736 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15737 i.e. non-zero for export ciphersuites, zero otherwise.
15738 Previous versions had this flag inverted, inconsistent with
15739 rsa_tmp_cb (..._TMP_RSA_CB).
15741 *Bodo Moeller; problem reported by Amit Chopra*
15743 * Add missing DSA library text string. Work around for some IIS
15744 key files with invalid SEQUENCE encoding.
15748 * Add a document (doc/standards.txt) that list all kinds of standards
15749 and so on that are implemented in OpenSSL.
15753 * Enhance c_rehash script. Old version would mishandle certificates
15754 with the same subject name hash and wouldn't handle CRLs at all.
15755 Added -fingerprint option to crl utility, to support new c_rehash
15760 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15764 * Fix for SSL server purpose checking. Server checking was
15765 rejecting certificates which had extended key usage present
15766 but no ssl client purpose.
15768 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15770 * Make PKCS#12 code work with no password. The PKCS#12 spec
15771 is a little unclear about how a blank password is handled.
15772 Since the password in encoded as a BMPString with terminating
15773 double NULL a zero length password would end up as just the
15774 double NULL. However no password at all is different and is
15775 handled differently in the PKCS#12 key generation code. NS
15776 treats a blank password as zero length. MSIE treats it as no
15777 password on export: but it will try both on import. We now do
15778 the same: PKCS12_parse() tries zero length and no password if
15779 the password is set to "" or NULL (NULL is now a valid password:
15780 it wasn't before) as does the pkcs12 application.
15784 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15785 perror when PEM_read_bio_X509_REQ fails, the error message must
15786 be obtained from the error queue.
15790 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15791 it in ERR_remove_state if appropriate, and change ERR_get_state
15792 accordingly to avoid race conditions (this is necessary because
15793 thread_hash is no longer constant once set).
15797 * Bugfix for linux-elf makefile.one.
15801 * RSA_get_default_method() will now cause a default
15802 RSA_METHOD to be chosen if one doesn't exist already.
15803 Previously this was only set during a call to RSA_new()
15804 or RSA_new_method(NULL) meaning it was possible for
15805 RSA_get_default_method() to return NULL.
15809 * Added native name translation to the existing DSO code
15810 that will convert (if the flag to do so is set) filenames
15811 that are sufficiently small and have no path information
15812 into a canonical native form. Eg. "blah" converted to
15813 "libblah.so" or "blah.dll" etc.
15817 * New function ERR_error_string_n(e, buf, len) which is like
15818 ERR_error_string(e, buf), but writes at most 'len' bytes
15819 including the 0 terminator. For ERR_error_string_n, 'buf'
15822 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15824 * CONF library reworked to become more general. A new CONF
15825 configuration file reader "class" is implemented as well as a
15826 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15827 old `CONF_*` functions are still there, but are reimplemented to
15828 work in terms of the new functions. Also, a set of functions
15829 to handle the internal storage of the configuration data is
15830 provided to make it easier to write new configuration file
15831 reader "classes" (I can definitely see something reading a
15832 configuration file in XML format, for example), called `_CONF_*`,
15833 or "the configuration storage API"...
15835 The new configuration file reading functions are:
15837 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15838 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15840 NCONF_default, NCONF_WIN32
15842 NCONF_dump_fp, NCONF_dump_bio
15844 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15845 NCONF_new creates a new CONF object. This works in the same way
15846 as other interfaces in OpenSSL, like the BIO interface.
15847 `NCONF_dump_*` dump the internal storage of the configuration file,
15848 which is useful for debugging. All other functions take the same
15849 arguments as the old `CONF_*` functions with the exception of the
15850 first that must be a `CONF *` instead of a `LHASH *`.
15852 To make it easier to use the new classes with the old `CONF_*` functions,
15853 the function CONF_set_default_method is provided.
15857 * Add '-tls1' option to 'openssl ciphers', which was already
15858 mentioned in the documentation but had not been implemented.
15859 (This option is not yet really useful because even the additional
15860 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15864 * Initial DSO code added into libcrypto for letting OpenSSL (and
15865 OpenSSL-based applications) load shared libraries and bind to
15866 them in a portable way.
15868 *Geoff Thorpe, with contributions from Richard Levitte*
15870 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15872 * Make sure _lrotl and _lrotr are only used with MSVC.
15874 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15875 (the default implementation of RAND_status).
15877 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15878 to '-clrext' (= clear extensions), as intended and documented.
15879 *Bodo Moeller; inconsistency pointed out by Michael Attili
15880 <attili@amaxo.com>*
15882 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15883 was larger than the MD block size.
15885 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15887 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15888 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15889 using the passed key: if the passed key was a private key the result
15890 of X509_print(), for example, would be to print out all the private key
15895 * des_quad_cksum() byte order bug fix.
15896 *Ulf Möller, using the problem description in krb4-0.9.7, where
15897 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15899 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15902 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15904 * For easily testing in shell scripts whether some command
15905 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15906 returns with exit code 0 iff no command of the given name is available.
15907 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15908 the output goes to stdout and nothing is printed to stderr.
15909 Additional arguments are always ignored.
15911 Since for each cipher there is a command of the same name,
15912 the 'no-cipher' compilation switches can be tested this way.
15914 ('openssl no-XXX' is not able to detect pseudo-commands such
15915 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15919 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15923 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15924 is set; it will be thrown away anyway because each handshake creates
15926 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15927 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15928 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15929 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15933 * New s_client option -ign_eof: EOF at stdin is ignored, and
15934 'Q' and 'R' lose their special meanings (quit/renegotiate).
15935 This is part of what -quiet does; unlike -quiet, -ign_eof
15936 does not suppress any output.
15940 * Add compatibility options to the purpose and trust code. The
15941 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15942 accepts a certificate or CA, this was the previous behaviour,
15943 with all the associated security issues.
15945 X509_TRUST_COMPAT is the old trust behaviour: only and
15946 automatically trust self signed roots in certificate store. A
15947 new trust setting X509_TRUST_DEFAULT is used to specify that
15948 a purpose has no associated trust setting and it should instead
15949 use the value in the default purpose.
15953 * Fix the PKCS#8 DSA private key code so it decodes keys again
15954 and fix a memory leak.
15958 * In util/mkerr.pl (which implements 'make errors'), preserve
15959 reason strings from the previous version of the .c file, as
15960 the default to have only downcase letters (and digits) in
15961 automatically generated reasons codes is not always appropriate.
15965 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15966 using strerror. Previously, ERR_reason_error_string() returned
15967 library names as reason strings for SYSerr; but SYSerr is a special
15968 case where small numbers are errno values, not library numbers.
15972 * Add '-dsaparam' option to 'openssl dhparam' application. This
15973 converts DSA parameters into DH parameters. (When creating parameters,
15974 DSA_generate_parameters is used.)
15978 * Include 'length' (recommended exponent length) in C code generated
15979 by 'openssl dhparam -C'.
15983 * The second argument to set_label in perlasm was already being used
15984 so couldn't be used as a "file scope" flag. Moved to third argument
15989 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15990 instead of RAND_bytes for encryption IVs and salts.
15994 * Include RAND_status() into RAND_METHOD instead of implementing
15995 it only for md_rand.c Otherwise replacing the PRNG by calling
15996 RAND_set_rand_method would be impossible.
16000 * Don't let DSA_generate_key() enter an infinite loop if the random
16001 number generation fails.
16005 * New 'rand' application for creating pseudo-random output.
16009 * Added configuration support for Linux/IA64
16011 *Rolf Haberrecker <rolf@suse.de>*
16013 * Assembler module support for Mingw32.
16017 * Shared library support for HPUX (in shlib/).
16019 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16021 * Shared library support for Solaris gcc.
16023 *Lutz Behnke <behnke@trustcenter.de>*
16025 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16027 * PKCS7_encrypt() was adding text MIME headers twice because they
16028 were added manually and by SMIME_crlf_copy().
16032 * In bntest.c don't call BN_rand with zero bits argument.
16034 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16036 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16037 case was implemented. This caused BN_div_recp() to fail occasionally.
16041 * Add an optional second argument to the set_label() in the perl
16042 assembly language builder. If this argument exists and is set
16043 to 1 it signals that the assembler should use a symbol whose
16044 scope is the entire file, not just the current function. This
16045 is needed with MASM which uses the format label:: for this scope.
16047 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16049 * Change the ASN1 types so they are typedefs by default. Before
16050 almost all types were #define'd to ASN1_STRING which was causing
16051 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16056 * Change names of new functions to the new get1/get0 naming
16057 convention: After 'get1', the caller owns a reference count
16058 and has to call `..._free`; 'get0' returns a pointer to some
16059 data structure without incrementing reference counters.
16060 (Some of the existing 'get' functions increment a reference
16061 counter, some don't.)
16062 Similarly, 'set1' and 'add1' functions increase reference
16063 counters or duplicate objects.
16067 * Allow for the possibility of temp RSA key generation failure:
16068 the code used to assume it always worked and crashed on failure.
16072 * Fix potential buffer overrun problem in BIO_printf().
16073 *Ulf Möller, using public domain code by Patrick Powell; problem
16074 pointed out by David Sacerdote <das33@cornell.edu>*
16076 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16077 RAND_egd() and RAND_status(). In the command line application,
16078 the EGD socket can be specified like a seed file using RANDFILE
16083 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16084 Some CAs (e.g. Verisign) distribute certificates in this form.
16088 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16089 list to exclude them. This means that no special compilation option
16090 is needed to use anonymous DH: it just needs to be included in the
16095 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16096 EVP_MD_type. The old functionality is available in a new macro called
16097 EVP_MD_md(). Change code that uses it and update docs.
16101 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16102 where the `void *` argument is replaced by a function pointer argument.
16103 Previously `void *` was abused to point to functions, which works on
16104 many platforms, but is not correct. As these functions are usually
16105 called by macros defined in OpenSSL header files, most source code
16106 should work without changes.
16110 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16111 sections with information on -D... compiler switches used for
16112 compiling the library so that applications can see them. To enable
16113 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16114 must be defined. E.g.,
16115 #define OPENSSL_ALGORITHM_DEFINES
16116 #include <openssl/opensslconf.h>
16117 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16119 *Richard Levitte, Ulf and Bodo Möller*
16121 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16126 * Change the 'other' type in certificate aux info to a STACK_OF
16127 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16128 the required ASN1 format: arbitrary types determined by an OID.
16132 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16133 argument to 'req'. This is not because the function is newer or
16134 better than others it just uses the work 'NEW' in the certificate
16135 request header lines. Some software needs this.
16139 * Reorganise password command line arguments: now passwords can be
16140 obtained from various sources. Delete the PEM_cb function and make
16141 it the default behaviour: i.e. if the callback is NULL and the
16142 usrdata argument is not NULL interpret it as a null terminated pass
16143 phrase. If usrdata and the callback are NULL then the pass phrase
16144 is prompted for as usual.
