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 * The openssl commands that read keys, certificates, and CRLs now
55 automatically detect the PEM or DER format of the input files so it is not
56 necessary to explicitly specify the input format anymore. However if the
57 input format option is used the specified format will be required.
59 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
61 * Added enhanced PKCS#12 APIs which accept a library context `OSSL_LIB_CTX`
62 and (where relevant) a property query. Other APIs which handle PKCS#7 and
63 PKCS#8 objects have also been enhanced where required. This includes:
65 PKCS12_add_key_ex(), PKCS12_add_safe_ex(), PKCS12_add_safes_ex(),
66 PKCS12_create_ex(), PKCS12_decrypt_skey_ex(), PKCS12_init_ex(),
67 PKCS12_item_decrypt_d2i_ex(), PKCS12_item_i2d_encrypt_ex(),
68 PKCS12_key_gen_asc_ex(), PKCS12_key_gen_uni_ex(), PKCS12_key_gen_utf8_ex(),
69 PKCS12_pack_p7encdata_ex(), PKCS12_pbe_crypt_ex(), PKCS12_PBE_keyivgen_ex(),
70 PKCS12_SAFEBAG_create_pkcs8_encrypt_ex(), PKCS5_pbe2_set_iv_ex(),
71 PKCS5_pbe_set0_algor_ex(), PKCS5_pbe_set_ex(), PKCS5_pbkdf2_set_ex(),
72 PKCS5_v2_PBE_keyivgen_ex(), PKCS5_v2_scrypt_keyivgen_ex(),
73 PKCS8_decrypt_ex(), PKCS8_encrypt_ex(), PKCS8_set0_pbe_ex().
75 As part of this change the EVP_PBE_xxx APIs can also accept a library
76 context and property query and will call an extended version of the key/IV
77 derivation function which supports these parameters. This includes
78 EVP_PBE_CipherInit_ex(), EVP_PBE_find_ex() and EVP_PBE_scrypt_ex().
82 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
86 * Added support for Kernel TLS (KTLS). In order to use KTLS, support for it
87 must be compiled in using the "enable-ktls" compile time option. It must
88 also be enabled at run time using the SSL_OP_ENABLE_KTLS option.
90 *Boris Pismenny, John Baldwin and Andrew Gallatin*
92 * Support for RFC 5746 secure renegotiation is now required by default for
93 SSL or TLS connections to succeed. Applications that require the ability
94 to connect to legacy peers will need to explicitly set
95 SSL_OP_LEGACY_SERVER_CONNECT. Accordingly, SSL_OP_LEGACY_SERVER_CONNECT
96 is no longer set as part of SSL_OP_ALL.
100 * The signature of the `copy` functional parameter of the
101 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
102 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
103 the signature of the `pub_decode` functional parameter of the
104 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
105 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
109 * The error return values from some control calls (ctrl) have changed.
110 One significant change is that controls which used to return -2 for
111 invalid inputs, now return -1 indicating a generic error condition instead.
115 * A public key check is now performed during EVP_PKEY_derive_set_peer().
116 Previously DH was internally doing this during EVP_PKEY_derive().
117 To disable this check use EVP_PKEY_derive_set_peer_ex(dh, peer, 0). This
118 may mean that an error can occur in EVP_PKEY_derive_set_peer() rather than
119 during EVP_PKEY_derive().
123 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
124 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
125 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
126 are deprecated. They are not invoked by the OpenSSL library anymore and
127 are replaced by direct checks of the key operation against the key type
128 when the operation is initialized.
132 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
133 more key types including RSA, DSA, ED25519, X25519, ED448 and X448.
134 Previously (in 1.1.1) they would return -2. For key types that do not have
135 parameters then EVP_PKEY_param_check() will always return 1.
137 * The output from the command line applications may have minor
138 changes. These are primarily changes in capitalisation and white
139 space. However, in some cases, there are additional differences.
140 For example, the DH parameters output from `dhparam` now lists 'P',
141 'Q', 'G' and 'pcounter' instead of 'prime', 'generator', 'subgroup
142 order' and 'counter' respectively.