16148 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16149 the support is automatically enabled. The resulting binaries will
16150 autodetect the card and use it if present.
16152 *Ben Laurie and Compaq Inc.*
16154 * Work around for Netscape hang bug. This sends certificate request
16155 and server done in one record. Since this is perfectly legal in the
16156 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16157 the bugs/SSLv3 entry for more info.
16161 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16165 * Add -rand argument to smime and pkcs12 applications and read/write
16170 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16174 * Add command line password options to the remaining applications.
16178 * Bug fix for BN_div_recp() for numerators with an even number of
16183 * More tests in bntest.c, and changed test_bn output.
16187 * ./config recognizes MacOS X now.
16191 * Bug fix for BN_div() when the first words of num and divisor are
16192 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16196 * Add support for various broken PKCS#8 formats, and command line
16197 options to produce them.
16201 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16202 get temporary BIGNUMs from a BN_CTX.
16206 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16211 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16212 include a #define from the old name to the new. The original intent
16213 was that statically linked binaries could for example just call
16214 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16215 link with digests. This never worked because SSLeay_add_all_digests()
16216 and SSLeay_add_all_ciphers() were in the same source file so calling
16217 one would link with the other. They are now in separate source files.
16221 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16225 * Use a less unusual form of the Miller-Rabin primality test (it used
16226 a binary algorithm for exponentiation integrated into the Miller-Rabin
16227 loop, our standard modexp algorithms are faster).
16231 * Support for the EBCDIC character set completed.
16233 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16235 * Source code cleanups: use const where appropriate, eliminate casts,
16236 use `void *` instead of `char *` in lhash.
16240 * Bugfix: ssl3_send_server_key_exchange was not restartable
16241 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16242 this the server could overwrite ephemeral keys that the client
16247 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16248 using 50 iterations of the Rabin-Miller test.
16250 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16251 iterations of the Rabin-Miller test as required by the appendix
16252 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16253 As BN_is_prime_fasttest includes trial division, DSA parameter
16254 generation becomes much faster.
16256 This implies a change for the callback functions in DSA_is_prime
16257 and DSA_generate_parameters: The callback function is called once
16258 for each positive witness in the Rabin-Miller test, not just
16259 occasionally in the inner loop; and the parameters to the
16260 callback function now provide an iteration count for the outer
16261 loop rather than for the current invocation of the inner loop.
16262 DSA_generate_parameters additionally can call the callback
16263 function with an 'iteration count' of -1, meaning that a
16264 candidate has passed the trial division test (when q is generated
16265 from an application-provided seed, trial division is skipped).
16269 * New function BN_is_prime_fasttest that optionally does trial
16270 division before starting the Rabin-Miller test and has
16271 an additional BN_CTX * argument (whereas BN_is_prime always
16272 has to allocate at least one BN_CTX).
16273 'callback(1, -1, cb_arg)' is called when a number has passed the
16274 trial division stage.
16278 * Fix for bug in CRL encoding. The validity dates weren't being handled
16283 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16287 * New function BN_pseudo_rand().
16291 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16292 bignum version of BN_from_montgomery() with the working code from
16293 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16298 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16299 made it impossible to use the same SSL_SESSION data structure in
16300 SSL2 clients in multiple threads.
16304 * The return value of RAND_load_file() no longer counts bytes obtained
16305 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16306 to seed the PRNG (previously an explicit byte count was required).
16308 *Ulf Möller, Bodo Möller*
16310 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16311 used `char *` instead of `void *` and had casts all over the place.
16315 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16319 * Retain source code compatibility for BN_prime_checks macro:
16320 BN_is_prime(..., BN_prime_checks, ...) now uses
16321 BN_prime_checks_for_size to determine the appropriate number of
16322 Rabin-Miller iterations.
16326 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16327 DH_CHECK_P_NOT_SAFE_PRIME.
16328 (Check if this is true? OpenPGP calls them "strong".)
16332 * Merge the functionality of "dh" and "gendh" programs into a new program
16333 "dhparam". The old programs are retained for now but will handle DH keys
16334 (instead of parameters) in future.
16338 * Make the ciphers, s_server and s_client programs check the return values
16339 when a new cipher list is set.
16343 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16344 ciphers. Before when the 56bit ciphers were enabled the sorting was
16347 The syntax for the cipher sorting has been extended to support sorting by
16348 cipher-strength (using the strength_bits hard coded in the tables).
16349 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16351 Fix a bug in the cipher-command parser: when supplying a cipher command
16352 string with an "undefined" symbol (neither command nor alphanumeric
16353 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16354 an error is flagged.
16356 Due to the strength-sorting extension, the code of the
16357 ssl_create_cipher_list() function was completely rearranged. I hope that
16358 the readability was also increased :-)
16360 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16362 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16363 for the first serial number and places 2 in the serial number file. This
16364 avoids problems when the root CA is created with serial number zero and
16365 the first user certificate has the same issuer name and serial number
16370 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16371 the new code. Add documentation for this stuff.
16375 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16376 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16377 structures and behave in an analogous way to the X509v3 functions:
16378 they shouldn't be called directly but wrapper functions should be used
16381 So we also now have some wrapper functions that call the X509at functions
16382 when passed certificate requests. (TO DO: similar things can be done with
16383 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16384 things. Some of these need some d2i or i2d and print functionality
16385 because they handle more complex structures.)
16389 * Add missing #ifndefs that caused missing symbols when building libssl
16390 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16391 NO_RSA in `ssl/s2*.c`.
16393 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16395 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16396 has a return value which indicates the quality of the random data
16397 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16398 error queue. New function RAND_pseudo_bytes() generates output that is
16399 guaranteed to be unique but not unpredictable. RAND_add is like
16400 RAND_seed, but takes an extra argument for an entropy estimate
16401 (RAND_seed always assumes full entropy).
16405 * Do more iterations of Rabin-Miller probable prime test (specifically,
16406 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16407 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16408 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16409 false-positive rate of at most 2^-80 for random input.
16413 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16417 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16418 in the 0.9.5 release), this returns the chain
16419 from an X509_CTX structure with a dup of the stack and all
16420 the X509 reference counts upped: so the stack will exist
16421 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16424 Also make SSL_SESSION_print() print out the verify return
16429 * Add manpage for the pkcs12 command. Also change the default
16430 behaviour so MAC iteration counts are used unless the new
16431 -nomaciter option is used. This improves file security and
16432 only older versions of MSIE (4.0 for example) need it.
16436 * Honor the no-xxx Configure options when creating .DEF files.
16440 * Add PKCS#10 attributes to field table: challengePassword,
16441 unstructuredName and unstructuredAddress. These are taken from
16442 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16443 international characters are used.
16445 More changes to X509_ATTRIBUTE code: allow the setting of types
16446 based on strings. Remove the 'loc' parameter when adding
16447 attributes because these will be a SET OF encoding which is sorted
16452 * Initial changes to the 'req' utility to allow request generation
16453 automation. This will allow an application to just generate a template
16454 file containing all the field values and have req construct the
16457 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16458 used all over the place including certificate requests and PKCS#7
16459 structures. They are currently handled manually where necessary with
16460 some primitive wrappers for PKCS#7. The new functions behave in a
16461 manner analogous to the X509 extension functions: they allow
16462 attributes to be looked up by NID and added.
16464 Later something similar to the X509V3 code would be desirable to
16465 automatically handle the encoding, decoding and printing of the
16466 more complex types. The string types like challengePassword can
16467 be handled by the string table functions.
16469 Also modified the multi byte string table handling. Now there is
16470 a 'global mask' which masks out certain types. The table itself
16471 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16472 is useful when for example there is only one permissible type
16473 (as in countryName) and using the mask might result in no valid
16478 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16479 SSL_get_peer_finished to allow applications to obtain the latest
16480 Finished messages sent to the peer or expected from the peer,
16481 respectively. (SSL_get_peer_finished is usually the Finished message
16482 actually received from the peer, otherwise the protocol will be aborted.)
16484 As the Finished message are message digests of the complete handshake
16485 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16486 be used for external authentication procedures when the authentication
16487 provided by SSL/TLS is not desired or is not enough.
16491 * Enhanced support for Alpha Linux is added. Now ./config checks if
16492 the host supports BWX extension and if Compaq C is present on the
16493 $PATH. Just exploiting of the BWX extension results in 20-30%
16494 performance kick for some algorithms, e.g. DES and RC4 to mention
16495 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16500 * Add support for MS "fast SGC". This is arguably a violation of the
16501 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16502 weak crypto and after checking the certificate is SGC a second one
16503 with strong crypto. MS SGC stops the first handshake after receiving
16504 the server certificate message and sends a second client hello. Since
16505 a server will typically do all the time consuming operations before
16506 expecting any further messages from the client (server key exchange
16507 is the most expensive) there is little difference between the two.
16509 To get OpenSSL to support MS SGC we have to permit a second client
16510 hello message after we have sent server done. In addition we have to
16511 reset the MAC if we do get this second client hello.
16515 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16516 if a DER encoded private key is RSA or DSA traditional format. Changed
16517 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16518 format DER encoded private key. Newer code should use PKCS#8 format which
16519 has the key type encoded in the ASN1 structure. Added DER private key
16520 support to pkcs8 application.
16524 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16525 ciphersuites has been selected (as required by the SSL 3/TLS 1
16526 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16527 is set, we interpret this as a request to violate the specification
16528 (the worst that can happen is a handshake failure, and 'correct'
16529 behaviour would result in a handshake failure anyway).
16533 * In SSL_CTX_add_session, take into account that there might be multiple
16534 SSL_SESSION structures with the same session ID (e.g. when two threads
16535 concurrently obtain them from an external cache).
16536 The internal cache can handle only one SSL_SESSION with a given ID,
16537 so if there's a conflict, we now throw out the old one to achieve
16542 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16543 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16544 some routines that use cipher OIDs: some ciphers do not have OIDs
16545 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16550 * Simplify the trust setting structure and code. Now we just have
16551 two sequences of OIDs for trusted and rejected settings. These will
16552 typically have values the same as the extended key usage extension
16553 and any application specific purposes.
16555 The trust checking code now has a default behaviour: it will just
16556 check for an object with the same NID as the passed id. Functions can
16557 be provided to override either the default behaviour or the behaviour
16558 for a given id. SSL client, server and email already have functions
16559 in place for compatibility: they check the NID and also return "trusted"
16560 if the certificate is self signed.
16564 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16565 traditional format into an EVP_PKEY structure.
16569 * Add a password callback function PEM_cb() which either prompts for
16570 a password if usr_data is NULL or otherwise assumes it is a null
16571 terminated password. Allow passwords to be passed on command line
16572 environment or config files in a few more utilities.
16576 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16577 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16578 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16579 Update documentation.
16583 * Support for ASN1 "NULL" type. This could be handled before by using
16584 ASN1_TYPE but there wasn't any function that would try to read a NULL
16585 and produce an error if it couldn't. For compatibility we also have
16586 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16587 don't allocate anything because they don't need to.
16591 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16594 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16596 * Rebuild of the memory allocation routines used by OpenSSL code and
16597 possibly others as well. The purpose is to make an interface that
16598 provide hooks so anyone can build a separate set of allocation and
16599 deallocation routines to be used by OpenSSL, for example memory
16600 pool implementations, or something else, which was previously hard
16601 since Malloc(), Realloc() and Free() were defined as macros having
16602 the values malloc, realloc and free, respectively (except for Win32
16603 compilations). The same is provided for memory debugging code.
16604 OpenSSL already comes with functionality to find memory leaks, but
16605 this gives people a chance to debug other memory problems.
16607 With these changes, a new set of functions and macros have appeared:
16609 CRYPTO_set_mem_debug_functions() [F]
16610 CRYPTO_get_mem_debug_functions() [F]
16611 CRYPTO_dbg_set_options() [F]
16612 CRYPTO_dbg_get_options() [F]
16613 CRYPTO_malloc_debug_init() [M]
16615 The memory debug functions are NULL by default, unless the library
16616 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16617 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16618 gives the standard debugging functions that come with OpenSSL) or
16619 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16620 provided by the library user) must be used. When the standard
16621 debugging functions are used, CRYPTO_dbg_set_options can be used to
16622 request additional information:
16623 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16624 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16626 Also, things like CRYPTO_set_mem_functions will always give the
16627 expected result (the new set of functions is used for allocation
16628 and deallocation) at all times, regardless of platform and compiler
16631 To finish it up, some functions that were never use in any other
16632 way than through macros have a new API and new semantic:
16634 CRYPTO_dbg_malloc()
16635 CRYPTO_dbg_realloc()
16638 All macros of value have retained their old syntax.
16640 *Richard Levitte and Bodo Moeller*
16642 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16643 ordering of SMIMECapabilities wasn't in "strength order" and there
16644 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16649 * Some ASN1 types with illegal zero length encoding (INTEGER,
16650 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16652 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16654 * Merge in my S/MIME library for OpenSSL. This provides a simple
16655 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16656 functionality to handle multipart/signed properly) and a utility
16657 called 'smime' to call all this stuff. This is based on code I
16658 originally wrote for Celo who have kindly allowed it to be
16659 included in OpenSSL.
16663 * Add variants des_set_key_checked and des_set_key_unchecked of
16664 des_set_key (aka des_key_sched). Global variable des_check_key
16665 decides which of these is called by des_set_key; this way
16666 des_check_key behaves as it always did, but applications and
16667 the library itself, which was buggy for des_check_key == 1,
16668 have a cleaner way to pick the version they need.
16672 * New function PKCS12_newpass() which changes the password of a
16677 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16678 dynamic mix. In both cases the ids can be used as an index into the
16679 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16680 functions so they accept a list of the field values and the
16681 application doesn't need to directly manipulate the X509_TRUST
16686 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16691 * Modify the way the V3 extension code looks up extensions. This now
16692 works in a similar way to the object code: we have some "standard"
16693 extensions in a static table which is searched with OBJ_bsearch()
16694 and the application can add dynamic ones if needed. The file
16695 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16696 updated whenever a new extension is added to the core code and kept
16697 in ext_nid order. There is a simple program 'tabtest.c' which checks
16698 this. New extensions are not added too often so this file can readily
16699 be maintained manually.
16701 There are two big advantages in doing things this way. The extensions
16702 can be looked up immediately and no longer need to be "added" using
16703 X509V3_add_standard_extensions(): this function now does nothing.
16704 Side note: I get *lots* of email saying the extension code doesn't
16705 work because people forget to call this function.
16706 Also no dynamic allocation is done unless new extensions are added:
16707 so if we don't add custom extensions there is no need to call
16708 X509V3_EXT_cleanup().
16712 * Modify enc utility's salting as follows: make salting the default. Add a
16713 magic header, so unsalted files fail gracefully instead of just decrypting
16714 to garbage. This is because not salting is a big security hole, so people
16715 should be discouraged from doing it.
16719 * Fixes and enhancements to the 'x509' utility. It allowed a message
16720 digest to be passed on the command line but it only used this
16721 parameter when signing a certificate. Modified so all relevant
16722 operations are affected by the digest parameter including the
16723 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16724 DSA key was used because it didn't fix the digest.
16728 * Initial certificate chain verify code. Currently tests the untrusted
16729 certificates for consistency with the verify purpose (which is set
16730 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16732 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16733 this is because it will reject chains with invalid extensions whereas
16734 every previous version of OpenSSL and SSLeay made no checks at all.
16736 Trust code: checks the root CA for the relevant trust settings. Trust
16737 settings have an initial value consistent with the verify purpose: e.g.
16738 if the verify purpose is for SSL client use it expects the CA to be
16739 trusted for SSL client use. However the default value can be changed to
16740 permit custom trust settings: one example of this would be to only trust
16741 certificates from a specific "secure" set of CAs.
16743 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16744 which should be used for version portability: especially since the
16745 verify structure is likely to change more often now.
16747 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16748 to set them. If not set then assume SSL clients will verify SSL servers
16751 Two new options to the verify program: -untrusted allows a set of
16752 untrusted certificates to be passed in and -purpose which sets the
16753 intended purpose of the certificate. If a purpose is set then the
16754 new chain verify code is used to check extension consistency.
16758 * Support for the authority information access extension.
16762 * Modify RSA and DSA PEM read routines to transparently handle
16763 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16764 public keys in a format compatible with certificate
16765 SubjectPublicKeyInfo structures. Unfortunately there were already
16766 functions called *_PublicKey_* which used various odd formats so
16767 these are retained for compatibility: however the DSA variants were
16768 never in a public release so they have been deleted. Changed dsa/rsa
16769 utilities to handle the new format: note no releases ever handled public
16770 keys so we should be OK.
16772 The primary motivation for this change is to avoid the same fiasco
16773 that dogs private keys: there are several incompatible private key
16774 formats some of which are standard and some OpenSSL specific and
16775 require various evil hacks to allow partial transparent handling and
16776 even then it doesn't work with DER formats. Given the option anything
16777 other than PKCS#8 should be dumped: but the other formats have to
16778 stay in the name of compatibility.
16780 With public keys and the benefit of hindsight one standard format
16781 is used which works with EVP_PKEY, RSA or DSA structures: though
16782 it clearly returns an error if you try to read the wrong kind of key.
16784 Added a -pubkey option to the 'x509' utility to output the public key.
16785 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16786 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16787 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16788 that do the same as the `EVP_PKEY_assign_*()` except they up the
16789 reference count of the added key (they don't "swallow" the
16794 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16795 CRLs would fail if the file contained no certificates or no CRLs:
16796 added a new function to read in both types and return the number
16797 read: this means that if none are read it will be an error. The
16798 DER versions of the certificate and CRL reader would always fail
16799 because it isn't possible to mix certificates and CRLs in DER format
16800 without choking one or the other routine. Changed this to just read
16801 a certificate: this is the best we can do. Also modified the code
16802 in `apps/verify.c` to take notice of return codes: it was previously
16803 attempting to read in certificates from NULL pointers and ignoring
16804 any errors: this is one reason why the cert and CRL reader seemed
16805 to work. It doesn't check return codes from the default certificate
16806 routines: these may well fail if the certificates aren't installed.
16810 * Code to support otherName option in GeneralName.
16814 * First update to verify code. Change the verify utility
16815 so it warns if it is passed a self signed certificate:
16816 for consistency with the normal behaviour. X509_verify
16817 has been modified to it will now verify a self signed
16818 certificate if *exactly* the same certificate appears
16819 in the store: it was previously impossible to trust a
16820 single self signed certificate. This means that:
16821 openssl verify ss.pem
16822 now gives a warning about a self signed certificate but
16823 openssl verify -CAfile ss.pem ss.pem
16828 * For servers, store verify_result in SSL_SESSION data structure
16829 (and add it to external session representation).
16830 This is needed when client certificate verifications fails,
16831 but an application-provided verification callback (set by
16832 SSL_CTX_set_cert_verify_callback) allows accepting the session
16833 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16834 but returns 1): When the session is reused, we have to set
16835 ssl->verify_result to the appropriate error code to avoid
16838 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16840 * Fix a bug in the new PKCS#7 code: it didn't consider the
16841 case in PKCS7_dataInit() where the signed PKCS7 structure
16842 didn't contain any existing data because it was being created.
16844 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16846 * Add a salt to the key derivation routines in enc.c. This
16847 forms the first 8 bytes of the encrypted file. Also add a
16848 -S option to allow a salt to be input on the command line.
16852 * New function X509_cmp(). Oddly enough there wasn't a function
16853 to compare two certificates. We do this by working out the SHA1
16854 hash and comparing that. X509_cmp() will be needed by the trust
16859 * SSL_get1_session() is like SSL_get_session(), but increments
16860 the reference count in the SSL_SESSION returned.
16862 *Geoff Thorpe <geoff@eu.c2.net>*
16864 * Fix for 'req': it was adding a null to request attributes.
16865 Also change the X509_LOOKUP and X509_INFO code to handle
16866 certificate auxiliary information.
16870 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16875 * Add the possibility to add extra information to the memory leak
16876 detecting output, to form tracebacks, showing from where each
16877 allocation was originated: CRYPTO_push_info("constant string") adds
16878 the string plus current file name and line number to a per-thread
16879 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16880 is like calling CYRPTO_pop_info() until the stack is empty.
16881 Also updated memory leak detection code to be multi-thread-safe.
16885 * Add options -text and -noout to pkcs7 utility and delete the
16886 encryption options which never did anything. Update docs.
16890 * Add options to some of the utilities to allow the pass phrase
16891 to be included on either the command line (not recommended on
16892 OSes like Unix) or read from the environment. Update the
16893 manpages and fix a few bugs.
16897 * Add a few manpages for some of the openssl commands.
16901 * Fix the -revoke option in ca. It was freeing up memory twice,
16902 leaking and not finding already revoked certificates.
16906 * Extensive changes to support certificate auxiliary information.
16907 This involves the use of X509_CERT_AUX structure and X509_AUX
16908 functions. An X509_AUX function such as PEM_read_X509_AUX()
16909 can still read in a certificate file in the usual way but it
16910 will also read in any additional "auxiliary information". By
16911 doing things this way a fair degree of compatibility can be
16912 retained: existing certificates can have this information added
16913 using the new 'x509' options.