146 * The output from numerous "printing" functions such as X509_signature_print(),
147 X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been
148 amended such that there may be cosmetic differences between the output
149 observed in 1.1.1 and 3.0. This also applies to the "-text" output from the
150 x509 and crl applications.
154 * Windows thread synchronization uses read/write primitives (SRWLock) when
155 supported by the OS, otherwise CriticalSection continues to be used.
159 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
160 work on read only BIO source/sinks that do not support these functions.
161 This allows piping or redirection of a file BIO using stdin to be buffered
162 into memory. This is used internally in OSSL_DECODER_from_bio().
166 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
167 this function would return one of the values OSSL_STORE_INFO_NAME,
168 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
169 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
170 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
171 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
172 using this function should be amended to handle the changed return value.
176 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
177 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
178 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
179 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
180 contains more than one certificate identifier: This means that all
181 certificates referenced there MUST be part of the validation chain.
185 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
186 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
187 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
188 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
189 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
190 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
191 using the EVP APIs to access these ciphers should instead use more modern
192 ciphers. If that is not possible then these applications should ensure that
193 the legacy provider has been loaded. This can be achieved either
194 programmatically or via configuration. See the provider(7) man page for
199 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
200 RIPEMD-160 have been moved to the legacy provider. Applications using the
201 EVP APIs to access these digests should instead use more modern digests. If
202 that is not possible then these applications should ensure that the legacy
203 provider has been loaded. This can be achieved either programmatically or via
204 configuration. See the provider(7) man page for further details.
208 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
213 * The deprecated functions EVP_PKEY_get0_RSA(),
214 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
215 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
216 well as the similarly named "get1" functions behave slightly differently in
217 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
218 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
219 provider. Calling these functions will only return a handle on the internal
220 key where the EVP_PKEY was constructed using this key in the first place, for
221 example using a function or macro such as EVP_PKEY_assign_RSA(),
222 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
223 then these functions now return a cached copy of the key. Changes to
224 the internal provider key that take place after the first time the cached key
225 is accessed will not be reflected back in the cached copy. Similarly any
226 changes made to the cached copy by application code will not be reflected
227 back in the internal provider key.
229 For the above reasons the keys returned from these functions should typically
230 be treated as read-only. To emphasise this the value returned from
231 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
232 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
233 break some existing code. Applications broken by this change should be
234 modified. The preferred solution is to refactor the code to avoid the use of
235 these deprecated functions. Failing this the code should be modified to use a
236 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
237 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
238 non-const pointer to enable them to be "freed". However they should also be
239 treated as read-only.
243 * A number of functions handling low-level keys or engines were deprecated
244 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
245 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
246 EVP_PKEY_get0_siphash(). Applications using engines should instead use
247 providers. Applications getting or setting low-level keys in an EVP_PKEY
248 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
249 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
253 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
254 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
255 and require returning octet ptr parameters from providers that
256 would like to support them which complicates provider implementations.
260 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
261 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
262 RAND_set_rand_method(). Provider based random number generators should
263 be used instead via EVP_RAND(3).
267 * The SRP APIs have been deprecated. The old APIs do not work via providers,
268 and there is no EVP interface to them. Unfortunately there is no replacement
269 for these APIs at this time.
273 * Add a compile time option to prevent the caching of provider fetched
274 algorithms. This is enabled by including the no-cached-fetch option
275 at configuration time.
279 * The default algorithms for pkcs12 creation with the PKCS12_create() function
280 were changed to more modern PBKDF2 and AES based algorithms. The default
281 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
282 with the password-based encryption iteration count. The default digest
283 algorithm for the MAC computation was changed to SHA-256. The pkcs12
284 application now supports -legacy option that restores the previous
285 default algorithms to support interoperability with legacy systems.
287 *Tomáš Mráz and Sahana Prasad*
289 * The openssl speed command does not use low-level API calls anymore. This
290 implies some of the performance numbers might not be fully comparable
291 with the previous releases due to higher overhead. This applies
292 particularly to measuring performance on smaller data chunks.