16915 Current auxiliary information includes an "alias" and some trust
16916 settings. The trust settings will ultimately be used in enhanced
16917 certificate chain verification routines: currently a certificate
16918 can only be trusted if it is self signed and then it is trusted
16923 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16924 The problem was that one of the replacement routines had not been working
16925 since SSLeay releases. For now the offending routine has been replaced
16926 with non-optimised assembler. Even so, this now gives around 95%
16927 performance improvement for 1024 bit RSA signs.
16931 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16932 handling. Most clients have the effective key size in bits equal to
16933 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16934 A few however don't do this and instead use the size of the decrypted key
16935 to determine the RC2 key length and the AlgorithmIdentifier to determine
16936 the effective key length. In this case the effective key length can still
16937 be 40 bits but the key length can be 168 bits for example. This is fixed
16938 by manually forcing an RC2 key into the EVP_PKEY structure because the
16939 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16940 the key length and effective key length are equal.
16944 * Add a bunch of functions that should simplify the creation of
16945 X509_NAME structures. Now you should be able to do:
16946 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16947 and have it automatically work out the correct field type and fill in
16948 the structures. The more adventurous can try:
16949 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16950 and it will (hopefully) work out the correct multibyte encoding.
16954 * Change the 'req' utility to use the new field handling and multibyte
16955 copy routines. Before the DN field creation was handled in an ad hoc
16956 way in req, ca, and x509 which was rather broken and didn't support
16957 BMPStrings or UTF8Strings. Since some software doesn't implement
16958 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16959 using the dirstring_type option. See the new comment in the default
16960 openssl.cnf for more info.
16964 * Make crypto/rand/md_rand.c more robust:
16965 - Assure unique random numbers after fork().
16966 - Make sure that concurrent threads access the global counter and
16967 md serializably so that we never lose entropy in them
16968 or use exactly the same state in multiple threads.
16969 Access to the large state is not always serializable because
16970 the additional locking could be a performance killer, and
16971 md should be large enough anyway.
16975 * New file `apps/app_rand.c` with commonly needed functionality
16976 for handling the random seed file.
16978 Use the random seed file in some applications that previously did not:
16980 dsaparam -genkey (which also ignored its '-rand' option),
16983 x509 (when signing).
16984 Except on systems with /dev/urandom, it is crucial to have a random
16985 seed file at least for key creation, DSA signing, and for DH exchanges;
16986 for RSA signatures we could do without one.
16988 gendh and gendsa (unlike genrsa) used to read only the first byte
16989 of each file listed in the '-rand' option. The function as previously
16990 found in genrsa is now in app_rand.c and is used by all programs
16991 that support '-rand'.
16995 * In RAND_write_file, use mode 0600 for creating files;
16996 don't just chmod when it may be too late.
17000 * Report an error from X509_STORE_load_locations
17001 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17005 * New function ASN1_mbstring_copy() this copies a string in either
17006 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17007 into an ASN1_STRING type. A mask of permissible types is passed
17008 and it chooses the "minimal" type to use or an error if not type
17013 * Add function equivalents to the various macros in asn1.h. The old
17014 macros are retained with an `M_` prefix. Code inside the library can
17015 use the `M_` macros. External code (including the openssl utility)
17016 should *NOT* in order to be "shared library friendly".
17020 * Add various functions that can check a certificate's extensions
17021 to see if it usable for various purposes such as SSL client,
17022 server or S/MIME and CAs of these types. This is currently
17023 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17024 verification. Also added a -purpose flag to x509 utility to
17025 print out all the purposes.
17029 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17034 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17035 for, obtain and decode and extension and obtain its critical flag.
17036 This allows all the necessary extension code to be handled in a
17037 single function call.
17041 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17042 platforms. See crypto/rc4/rc4_enc.c for further details.
17046 * New -noout option to asn1parse. This causes no output to be produced
17047 its main use is when combined with -strparse and -out to extract data
17048 from a file (which may not be in ASN.1 format).
17052 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17053 when producing the local key id.
17055 *Richard Levitte <levitte@stacken.kth.se>*
17057 * New option -dhparam in s_server. This allows a DH parameter file to be
17058 stated explicitly. If it is not stated then it tries the first server
17059 certificate file. The previous behaviour hard coded the filename
17064 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17065 a public key to be input or output. For example:
17066 openssl rsa -in key.pem -pubout -out pubkey.pem
17067 Also added necessary DSA public key functions to handle this.
17071 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17072 in the message. This was handled by allowing
17073 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17075 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17077 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17078 to the end of the strings whereas this didn't. This would cause problems
17079 if strings read with d2i_ASN1_bytes() were later modified.
17081 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17083 * Fix for base64 decode bug. When a base64 bio reads only one line of
17084 data and it contains EOF it will end up returning an error. This is
17085 caused by input 46 bytes long. The cause is due to the way base64
17086 BIOs find the start of base64 encoded data. They do this by trying a
17087 trial decode on each line until they find one that works. When they
17088 do a flag is set and it starts again knowing it can pass all the
17089 data directly through the decoder. Unfortunately it doesn't reset
17090 the context it uses. This means that if EOF is reached an attempt
17091 is made to pass two EOFs through the context and this causes the
17092 resulting error. This can also cause other problems as well. As is
17093 usual with these problems it takes *ages* to find and the fix is
17094 trivial: move one line.
17096 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17098 * Ugly workaround to get s_client and s_server working under Windows. The
17099 old code wouldn't work because it needed to select() on sockets and the
17100 tty (for keypresses and to see if data could be written). Win32 only
17101 supports select() on sockets so we select() with a 1s timeout on the
17102 sockets and then see if any characters are waiting to be read, if none
17103 are present then we retry, we also assume we can always write data to
17104 the tty. This isn't nice because the code then blocks until we've
17105 received a complete line of data and it is effectively polling the
17106 keyboard at 1s intervals: however it's quite a bit better than not
17107 working at all :-) A dedicated Windows application might handle this
17108 with an event loop for example.
17112 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17113 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17114 will be called when RSA_sign() and RSA_verify() are used. This is useful
17115 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17116 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17117 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17118 This necessitated the support of an extra signature type NID_md5_sha1
17119 for SSL signatures and modifications to the SSL library to use it instead
17120 of calling RSA_public_decrypt() and RSA_private_encrypt().
17124 * Add new -verify -CAfile and -CApath options to the crl program, these
17125 will lookup a CRL issuers certificate and verify the signature in a
17126 similar way to the verify program. Tidy up the crl program so it
17127 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17128 less strict. It will now permit CRL extensions even if it is not
17129 a V2 CRL: this will allow it to tolerate some broken CRLs.
17133 * Initialize all non-automatic variables each time one of the openssl
17134 sub-programs is started (this is necessary as they may be started
17135 multiple times from the "OpenSSL>" prompt).
17137 *Lennart Bang, Bodo Moeller*
17139 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17140 removing all other RSA functionality (this is what NO_RSA does). This
17141 is so (for example) those in the US can disable those operations covered
17142 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17147 * Non-copying interface to BIO pairs.
17148 (still largely untested)
17152 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17153 ASCII string. This was handled independently in various places before.
17157 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17158 UTF8 strings a character at a time.
17162 * Use client_version from client hello to select the protocol
17163 (s23_srvr.c) and for RSA client key exchange verification
17164 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17168 * Add various utility functions to handle SPKACs, these were previously
17169 handled by poking round in the structure internals. Added new function
17170 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17171 print, verify and generate SPKACs. Based on an original idea from
17172 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17176 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17180 * Allow the config file extension section to be overwritten on the
17181 command line. Based on an original idea from Massimiliano Pala
17182 <madwolf@comune.modena.it>. The new option is called -extensions
17183 and can be applied to ca, req and x509. Also -reqexts to override
17184 the request extensions in req and -crlexts to override the crl extensions
17189 * Add new feature to the SPKAC handling in ca. Now you can include
17190 the same field multiple times by preceding it by "XXXX." for example:
17193 this is the same syntax as used in the req config file.
17197 * Allow certificate extensions to be added to certificate requests. These
17198 are specified in a 'req_extensions' option of the req section of the
17199 config file. They can be printed out with the -text option to req but
17200 are otherwise ignored at present.
17204 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17205 data read consists of only the final block it would not decrypted because
17206 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17207 A misplaced 'break' also meant the decrypted final block might not be
17208 copied until the next read.
17212 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17213 a few extra parameters to the DH structure: these will be useful if
17214 for example we want the value of 'q' or implement X9.42 DH.
17218 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17219 provides hooks that allow the default DSA functions or functions on a
17220 "per key" basis to be replaced. This allows hardware acceleration and
17221 hardware key storage to be handled without major modification to the
17222 library. Also added low level modexp hooks and CRYPTO_EX structure and
17223 associated functions.
17227 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17228 as "read only": it can't be written to and the buffer it points to will
17229 not be freed. Reading from a read only BIO is much more efficient than
17230 a normal memory BIO. This was added because there are several times when
17231 an area of memory needs to be read from a BIO. The previous method was
17232 to create a memory BIO and write the data to it, this results in two
17233 copies of the data and an O(n^2) reading algorithm. There is a new
17234 function BIO_new_mem_buf() which creates a read only memory BIO from
17235 an area of memory. Also modified the PKCS#7 routines to use read only
17240 * Bugfix: ssl23_get_client_hello did not work properly when called in
17241 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17242 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17243 but a retry condition occurred while trying to read the rest.
17247 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17248 NID_pkcs7_encrypted by default: this was wrong since this should almost
17249 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17250 the encrypted data type: this is a more sensible place to put it and it
17251 allows the PKCS#12 code to be tidied up that duplicated this
17256 * Changed obj_dat.pl script so it takes its input and output files on
17257 the command line. This should avoid shell escape redirection problems
17262 * Initial support for certificate extension requests, these are included
17263 in things like Xenroll certificate requests. Included functions to allow
17264 extensions to be obtained and added.
17268 * -crlf option to s_client and s_server for sending newlines as
17269 CRLF (as required by many protocols).
17273 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17275 * Install libRSAglue.a when OpenSSL is built with RSAref.
17277 *Ralf S. Engelschall*
17279 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17281 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17283 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17288 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17289 DH parameters/keys (q is lost during that conversion, but the resulting
17290 DH parameters contain its length).
17292 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17293 much faster than DH_generate_parameters (which creates parameters
17294 where `p = 2*q + 1`), and also the smaller q makes DH computations
17295 much more efficient (160-bit exponentiation instead of 1024-bit
17296 exponentiation); so this provides a convenient way to support DHE
17297 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17298 utter importance to use
17299 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17301 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17302 when such DH parameters are used, because otherwise small subgroup
17303 attacks may become possible!
17307 * Avoid memory leak in i2d_DHparams.
17311 * Allow the -k option to be used more than once in the enc program:
17312 this allows the same encrypted message to be read by multiple recipients.
17316 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17317 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17318 it will always use the numerical form of the OID, even if it has a short
17323 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17324 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17325 otherwise bn_mod_exp was called. In the case of hardware keys for example
17326 no private key components need be present and it might store extra data
17327 in the RSA structure, which cannot be accessed from bn_mod_exp.
17328 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17329 private key operations.
17333 * Added support for SPARC Linux.
17337 * pem_password_cb function type incompatibly changed from
17338 typedef int pem_password_cb(char *buf, int size, int rwflag);
17340 ....(char *buf, int size, int rwflag, void *userdata);
17341 so that applications can pass data to their callbacks:
17342 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17343 additional void * argument, which is just handed through whenever
17344 the password callback is called.
17346 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17348 New function SSL_CTX_set_default_passwd_cb_userdata.
17350 Compatibility note: As many C implementations push function arguments
17351 onto the stack in reverse order, the new library version is likely to
17352 interoperate with programs that have been compiled with the old
17353 pem_password_cb definition (PEM_whatever takes some data that
17354 happens to be on the stack as its last argument, and the callback
17355 just ignores this garbage); but there is no guarantee whatsoever that
17358 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17359 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17360 problems not only on Windows, but also on some Unix platforms.
17361 To avoid problematic command lines, these definitions are now in an
17362 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17363 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17367 * MIPS III/IV assembler module is reimplemented.
17371 * More DES library cleanups: remove references to srand/rand and
17372 delete an unused file.
17376 * Add support for the free Netwide assembler (NASM) under Win32,
17377 since not many people have MASM (ml) and it can be hard to obtain.
17378 This is currently experimental but it seems to work OK and pass all
17379 the tests. Check out INSTALL.W32 for info.
17383 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17384 without temporary keys kept an extra copy of the server key,
17385 and connections with temporary keys did not free everything in case
17390 * New function RSA_check_key and new openssl rsa option -check
17391 for verifying the consistency of RSA keys.
17393 *Ulf Moeller, Bodo Moeller*
17395 * Various changes to make Win32 compile work:
17396 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17397 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17398 comparison" warnings.
17399 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17403 * Add a debugging option to PKCS#5 v2 key generation function: when
17404 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17405 derived keys are printed to stderr.
17409 * Copy the flags in ASN1_STRING_dup().
17411 *Roman E. Pavlov <pre@mo.msk.ru>*
17413 * The x509 application mishandled signing requests containing DSA
17414 keys when the signing key was also DSA and the parameters didn't match.
17416 It was supposed to omit the parameters when they matched the signing key:
17417 the verifying software was then supposed to automatically use the CA's
17418 parameters if they were absent from the end user certificate.
17420 Omitting parameters is no longer recommended. The test was also
17421 the wrong way round! This was probably due to unusual behaviour in
17422 EVP_cmp_parameters() which returns 1 if the parameters match.
17423 This meant that parameters were omitted when they *didn't* match and
17424 the certificate was useless. Certificates signed with 'ca' didn't have
17427 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17429 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17430 The interface is as follows:
17431 Applications can use
17432 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17433 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17434 "off" is now the default.
17435 The library internally uses
17436 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17437 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17438 to disable memory-checking temporarily.
17440 Some inconsistent states that previously were possible (and were
17441 even the default) are now avoided.
17443 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17444 with each memory chunk allocated; this is occasionally more helpful
17445 than just having a counter.
17447 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17449 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17454 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17455 which largely parallels "options", but is for changing API behaviour,
17456 whereas "options" are about protocol behaviour.
17457 Initial "mode" flags are:
17459 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17460 a single record has been written.
17461 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17462 retries use the same buffer location.
17463 (But all of the contents must be
17468 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17471 * Fix problems with no-hmac etc.
17473 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17475 * New functions RSA_get_default_method(), RSA_set_method() and
17476 RSA_get_method(). These allows replacement of RSA_METHODs without having
17477 to mess around with the internals of an RSA structure.
17481 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17482 Also really enable memory leak checks in openssl.c and in some
17485 *Chad C. Mulligan, Bodo Moeller*
17487 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17488 up the length of negative integers. This has now been simplified to just
17489 store the length when it is first determined and use it later, rather
17490 than trying to keep track of where data is copied and updating it to
17492 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17494 * Add a new function PKCS7_signatureVerify. This allows the verification
17495 of a PKCS#7 signature but with the signing certificate passed to the
17496 function itself. This contrasts with PKCS7_dataVerify which assumes the
17497 certificate is present in the PKCS#7 structure. This isn't always the
17498 case: certificates can be omitted from a PKCS#7 structure and be
17499 distributed by "out of band" means (such as a certificate database).
17503 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17504 function prototypes in pem.h, also change util/mkdef.pl to add the
17505 necessary function names.
17509 * mk1mf.pl (used by Windows builds) did not properly read the
17510 options set by Configure in the top level Makefile, and Configure
17511 was not even able to write more than one option correctly.
17512 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17516 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17517 file to be loaded from a BIO or FILE pointer. The BIO version will
17518 for example allow memory BIOs to contain config info.
17522 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17523 Whoever hopes to achieve shared-library compatibility across versions
17524 must use this, not the compile-time macro.
17525 (Exercise 0.9.4: Which is the minimum library version required by
17527 Note: All this applies only to multi-threaded programs, others don't
17532 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17533 through a BIO pair triggered the default case, i.e.
17534 SSLerr(...,SSL_R_UNKNOWN_STATE).
17538 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17539 can use the SSL library even if none of the specific BIOs is
17544 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17545 for the encoded length.
17547 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17549 * Add initial documentation of the X509V3 functions.
17553 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17554 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17555 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17556 secure PKCS#8 private key format with a high iteration count.
17560 * Fix determination of Perl interpreter: A perl or perl5
17561 *directory* in $PATH was also accepted as the interpreter.
17563 *Ralf S. Engelschall*
17565 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17566 wrong with it but it was very old and did things like calling
17567 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17568 unusual formatting.
17572 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17573 to use the new extension code.
17577 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17578 with macros. This should make it easier to change their form, add extra
17579 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17584 * Add to configuration table a new entry that can specify an alternative
17585 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17586 according to Mark Crispin <MRC@Panda.COM>.
17590 * DES CBC did not update the IV. Weird.
17594 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17595 Changing the behaviour of the former might break existing programs --
17596 where IV updating is needed, des_ncbc_encrypt can be used.
17599 * When bntest is run from "make test" it drives bc to check its
17600 calculations, as well as internally checking them. If an internal check
17601 fails, it needs to cause bc to give a non-zero result or make test carries
17602 on without noticing the failure. Fixed.
17606 * DES library cleanups.
17610 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17611 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17612 ciphers. NOTE: although the key derivation function has been verified
17613 against some published test vectors it has not been extensively tested
17614 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17619 * Instead of "mkdir -p", which is not fully portable, use new
17620 Perl script "util/mkdir-p.pl".
17624 * Rewrite the way password based encryption (PBE) is handled. It used to
17625 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17626 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17627 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17628 the 'parameter' field of the AlgorithmIdentifier is passed to the
17629 underlying key generation function so it must do its own ASN1 parsing.
17630 This has also changed the EVP_PBE_CipherInit() function which now has a
17631 'parameter' argument instead of literal salt and iteration count values
17632 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17636 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17637 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17638 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17639 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17640 value was just used as a "magic string" and not used directly its
17641 value doesn't matter.
17645 * Introduce some semblance of const correctness to BN. Shame C doesn't
17650 * "linux-sparc64" configuration (ultrapenguin).
17652 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17653 "linux-sparc" configuration.
17655 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17657 * config now generates no-xxx options for missing ciphers.
17661 * Support the EBCDIC character set (work in progress).
17662 File ebcdic.c not yet included because it has a different license.
17664 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17666 * Support BS2000/OSD-POSIX.
17668 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17670 * Make callbacks for key generation use `void *` instead of `char *`.
17674 * Make S/MIME samples compile (not yet tested).
17678 * Additional typesafe stacks.
17682 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17686 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17688 * New configuration variant "sco5-gcc".
17690 * Updated some demos.
17692 *Sean O Riordain, Wade Scholine*
17694 * Add missing BIO_free at exit of pkcs12 application.
17698 * Fix memory leak in conf.c.
17702 * Updates for Win32 to assembler version of MD5.
17706 * Set #! path to perl in `apps/der_chop` to where we found it
17707 instead of using a fixed path.
17711 * SHA library changes for irix64-mips4-cc.
17715 * Improvements for VMS support.
17719 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17721 * Bignum library bug fix. IRIX 6 passes "make test" now!
17722 This also avoids the problems with SC4.2 and unpatched SC5.
17724 *Andy Polyakov <appro@fy.chalmers.se>*
17726 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17727 These are required because of the typesafe stack would otherwise break
17728 existing code. If old code used a structure member which used to be STACK
17729 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17730 sk_num or sk_value it would produce an error because the num, data members
17731 are not present in STACK_OF. Now it just produces a warning. sk_set
17732 replaces the old method of assigning a value to sk_value
17733 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17734 that does this will no longer work (and should use sk_set instead) but
17735 this could be regarded as a "questionable" behaviour anyway.
17739 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17740 correctly handle encrypted S/MIME data.
17744 * Change type of various DES function arguments from des_cblock
17745 (which means, in function argument declarations, pointer to char)
17746 to des_cblock * (meaning pointer to array with 8 char elements),
17747 which allows the compiler to do more typechecking; it was like
17748 that back in SSLeay, but with lots of ugly casts.
17750 Introduce new type const_des_cblock.
17754 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17755 problems: find RecipientInfo structure that matches recipient certificate
17756 and initialise the ASN1 structures properly based on passed cipher.
17760 * Belatedly make the BN tests actually check the results.
17764 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17765 to and from BNs: it was completely broken. New compilation option
17766 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17767 key elements as negative integers.
17771 * Reorganize and speed up MD5.
17773 *Andy Polyakov <appro@fy.chalmers.se>*
17777 *Richard Levitte <richard@levitte.org>*
17779 * New option -out to asn1parse to allow the parsed structure to be
17780 output to a file. This is most useful when combined with the -strparse
17781 option to examine the output of things like OCTET STRINGS.
17785 * Make SSL library a little more fool-proof by not requiring any longer
17786 that `SSL_set_{accept,connect}_state` be called before
17787 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17788 in many applications because usually everything *appeared* to work as
17789 intended anyway -- now it really works as intended).
17793 * Move openssl.cnf out of lib/.