296 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
299 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
301 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
302 Typically if OpenSSL has no EC or DH algorithms then it cannot support
303 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
304 through providers. Therefore third party providers may supply group
305 implementations even where there are no built-in ones. Attempting to create
306 TLS connections in such a build without also disabling TLSv1.3 at run time or
307 using third party provider groups may result in handshake failures. TLSv1.3
308 can be disabled at compile time using the "no-tls1_3" Configure option.
312 * The undocumented function X509_certificate_type() has been deprecated;
313 applications can use X509_get0_pubkey() and X509_get0_signature() to
314 get the same information.
318 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
319 functions. They are identical to BN_rand() and BN_rand_range()
324 * Removed RSA padding mode for SSLv23 (which was only used for
325 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
326 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
331 * Deprecated the obsolete X9.31 RSA key generation related functions
332 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
333 BN_X931_generate_prime_ex().
337 * The default key generation method for the regular 2-prime RSA keys was
338 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
339 Conditions Based on Auxiliary Probable Primes). This method is slower
340 than the original method.
344 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
345 They are replaced with the BN_check_prime() function that avoids possible
346 misuse and always uses at least 64 rounds of the Miller-Rabin
347 primality test. At least 64 rounds of the Miller-Rabin test are now also
348 used for all prime generation, including RSA key generation.
349 This increases key generation time, especially for larger keys.
353 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
354 as they are not useful with non-deprecated functions.
358 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
359 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
360 OCSP_REQ_CTX_i2d() and its special form OCSP_REQ_CTX_set1_req(),
361 OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
362 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
363 were used to collect all necessary data to form a HTTP request, and to
364 perform the HTTP transfer with that request. With OpenSSL 3.0, the
365 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
366 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
367 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
368 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
369 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
370 OSSL_HTTP_REQ_CTX_set_max_response_length().
372 *Rich Salz and Richard Levitte*
374 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`,
375 which are superseded by `X509_load_http()` and `X509_CRL_load_http()`.
379 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
383 * Validation of SM2 keys has been separated from the validation of regular EC
384 keys, allowing to improve the SM2 validation process to reject loaded private
385 keys that are not conforming to the SM2 ISO standard.
386 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
391 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
392 switches: a validation failure triggers an early exit, returning a failure
393 exit status to the parent process.
397 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
398 to ignore unknown ciphers.
402 * The `-cipher-commands` and `-digest-commands` options
403 of the command line utility `list` have been deprecated.
404 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
408 * All of the low-level EC_KEY functions have been deprecated including:
410 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
411 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
412 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
413 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
414 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
415 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
416 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
417 EC_KEY_METHOD_get_verify,
418 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
419 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
420 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
421 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
422 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
423 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
424 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
425 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
426 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
427 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
428 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
429 Applications that need to implement an EC_KEY_METHOD need to consider
430 implementation of the functionality in a special provider.
431 For replacement of the functions manipulating the EC_KEY objects
432 see the EVP_PKEY-EC(7) manual page.
434 Additionally functions that read and write EC_KEY objects such as
435 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
436 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
437 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
438 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
439 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
440 have also been deprecated. Applications should instead use the
441 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
443 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
444 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
445 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
446 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
447 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
449 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
451 * Deprecated all the libcrypto and libssl error string loading
452 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
453 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
454 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
455 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
456 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
457 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
458 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
459 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
460 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
461 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
463 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
464 now loads error strings automatically.
468 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
469 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
470 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
471 are to be used by servers requiring ephemeral DH keys. Instead applications
472 should consider using the built-in DH parameters that are available by
473 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
474 necessary then applications can use the alternative functions
475 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
476 replacement for the "callback" functions. The callback was originally useful
477 in order to have different parameters for export and non-export ciphersuites.
478 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
479 functions should be replaced by one of the other methods described above.
483 * The `-crypt` option to the `passwd` command line tool has been removed.
487 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
492 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
494 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
495 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
496 The inverse ciphers use AES decryption for wrapping, and
497 AES encryption for unwrapping.
501 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
502 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
503 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
504 OpenSSL 3.0 these are replaced by the more generic functions
505 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
506 The old versions have been converted to deprecated macros that just call the
511 * The security callback, which can be customised by application code, supports
512 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
513 in the "other" parameter. In most places this is what is passed. All these
514 places occur server side. However there was one client side call of this
515 security operation and it passed a DH object instead. This is incorrect
516 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
517 of the other locations. Therefore this client side call has been changed to
518 pass an EVP_PKEY instead.
522 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
523 interface. Their functionality remains unchanged.
527 * Added new option for 'openssl list', '-providers', which will display the
528 list of loaded providers, their names, version and status. It optionally
529 displays their gettable parameters.
533 * Removed EVP_PKEY_set_alias_type(). This function was previously
534 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
535 type is internally recognised so the workaround is no longer needed.
537 This is a breaking change from previous OpenSSL versions.
541 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
542 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
546 * Changed all "STACK" functions to be macros instead of inline functions. Macro
547 parameters are still checked for type safety at compile time via helper
552 * Remove the RAND_DRBG API
554 The RAND_DRBG API did not fit well into the new provider concept as
555 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
556 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
557 and some of its API calls are rather low-level. This holds in particular
558 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
560 Adding a compatibility layer to continue supporting the RAND_DRBG API as
561 a legacy API for a regular deprecation period turned out to come at the
562 price of complicating the new provider API unnecessarily. Since the
563 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
566 *Paul Dale and Matthias St. Pierre*
568 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
569 as well as actual hostnames.
573 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
574 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
575 conversely, silently ignore DTLS protocol version bounds when configuring
576 TLS-based contexts. The commands can be repeated to set bounds of both
577 types. The same applies with the corresponding "min_protocol" and
578 "max_protocol" command-line switches, in case some application uses both TLS
581 SSL_CTX instances that are created for a fixed protocol version (e.g.
582 `TLSv1_server_method()`) also silently ignore version bounds. Previously
583 attempts to apply bounds to these protocol versions would result in an
584 error. Now only the "version-flexible" SSL_CTX instances are subject to
585 limits in configuration files in command-line options.
589 * Deprecated the `ENGINE` API. Engines should be replaced with providers
594 * Reworked the recorded ERR codes to make better space for system errors.
595 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
596 given code is a system error (true) or an OpenSSL error (false).
600 * Reworked the test perl framework to better allow parallel testing.
602 *Nicola Tuveri and David von Oheimb*
604 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
605 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
609 * 'Configure' has been changed to figure out the configuration target if
610 none is given on the command line. Consequently, the 'config' script is
611 now only a mere wrapper. All documentation is changed to only mention
614 *Rich Salz and Richard Levitte*
616 * Added a library context `OSSL_LIB_CTX` that applications as well as
617 other libraries can use to form a separate context within which
618 libcrypto operations are performed.
620 There are two ways this can be used:
622 - Directly, by passing a library context to functions that take
623 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
625 - Indirectly, by creating a new library context and then assigning
626 it as the new default, with `OSSL_LIB_CTX_set0_default`.
628 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
629 apart from the functions directly related to `OSSL_LIB_CTX`, accept
630 NULL to indicate that the default library context should be used.
632 Library code that changes the default library context using
633 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
634 second call before returning to the caller.
636 _(Note: the library context was initially called `OPENSSL_CTX` and
637 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
641 * Handshake now fails if Extended Master Secret extension is dropped
646 * Dropped interactive mode from the `openssl` program. From now on,
647 running it without arguments is equivalent to `openssl help`.
651 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()` since their
652 return values were confusing: Unlike other `_cmp()` functions
653 they do not return 0 when their arguments are equal.
654 The new replacement functions `EVP_PKEY_eq()` and `EVP_PKEY_parameters_eq()`
657 *David von Oheimb and Shane Lontis*
659 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
660 `EC_GROUP_get_field_type()`.
664 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
665 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
666 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
667 Applications should rely on the library automatically assigning a suitable
668 EC_METHOD internally upon EC_GROUP construction.
672 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
673 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
674 assigned internally without application intervention.