17797 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17798 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17799 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17801 *Ralf S. Engelschall*
17803 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17804 handle PKCS#7 enveloped data properly.
17806 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17808 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17809 copying pointers. The cert_st handling is changed by this in
17810 various ways (and thus what used to be known as ctx->default_cert
17811 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17812 any longer when s->cert does not give us what we need).
17813 ssl_cert_instantiate becomes obsolete by this change.
17814 As soon as we've got the new code right (possibly it already is?),
17815 we have solved a couple of bugs of the earlier code where s->cert
17816 was used as if it could not have been shared with other SSL structures.
17818 Note that using the SSL API in certain dirty ways now will result
17819 in different behaviour than observed with earlier library versions:
17820 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17821 does not influence s as it used to.
17823 In order to clean up things more thoroughly, inside SSL_SESSION
17824 we don't use CERT any longer, but a new structure SESS_CERT
17825 that holds per-session data (if available); currently, this is
17826 the peer's certificate chain and, for clients, the server's certificate
17827 and temporary key. CERT holds only those values that can have
17828 meaningful defaults in an SSL_CTX.
17832 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17833 from the internal representation. Various PKCS#7 fixes: remove some
17834 evil casts and set the enc_dig_alg field properly based on the signing
17839 * Allow PKCS#12 password to be set from the command line or the
17840 environment. Let 'ca' get its config file name from the environment
17841 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17846 * Allow certificate policies extension to use an IA5STRING for the
17847 organization field. This is contrary to the PKIX definition but
17848 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17853 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17854 without disallowing inline assembler and the like for non-pedantic builds.
17858 * Support Borland C++ builder.
17860 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17866 * SHA-1 cleanups and performance enhancements.
17868 *Andy Polyakov <appro@fy.chalmers.se>*
17870 * Sparc v8plus assembler for the bignum library.
17872 *Andy Polyakov <appro@fy.chalmers.se>*
17874 * Accept any -xxx and +xxx compiler options in Configure.
17878 * Update HPUX configuration.
17882 * Add missing `sk_<type>_unshift()` function to safestack.h
17884 *Ralf S. Engelschall*
17886 * New function SSL_CTX_use_certificate_chain_file that sets the
17887 "extra_cert"s in addition to the certificate. (This makes sense
17888 only for "PEM" format files, as chains as a whole are not
17893 * Support verify_depth from the SSL API.
17894 x509_vfy.c had what can be considered an off-by-one-error:
17895 Its depth (which was not part of the external interface)
17896 was actually counting the number of certificates in a chain;
17897 now it really counts the depth.
17901 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17902 instead of X509err, which often resulted in confusing error
17903 messages since the error codes are not globally unique
17904 (e.g. an alleged error in ssl3_accept when a certificate
17905 didn't match the private key).
17907 * New function SSL_CTX_set_session_id_context that allows to set a default
17908 value (so that you don't need SSL_set_session_id_context for each
17909 connection using the SSL_CTX).
17913 * OAEP decoding bug fix.
17917 * Support INSTALL_PREFIX for package builders, as proposed by
17922 * New Configure options "threads" and "no-threads". For systems
17923 where the proper compiler options are known (currently Solaris
17924 and Linux), "threads" is the default.
17928 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17932 * Install various scripts to $(OPENSSLDIR)/misc, not to
17933 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17934 such as /usr/local/bin.
17938 * "make linux-shared" to build shared libraries.
17940 *Niels Poppe <niels@netbox.org>*
17942 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17946 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17947 extension adding in x509 utility.
17951 * Remove NOPROTO sections and error code comments.
17955 * Partial rewrite of the DEF file generator to now parse the ANSI
17960 * New Configure options --prefix=DIR and --openssldir=DIR.
17964 * Complete rewrite of the error code script(s). It is all now handled
17965 by one script at the top level which handles error code gathering,
17966 header rewriting and C source file generation. It should be much better
17967 than the old method: it now uses a modified version of Ulf's parser to
17968 read the ANSI prototypes in all header files (thus the old K&R definitions
17969 aren't needed for error creation any more) and do a better job of
17970 translating function codes into names. The old 'ASN1 error code embedded
17971 in a comment' is no longer necessary and it doesn't use .err files which
17972 have now been deleted. Also the error code call doesn't have to appear all
17973 on one line (which resulted in some large lines...).
17977 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17981 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17982 0 (which usually indicates a closed connection), but continue reading.
17986 * Fix some race conditions.
17990 * Add support for CRL distribution points extension. Add Certificate
17991 Policies and CRL distribution points documentation.
17995 * Move the autogenerated header file parts to crypto/opensslconf.h.
17999 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18000 8 of keying material. Merlin has also confirmed interop with this fix
18001 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18003 *Merlin Hughes <merlin@baltimore.ie>*
18005 * Fix lots of warnings.
18007 *Richard Levitte <levitte@stacken.kth.se>*
18009 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18010 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18012 *Richard Levitte <levitte@stacken.kth.se>*
18014 * Fix problems with sizeof(long) == 8.
18016 *Andy Polyakov <appro@fy.chalmers.se>*
18018 * Change functions to ANSI C.
18022 * Fix typos in error codes.
18024 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18026 * Remove defunct assembler files from Configure.
18030 * SPARC v8 assembler BIGNUM implementation.
18032 *Andy Polyakov <appro@fy.chalmers.se>*
18034 * Support for Certificate Policies extension: both print and set.
18035 Various additions to support the r2i method this uses.
18039 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18040 return a const string when you are expecting an allocated buffer.
18044 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18045 types DirectoryString and DisplayText.
18049 * Add code to allow r2i extensions to access the configuration database,
18050 add an LHASH database driver and add several ctx helper functions.
18054 * Fix an evil bug in bn_expand2() which caused various BN functions to
18055 fail when they extended the size of a BIGNUM.
18059 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18060 support typesafe stack.
18064 * Fix typo in SSL_[gs]et_options().
18066 *Nils Frostberg <nils@medcom.se>*
18068 * Delete various functions and files that belonged to the (now obsolete)
18069 old X509V3 handling code.
18073 * New Configure option "rsaref".
18077 * Don't auto-generate pem.h.
18081 * Introduce type-safe ASN.1 SETs.
18085 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18087 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18089 * Introduce type-safe STACKs. This will almost certainly break lots of code
18090 that links with OpenSSL (well at least cause lots of warnings), but fear
18091 not: the conversion is trivial, and it eliminates loads of evil casts. A
18092 few STACKed things have been converted already. Feel free to convert more.
18093 In the fullness of time, I'll do away with the STACK type altogether.
18097 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18098 specified in `<certfile>` by updating the entry in the index.txt file.
18099 This way one no longer has to edit the index.txt file manually for
18100 revoking a certificate. The -revoke option does the gory details now.
18102 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18104 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18105 `-text` option at all and this way the `-noout -text` combination was
18106 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18108 *Ralf S. Engelschall*
18110 * Make sure a corresponding plain text error message exists for the
18111 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18112 verify callback function determined that a certificate was revoked.
18114 *Ralf S. Engelschall*
18116 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18117 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18118 all available ciphers including rc5, which was forgotten until now.
18119 In order to let the testing shell script know which algorithms
18120 are available, a new (up to now undocumented) command
18121 `openssl list-cipher-commands` is used.
18125 * Bugfix: s_client occasionally would sleep in select() when
18126 it should have checked SSL_pending() first.
18130 * New functions DSA_do_sign and DSA_do_verify to provide access to
18131 the raw DSA values prior to ASN.1 encoding.
18135 * Tweaks to Configure
18137 *Niels Poppe <niels@netbox.org>*
18139 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18144 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18148 * New config option to avoid instructions that are illegal on the 80386.
18149 The default code is faster, but requires at least a 486.
18153 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18154 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18155 same as SSL2_VERSION anyway.
18159 * New "-showcerts" option for s_client.
18163 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18164 application. Various cleanups and fixes.
18168 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18169 modify error routines to work internally. Add error codes and PBE init
18170 to library startup routines.
18174 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18175 packing functions to asn1 and evp. Changed function names and error
18176 codes along the way.
18180 * PKCS12 integration: and so it begins... First of several patches to
18181 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18182 objects to objects.h
18186 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18187 and display support for Thawte strong extranet extension.
18191 * Add LinuxPPC support.
18193 *Jeff Dubrule <igor@pobox.org>*
18195 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18196 bn_div_words in alpha.s.
18198 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18200 * Make sure the RSA OAEP test is skipped under -DRSAref because
18201 OAEP isn't supported when OpenSSL is built with RSAref.
18203 *Ulf Moeller <ulf@fitug.de>*
18205 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18206 so they no longer are missing under -DNOPROTO.
18208 *Soren S. Jorvang <soren@t.dk>*
18210 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18212 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18213 doesn't work when the session is reused. Coming soon!
18217 * Fix a security hole, that allows sessions to be reused in the wrong
18218 context thus bypassing client cert protection! All software that uses
18219 client certs and session caches in multiple contexts NEEDS PATCHING to
18220 allow session reuse! A fuller solution is in the works.
18222 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18224 * Some more source tree cleanups (removed obsolete files
18225 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18226 permission on "config" script to be executable) and a fix for the INSTALL
18229 *Ulf Moeller <ulf@fitug.de>*
18231 * Remove some legacy and erroneous uses of malloc, free instead of
18234 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18236 * Make rsa_oaep_test return non-zero on error.
18238 *Ulf Moeller <ulf@fitug.de>*
18240 * Add support for native Solaris shared libraries. Configure
18241 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18242 if someone would make that last step automatic.
18244 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18246 * ctx_size was not built with the right compiler during "make links". Fixed.
18250 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18251 except NULL ciphers". This means the default cipher list will no longer
18252 enable NULL ciphers. They need to be specifically enabled e.g. with
18253 the string "DEFAULT:eNULL".
18257 * Fix to RSA private encryption routines: if p < q then it would
18258 occasionally produce an invalid result. This will only happen with
18259 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18263 * Be less restrictive and allow also `perl util/perlpath.pl
18264 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18265 because this way one can also use an interpreter named `perl5` (which is
18266 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18267 installed as `perl`).
18269 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18271 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18273 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18275 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18276 advapi32.lib to Win32 build and change the pem test comparison
18277 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18278 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18279 and crypto/des/ede_cbcm_enc.c.
18283 * DES quad checksum was broken on big-endian architectures. Fixed.
18287 * Comment out two functions in bio.h that aren't implemented. Fix up the
18288 Win32 test batch file so it (might) work again. The Win32 test batch file
18289 is horrible: I feel ill....
18293 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18294 in e_os.h. Audit of header files to check ANSI and non ANSI
18295 sections: 10 functions were absent from non ANSI section and not exported
18296 from Windows DLLs. Fixed up libeay.num for new functions.