675 Users of EC_GROUP_new() should switch to a different suitable constructor.
679 * Add CAdES-BES signature verification support, mostly derived
680 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
682 *Filipe Raimundo da Silva*
684 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
688 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
689 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
690 is to support encryption and decryption of a digital envelope that is both
691 authenticated and encrypted using AES GCM mode.
695 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
696 functions are not widely used and now OpenSSL automatically perform this
697 conversion when needed.
701 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
702 EC_KEY_precompute_mult(). These functions are not widely used and
703 applications should instead switch to named curves which OpenSSL has
704 hardcoded lookup tables for.
708 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
709 should instead use the L<EC_POINT_mul(3)> function.
713 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
714 that are not applicable to the new provider model. Applications should
715 instead use EVP_default_properties_is_fips_enabled() and
716 EVP_default_properties_enable_fips().
720 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
721 is set, an unexpected EOF is ignored, it pretends a close notify was received
722 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
726 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
727 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
728 used and applications should instead use the
729 L<EC_POINT_set_affine_coordinates(3)> and
730 L<EC_POINT_get_affine_coordinates(3)> functions.
734 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
735 arrays to be more easily constructed via a series of utility functions.
736 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
737 the various push functions and finally convert to a passable OSSL_PARAM
738 array using OSSL_PARAM_BLD_to_param().
742 * The security strength of SHA1 and MD5 based signatures in TLS has been
743 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
744 working at the default security level of 1 and instead requires security
745 level 0. The security level can be changed either using the cipher string
746 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
747 that where the signature algorithms extension is missing from a ClientHello
748 then the handshake will fail in TLS 1.2 at security level 1. This is because,
749 although this extension is optional, failing to provide one means that
750 OpenSSL will fallback to a default set of signature algorithms. This default
751 set requires the availability of SHA1.
755 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
756 contain a provider side internal key.
760 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
761 They are old functions that we don't use, and that you could disable with
762 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
766 * Project text documents not yet having a proper file name extension
767 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
768 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
772 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
773 have been converted to Markdown with the goal to produce documents
774 which not only look pretty when viewed online in the browser, but
775 remain well readable inside a plain text editor.
777 To achieve this goal, a 'minimalistic' Markdown style has been applied
778 which avoids formatting elements that interfere too much with the
779 reading flow in the text file. For example, it
781 * avoids [ATX headings][] and uses [setext headings][] instead
782 (which works for `<h1>` and `<h2>` headings only).
783 * avoids [inline links][] and uses [reference links][] instead.
784 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
786 [ATX headings]: https://github.github.com/gfm/#atx-headings
787 [setext headings]: https://github.github.com/gfm/#setext-headings
788 [inline links]: https://github.github.com/gfm/#inline-link
789 [reference links]: https://github.github.com/gfm/#reference-link
790 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
791 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
793 *Matthias St. Pierre*
795 * The test suite is changed to preserve results of each test recipe.
796 A new directory test-runs/ with subdirectories named like the
797 test recipes are created in the build tree for this purpose.
801 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
802 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
803 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
805 *David von Oheimb, Martin Peylo*
807 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
808 The legacy OCSP-focused and only partly documented API is retained for
809 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
813 * Added `util/check-format.pl`, a tool for checking adherence to the
814 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
815 The checks performed are incomplete and yield some false positives.
816 Still the tool should be useful for detecting most typical glitches.
820 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
821 If domain name resolution yields multiple IP addresses all of them are tried
822 after `connect()` failures.
826 * All of the low-level RSA functions have been deprecated including:
828 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
829 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
830 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
831 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
832 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
833 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
834 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
835 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
836 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
837 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
838 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
839 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
840 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
841 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
842 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
843 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
844 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
845 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
846 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
847 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
848 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
849 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
850 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
851 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
852 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
853 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
854 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
855 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
857 Use of these low-level functions has been informally discouraged for a long
858 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
859 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
860 L<EVP_PKEY_decrypt(3)>.
862 All of these low-level RSA functions have been deprecated without
865 RSA_blinding_off, RSA_blinding_on, RSA_clear_flags, RSA_get_version,
866 RSAPrivateKey_dup, RSAPublicKey_dup, RSA_set_flags, RSA_setup_blinding and
869 All of these RSA flags have been deprecated without replacement:
871 RSA_FLAG_BLINDING, RSA_FLAG_CACHE_PRIVATE, RSA_FLAG_CACHE_PUBLIC,
872 RSA_FLAG_EXT_PKEY, RSA_FLAG_NO_BLINDING, RSA_FLAG_THREAD_SAFE and
873 RSA_METHOD_FLAG_NO_CHECK.
877 * X509 certificates signed using SHA1 are no longer allowed at security
879 In TLS/SSL the default security level is 1. It can be set either
880 using the cipher string with `@SECLEVEL`, or calling
881 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
882 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
884 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
885 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
886 options of the commands.
890 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
891 modified to use PKEY APIs. These commands are now in maintenance mode
892 and no new features will be added to them.
896 * The command line utility rsautl has been deprecated.
897 Instead use the pkeyutl program.
901 * The command line utilities genrsa and rsa have been modified to use PKEY
902 APIs. They now write PKCS#8 keys by default. These commands are now in
903 maintenance mode and no new features will be added to them.
907 * All of the low-level DH functions have been deprecated including:
909 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
910 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
911 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
912 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
913 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
914 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
915 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
916 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
917 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
918 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
919 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
920 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
921 DH_meth_get_generate_params and DH_meth_set_generate_params.
923 Use of these low-level functions has been informally discouraged for a long
924 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
925 and L<EVP_PKEY_derive(3)>.
927 These low-level DH functions have been deprecated without replacement:
929 DH_clear_flags, DH_get_1024_160, DH_get_2048_224, DH_get_2048_256,
930 DH_set_flags and DH_test_flags.
932 The DH_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
933 The DH_FLAG_TYPE_DH and DH_FLAG_TYPE_DHX have been deprecated. Use
934 EVP_PKEY_is_a() to determine the type of a key. There is no replacement for
937 Additionally functions that read and write DH objects such as d2i_DHparams,
938 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
939 functions have also been deprecated. Applications should instead use the
940 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
942 Finally functions that assign or obtain DH objects from an EVP_PKEY such as
943 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
944 `EVP_PKEY_set1_DH()` are also deprecated.
945 Applications should instead either read or write an
946 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
947 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
949 *Paul Dale and Matt Caswell*
951 * All of the low-level DSA functions have been deprecated including:
953 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
954 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
955 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
956 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
957 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
958 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
959 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
960 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
961 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
962 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
963 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
964 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
965 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
966 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
967 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
968 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
969 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
971 Use of these low-level functions has been informally discouraged for a long
972 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
973 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
975 These low-level DSA functions have been deprecated without replacement:
977 DSA_clear_flags, DSA_dup_DH, DSAparams_dup, DSA_set_flags and
980 The DSA_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
982 Finally functions that assign or obtain DSA objects from an EVP_PKEY such as
983 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
984 `EVP_PKEY_set1_DSA()` are also deprecated.
985 Applications should instead either read or write an
986 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
987 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
991 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
992 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. This is a breaking
993 change from previous OpenSSL versions.
995 Unlike in previous OpenSSL versions, this means that applications must not
996 call `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
997 The `EVP_PKEY_set_alias_type` function has now been removed.
999 Parameter and key generation is also reworked to make it possible
1000 to generate EVP_PKEY_SM2 parameters and keys. Applications must now generate
1001 SM2 keys directly and must not create an EVP_PKEY_EC key first.
1005 * Deprecated low-level ECDH and ECDSA functions. These include:
1007 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
1008 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
1011 Use of these low-level functions has been informally discouraged for a long
1012 time. Instead applications should use the EVP_PKEY_derive(3),
1013 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
1017 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
1018 and EVP_PKEY_decrypt() instead.
1019 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
1020 and EVP_PKEY_encrypt() instead.
1024 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
1025 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
1026 a new formulation to include all the things it can be used for,
1027 as well as words of caution.
1031 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1032 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
1036 * All low level HMAC functions except for HMAC have been deprecated including:
1038 HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
1039 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
1040 and HMAC_CTX_get_md.