18300 * Make `openssl version` output lines consistent.
18302 *Ralf S. Engelschall*
18304 * Fix Win32 symbol export lists for BIO functions: Added
18305 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18306 to ms/libeay{16,32}.def.
18308 *Ralf S. Engelschall*
18310 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18311 fine under Unix and passes some trivial tests I've now added. But the
18312 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18313 added to make sure no one expects that this stuff really works in the
18314 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18315 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18318 *Ralf S. Engelschall*
18320 * Fix the generation of two part addresses in perl.
18322 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18324 * Add config entry for Linux on MIPS.
18326 *John Tobey <jtobey@channel1.com>*
18328 * Make links whenever Configure is run, unless we are on Windoze.
18332 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18333 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18338 * Add a useful kludge to allow package maintainers to specify compiler and
18339 other platforms details on the command line without having to patch the
18340 Configure script every time: One now can use
18341 `perl Configure <id>:<details>`,
18342 i.e. platform ids are allowed to have details appended
18343 to them (separated by colons). This is treated as there would be a static
18344 pre-configured entry in Configure's %table under key `<id>` with value
18345 `<details>` and `perl Configure <id>` is called. So, when you want to
18346 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18347 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18348 now, which overrides the FreeBSD-elf entry on-the-fly.
18350 *Ralf S. Engelschall*
18352 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18356 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18357 on the `perl Configure ...` command line. This way one can compile
18358 OpenSSL libraries with Position Independent Code (PIC) which is needed
18359 for linking it into DSOs.
18361 *Ralf S. Engelschall*
18363 * Remarkably, export ciphers were totally broken and no-one had noticed!
18368 * Cleaned up the LICENSE document: The official contact for any license
18369 questions now is the OpenSSL core team under openssl-core@openssl.org.
18370 And add a paragraph about the dual-license situation to make sure people
18371 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18372 to the OpenSSL toolkit.
18374 *Ralf S. Engelschall*
18376 * General source tree makefile cleanups: Made `making xxx in yyy...`
18377 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18378 Additionally cleaned up the `make links` target: Remove unnecessary
18379 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18380 to speed processing and no longer clutter the display with confusing
18381 stuff. Instead only the actually done links are displayed.
18383 *Ralf S. Engelschall*
18385 * Permit null encryption ciphersuites, used for authentication only. It used
18386 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18387 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18392 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18393 signed attributes when verifying signatures (this would break them),
18394 the detached data encoding was wrong and public keys obtained using
18395 X509_get_pubkey() weren't freed.
18399 * Add text documentation for the BUFFER functions. Also added a work around
18400 to a Win95 console bug. This was triggered by the password read stuff: the
18401 last character typed gets carried over to the next fread(). If you were
18402 generating a new cert request using 'req' for example then the last
18403 character of the passphrase would be CR which would then enter the first
18408 * Added the new 'Includes OpenSSL Cryptography Software' button as
18409 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18410 button and can be used by applications based on OpenSSL to show the
18411 relationship to the OpenSSL project.
18413 *Ralf S. Engelschall*
18415 * Remove confusing variables in function signatures in files
18416 ssl/ssl_lib.c and ssl/ssl.h.
18418 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18420 * Don't install bss_file.c under PREFIX/include/
18422 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18424 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18425 functions that return function pointers and has support for NT specific
18426 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18427 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18428 unsigned to signed types: this was killing the Win32 compile.
18432 * Add new certificate file to stack functions,
18433 SSL_add_dir_cert_subjects_to_stack() and
18434 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18435 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18436 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18437 This means that Apache-SSL and similar packages don't have to mess around
18438 to add as many CAs as they want to the preferred list.
18442 * Experiment with doxygen documentation. Currently only partially applied to
18444 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18445 openssl.doxy as the configuration file.
18449 * Get rid of remaining C++-style comments which strict C compilers hate.
18451 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18453 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18454 compiled in by default: it has problems with large keys.
18458 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18459 DH private keys and/or callback functions which directly correspond to
18460 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18461 is needed for applications which have to configure certificates on a
18462 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18464 For the RSA certificate situation is makes no difference, but
18465 for the DSA certificate situation this fixes the "no shared cipher"
18466 problem where the OpenSSL cipher selection procedure failed because the
18467 temporary keys were not overtaken from the context and the API provided
18468 no way to reconfigure them.
18469 The new functions now let applications reconfigure the stuff and they
18470 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18471 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18472 non-public-API function ssl_cert_instantiate() is used as a helper
18473 function and also to reduce code redundancy inside ssl_rsa.c.
18475 *Ralf S. Engelschall*
18477 * Move s_server -dcert and -dkey options out of the undocumented feature
18478 area because they are useful for the DSA situation and should be
18479 recognized by the users.
18481 *Ralf S. Engelschall*
18483 * Fix the cipher decision scheme for export ciphers: the export bits are
18484 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18485 SSL_EXP_MASK. So, the original variable has to be used instead of the
18486 already masked variable.
18488 *Richard Levitte <levitte@stacken.kth.se>*
18490 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18492 *Richard Levitte <levitte@stacken.kth.se>*
18494 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18495 from `int` to `unsigned int` because it is a length and initialized by
18496 EVP_DigestFinal() which expects an `unsigned int *`.
18498 *Richard Levitte <levitte@stacken.kth.se>*
18500 * Don't hard-code path to Perl interpreter on shebang line of Configure
18501 script. Instead use the usual Shell->Perl transition trick.
18503 *Ralf S. Engelschall*
18505 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18506 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18507 -noout -modulus` as it's already the case for `openssl rsa -noout
18508 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18509 currently the public key is printed (a decision which was already done by
18510 `openssl dsa -modulus` in the past) which serves a similar purpose.
18511 Additionally the NO_RSA no longer completely removes the whole -modulus
18512 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18515 *Ralf S. Engelschall*
18517 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18518 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18520 *Arne Ansper <arne@ats.cyber.ee>*
18522 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18523 to be added. Now both 'req' and 'ca' can use new objects defined in the
18528 * Add cool BIO that does syslog (or event log on NT).
18530 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18532 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18533 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18534 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18535 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18539 * Add preliminary config info for new extension code.
18543 * Make RSA_NO_PADDING really use no padding.
18545 *Ulf Moeller <ulf@fitug.de>*
18547 * Generate errors when private/public key check is done.
18551 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18552 for some CRL extensions and new objects added.
18556 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18557 key usage extension and fuller support for authority key id.
18561 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18562 padding method for RSA, which is recommended for new applications in PKCS
18563 #1 v2.0 (RFC 2437, October 1998).
18564 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18565 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18566 against Bleichbacher's attack on RSA.
18567 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18570 * Updates to the new SSL compression code
18572 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18574 * Fix so that the version number in the master secret, when passed
18575 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18576 (because the server will not accept higher), that the version number
18577 is 0x03,0x01, not 0x03,0x00
18579 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18581 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18582 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18583 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18587 * Support for RAW extensions where an arbitrary extension can be
18588 created by including its DER encoding. See `apps/openssl.cnf` for
18593 * Make sure latest Perl versions don't interpret some generated C array
18594 code as Perl array code in the crypto/err/err_genc.pl script.
18596 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18598 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18599 not many people have the assembler. Various Win32 compilation fixes and
18600 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18601 build instructions.
18605 * Modify configure script 'Configure' to automatically create crypto/date.h
18606 file under Win32 and also build pem.h from pem.org. New script
18607 util/mkfiles.pl to create the MINFO file on environments that can't do a
18608 'make files': perl util/mkfiles.pl >MINFO should work.
18612 * Major rework of DES function declarations, in the pursuit of correctness
18613 and purity. As a result, many evil casts evaporated, and some weirdness,
18614 too. You may find this causes warnings in your code. Zapping your evil
18615 casts will probably fix them. Mostly.
18619 * Fix for a typo in asn1.h. Bug fix to object creation script
18620 obj_dat.pl. It considered a zero in an object definition to mean
18621 "end of object": none of the objects in objects.h have any zeros
18622 so it wasn't spotted.
18624 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18626 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18627 Masking (CBCM). In the absence of test vectors, the best I have been able
18628 to do is check that the decrypt undoes the encrypt, so far. Send me test
18629 vectors if you have them.
18633 * Correct calculation of key length for export ciphers (too much space was
18634 allocated for null ciphers). This has not been tested!
18638 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18639 message is now correct (it understands "crypto" and "ssl" on its
18640 command line). There is also now an "update" option. This will update
18641 the util/ssleay.num and util/libeay.num files with any new functions.
18643 perl util/mkdef.pl crypto ssl update
18644 it will update them.
18648 * Overhauled the Perl interface:
18649 - ported BN stuff to OpenSSL's different BN library
18650 - made the perl/ source tree CVS-aware
18651 - renamed the package from SSLeay to OpenSSL (the files still contain
18652 their history because I've copied them in the repository)
18653 - removed obsolete files (the test scripts will be replaced
18654 by better Test::Harness variants in the future)
18656 *Ralf S. Engelschall*
18658 * First cut for a very conservative source tree cleanup:
18659 1. merge various obsolete readme texts into doc/ssleay.txt
18660 where we collect the old documents and readme texts.
18661 2. remove the first part of files where I'm already sure that we no
18662 longer need them because of three reasons: either they are just temporary
18663 files which were left by Eric or they are preserved original files where
18664 I've verified that the diff is also available in the CVS via "cvs diff
18665 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18666 the crypto/md/ stuff).
18668 *Ralf S. Engelschall*
18670 * More extension code. Incomplete support for subject and issuer alt
18671 name, issuer and authority key id. Change the i2v function parameters
18672 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18673 what that's for :-) Fix to ASN1 macro which messed up
18674 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18678 * Preliminary support for ENUMERATED type. This is largely copied from the
18683 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18685 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18687 * Make sure `make rehash` target really finds the `openssl` program.
18689 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18691 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18692 like to hear about it if this slows down other processors.
18696 * Add CygWin32 platform information to Configure script.
18698 *Alan Batie <batie@aahz.jf.intel.com>*
18700 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18702 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18704 * New program nseq to manipulate netscape certificate sequences
18708 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18713 * Fixes to BN code. Previously the default was to define BN_RECURSION
18714 but the BN code had some problems that would cause failures when
18715 doing certificate verification and some other functions.
18717 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18719 * Add ASN1 and PEM code to support netscape certificate sequences.
18723 * Add ASN1 and PEM code to support netscape certificate sequences.
18727 * Add several PKIX and private extended key usage OIDs.
18731 * Modify the 'ca' program to handle the new extension code. Modify
18732 openssl.cnf for new extension format, add comments.
18736 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18737 and add a sample to openssl.cnf so req -x509 now adds appropriate
18742 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18743 error code, add initial support to X509_print() and x509 application.
18747 * Takes a deep breath and start adding X509 V3 extension support code. Add
18748 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18749 stuff is currently isolated and isn't even compiled yet.
18753 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18754 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18755 Removed the versions check from X509 routines when loading extensions:
18756 this allows certain broken certificates that don't set the version
18757 properly to be processed.
18761 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18762 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18763 can still be regenerated with "make depend".
18767 * Spelling mistake in C version of CAST-128.
18769 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18771 * Changes to the error generation code. The perl script err-code.pl
18772 now reads in the old error codes and retains the old numbers, only
18773 adding new ones if necessary. It also only changes the .err files if new
18774 codes are added. The makefiles have been modified to only insert errors
18775 when needed (to avoid needlessly modifying header files). This is done
18776 by only inserting errors if the .err file is newer than the auto generated
18777 C file. To rebuild all the error codes from scratch (the old behaviour)
18778 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18779 or delete all the .err files.
18783 * CAST-128 was incorrectly implemented for short keys. The C version has
18784 been fixed, but is untested. The assembler versions are also fixed, but
18785 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18786 to regenerate it if needed.
18787 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18788 Hagino <itojun@kame.net>*
18790 * File was opened incorrectly in randfile.c.
18792 *Ulf Möller <ulf@fitug.de>*
18794 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18795 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18796 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18797 al: it's just almost always a UTCTime. Note this patch adds new error
18798 codes so do a "make errors" if there are problems.
18802 * Correct Linux 1 recognition in config.
18804 *Ulf Möller <ulf@fitug.de>*
18806 * Remove pointless MD5 hash when using DSA keys in ca.
18808 *Anonymous <nobody@replay.com>*
18810 * Generate an error if given an empty string as a cert directory. Also
18811 generate an error if handed NULL (previously returned 0 to indicate an
18812 error, but didn't set one).
18814 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18816 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18820 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18821 parameters. This was causing a warning which killed off the Win32 compile.
18825 * Remove C++ style comments from crypto/bn/bn_local.h.
18827 *Neil Costigan <neil.costigan@celocom.com>*
18829 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18830 based on a text string, looking up short and long names and finally
18831 "dot" format. The "dot" format stuff didn't work. Added new function
18832 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18833 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18834 OID is not part of the table.
18838 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18839 X509_LOOKUP_by_alias().
18843 * Sort openssl functions by name.
18847 * Get the `gendsa` command working and add it to the `list` command. Remove
18848 encryption from sample DSA keys (in case anyone is interested the password
18853 * Make *all* `*_free` functions accept a NULL pointer.
18855 *Frans Heymans <fheymans@isaserver.be>*
18857 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18860 *Anonymous <nobody@replay.com>*
18862 * s_server should send the CAfile as acceptable CAs, not its own cert.
18864 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18866 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18868 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18870 * Temp key "for export" tests were wrong in s3_srvr.c.
18872 *Anonymous <nobody@replay.com>*
18874 * Add prototype for temp key callback functions
18875 SSL_CTX_set_tmp_{rsa,dh}_callback().
18879 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18880 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18884 * X509_name_add_entry() freed the wrong thing after an error.
18886 *Arne Ansper <arne@ats.cyber.ee>*
18888 * rsa_eay.c would attempt to free a NULL context.
18890 *Arne Ansper <arne@ats.cyber.ee>*
18892 * BIO_s_socket() had a broken should_retry() on Windoze.
18894 *Arne Ansper <arne@ats.cyber.ee>*
18896 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18898 *Arne Ansper <arne@ats.cyber.ee>*
18900 * Make sure the already existing X509_STORE->depth variable is initialized
18901 in X509_STORE_new(), but document the fact that this variable is still
18902 unused in the certificate verification process.
18904 *Ralf S. Engelschall*
18906 * Fix the various library and `apps/` files to free up pkeys obtained from
18907 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18911 * Fix reference counting in X509_PUBKEY_get(). This makes
18912 demos/maurice/example2.c work, amongst others, probably.
18914 *Steve Henson and Ben Laurie*
18916 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18917 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18918 are no longer created. This way we have a single and consistent command
18919 line interface `openssl <command>`, similar to `cvs <command>`.
18921 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18923 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18924 BIT STRING wrapper always have zero unused bits.
18928 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18932 * Make the top-level INSTALL documentation easier to understand.
18936 * Makefiles updated to exit if an error occurs in a sub-directory
18937 make (including if user presses ^C) [Paul Sutton]
18939 * Make Montgomery context stuff explicit in RSA data structure.
18943 * Fix build order of pem and err to allow for generated pem.h.
18947 * Fix renumbering bug in X509_NAME_delete_entry().
18951 * Enhanced the err-ins.pl script so it makes the error library number
18952 global and can add a library name. This is needed for external ASN1 and
18953 other error libraries.
18957 * Fixed sk_insert which never worked properly.
18961 * Fix ASN1 macros so they can handle indefinite length constructed
18962 EXPLICIT tags. Some non standard certificates use these: they can now
18967 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18968 into a single doc/ssleay.txt bundle. This way the information is still
18969 preserved but no longer messes up this directory. Now it's new room for
18970 the new set of documentation files.
18972 *Ralf S. Engelschall*
18974 * SETs were incorrectly DER encoded. This was a major pain, because they
18975 shared code with SEQUENCEs, which aren't coded the same. This means that
18976 almost everything to do with SETs or SEQUENCEs has either changed name or
18977 number of arguments.
18979 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18981 * Fix test data to work with the above.
18985 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18986 was already fixed by Eric for 0.9.1 it seems.
18988 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18990 * Autodetect FreeBSD3.
18994 * Fix various bugs in Configure. This affects the following platforms:
18998 unixware-2.0-pentium
19003 * Eliminate generated files from CVS. Reorder tests to regenerate files
19004 before they are needed.
19008 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19012 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19014 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19015 changed SSLeay to OpenSSL in version strings.
19017 *Ralf S. Engelschall*
19019 * Some fixups to the top-level documents.
19023 * Fixed the nasty bug where rsaref.h was not found under compile-time
19024 because the symlink to include/ was missing.
19026 *Ralf S. Engelschall*
19028 * Incorporated the popular no-RSA/DSA-only patches
19029 which allow to compile a RSA-free SSLeay.
19031 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19033 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19034 when "ssleay" is still not found.
19036 *Ralf S. Engelschall*
19038 * Added more platforms to Configure: Cray T3E, HPUX 11,
19040 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19042 * Updated the README file.
19044 *Ralf S. Engelschall*
19046 * Added various .cvsignore files in the CVS repository subdirs
19047 to make a "cvs update" really silent.
19049 *Ralf S. Engelschall*
19051 * Recompiled the error-definition header files and added
19052 missing symbols to the Win32 linker tables.
19054 *Ralf S. Engelschall*
19056 * Cleaned up the top-level documents;
19057 o new files: CHANGES and LICENSE
19058 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19059 o merged COPYRIGHT into LICENSE
19060 o removed obsolete TODO file
19061 o renamed MICROSOFT to INSTALL.W32
19063 *Ralf S. Engelschall*
19065 * Removed dummy files from the 0.9.1b source tree:
19067 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19068 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19069 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19070 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19071 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19074 *Ralf S. Engelschall*
19076 * Added various platform portability fixes.
19080 * The Genesis of the OpenSSL rpject:
19081 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19082 Young and Tim J. Hudson created while they were working for C2Net until
19085 *The OpenSSL Project*
19087 ### Changes between 0.9.0b and 0.9.1b [not released]
19089 * Updated a few CA certificates under certs/
19093 * Changed some BIGNUM api stuff.
19097 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19098 DGUX x86, Linux Alpha, etc.
19102 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19103 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19108 * Add -strparse option to asn1pars program which parses nested
19111 *Dr Stephen Henson <shenson@bigfoot.com>*
19113 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19117 * DSA fix for "ca" program.
19121 * Added "-genkey" option to "dsaparam" program.
19125 * Added RIPE MD160 (rmd160) message digest.
19129 * Added -a (all) option to "ssleay version" command.
19133 * Added PLATFORM define which is the id given to Configure.
19137 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19141 * Extended the ASN.1 parser routines.
19145 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19149 * Added a BN_CTX to the BN library.
19153 * Fixed the weak key values in DES library
19157 * Changed API in EVP library for cipher aliases.
19161 * Added support for RC2/64bit cipher.
19165 * Converted the lhash library to the crypto/mem.c functions.
19169 * Added more recognized ASN.1 object ids.
19173 * Added more RSA padding checks for SSL/TLS.
19177 * Added BIO proxy/filter functionality.
19181 * Added extra_certs to SSL_CTX which can be used
19182 send extra CA certificates to the client in the CA cert chain sending
19183 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19187 * Now Fortezza is denied in the authentication phase because
19188 this is key exchange mechanism is not supported by SSLeay at all.
19192 * Additional PKCS1 checks.
19196 * Support the string "TLSv1" for all TLS v1 ciphers.
19200 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19201 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19205 * Fixed a few memory leaks.
19209 * Fixed various code and comment typos.
19213 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19214 bytes sent in the client random.
19216 *Edward Bishop <ebishop@spyglass.com>*
19220 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19221 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19222 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19223 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19224 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19225 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19226 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19227 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19228 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19229 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19230 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19231 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19232 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19233 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19234 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19235 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19236 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19237 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19238 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19239 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19240 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19241 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19242 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19243 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19244 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19245 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19246 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19247 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19248 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19249 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19250 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19251 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19252 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19253 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19254 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19255 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19256 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19257 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19258 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19259 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19260 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19261 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19262 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19263 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19264 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19265 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19266 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19267 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19268 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19269 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19270 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19271 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19272 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19273 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19274 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19275 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19276 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19277 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19278 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19279 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19280 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19281 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19282 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19283 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19284 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19285 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19286 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19287 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19288 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19289 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19290 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19291 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19292 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19293 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19294 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19295 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19296 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19297 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19298 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19299 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19300 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19301 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19302 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19303 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19304 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19305 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19306 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19307 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19308 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19309 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19310 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19311 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19312 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19313 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19314 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19315 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19316 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19317 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19318 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19319 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19320 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19321 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19322 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19323 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19324 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19325 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19326 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19327 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19328 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19329 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19330 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19331 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19332 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19333 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19334 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19335 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19336 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19337 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19338 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19339 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19340 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19341 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19342 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19343 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19344 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19345 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19346 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19347 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19348 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19349 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19350 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19351 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19352 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19353 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19354 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19355 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19356 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19357 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19358 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19359 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19360 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19361 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19362 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19363 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19364 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19365 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19366 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19367 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19368 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19369 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19370 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19371 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19372 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19373 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19374 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19375 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19376 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19377 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19378 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19379 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19380 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19381 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655