1042 Use of these low-level functions has been informally discouraged for a long
1043 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1044 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1045 and L<EVP_MAC_final(3)> or the single-shot MAC function L<EVP_Q_mac(3)>.
1047 *Paul Dale and David von Oheimb*
1049 * Over two thousand fixes were made to the documentation, including:
1050 - Common options (such as -rand/-writerand, TLS version control, etc)
1051 were refactored and point to newly-enhanced descriptions in openssl.pod.
1052 - Added style conformance for all options (with help from Richard Levitte),
1053 documented all reported missing options, added a CI build to check
1054 that all options are documented and that no unimplemented options
1056 - Documented some internals, such as all use of environment variables.
1057 - Addressed all internal broken L<> references.
1061 * All of the low-level CMAC functions have been deprecated including:
1063 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
1064 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
1066 Use of these low-level functions has been informally discouraged for a long
1067 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1068 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1069 and L<EVP_MAC_final(3)>.
1073 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
1074 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
1077 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
1078 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
1079 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
1080 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
1081 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
1082 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform,
1083 SHA256_Init, SHA256_Update, SHA256_Final, SHA256_Transform,
1084 SHA384_Init, SHA384_Update, SHA384_Final,
1085 SHA512_Init, SHA512_Update, SHA512_Final, SHA512_Transform,
1086 WHIRLPOOL, WHIRLPOOL_Init,
1087 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
1089 Use of these low-level functions has been informally discouraged
1090 for a long time. Applications should use the L<EVP_DigestInit_ex(3)>,
1091 L<EVP_DigestUpdate(3)>, and L<EVP_DigestFinal_ex(3)> functions instead.
1092 Alternatively, the quick one-shot function L<EVP_Q_digest(3)> can be used.
1093 SHA1, SHA224, SHA256, SHA384 and SHA512 have changed from functions to macros
1094 like this: (EVP_Q_digest(NULL, "SHA256", NULL, d, n, md, NULL) ? md : NULL).
1096 *Paul Dale and David von Oheimb*
1098 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1099 set of functions. The documentation mentioned negative values for some
1100 errors, but this was never the case, so the mention of negative values
1103 Code that followed the documentation and thereby check with something
1104 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1108 * All of the low-level cipher functions have been deprecated including:
1110 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
1111 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
1112 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
1113 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
1114 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
1115 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
1116 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
1117 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
1118 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
1119 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
1120 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
1121 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
1122 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
1123 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
1124 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
1125 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
1126 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
1127 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
1128 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
1129 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
1130 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
1131 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
1132 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
1133 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
1134 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
1135 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
1136 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
1137 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
1138 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
1140 Use of these low-level functions has been informally discouraged for
1141 a long time. Applications should use the high level EVP APIs, e.g.
1142 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
1143 equivalently named decrypt functions instead.
1145 *Matt Caswell and Paul Dale*
1147 * Removed include/openssl/opensslconf.h.in and replaced it with
1148 include/openssl/configuration.h.in, which differs in not including
1149 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1150 was added to include both.
1152 This allows internal hacks where one might need to modify the set
1153 of configured macros, for example this if deprecated symbols are
1154 still supposed to be available internally:
1156 #include <openssl/configuration.h>
1158 #undef OPENSSL_NO_DEPRECATED
1159 #define OPENSSL_SUPPRESS_DEPRECATED
1161 #include <openssl/macros.h>
1163 This should not be used by applications that use the exported
1164 symbols, as that will lead to linking errors.
1168 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1169 used in exponentiation with 512-bit moduli. No EC algorithms are
1170 affected. Analysis suggests that attacks against 2-prime RSA1024,
1171 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1172 difficult to perform and are not believed likely. Attacks against DH512
1173 are considered just feasible. However, for an attack the target would
1174 have to re-use the DH512 private key, which is not recommended anyway.
1175 Also applications directly using the low-level API BN_mod_exp may be
1176 affected if they use BN_FLG_CONSTTIME.
1181 * Most memory-debug features have been deprecated, and the functionality
1182 replaced with no-ops.
1186 * Added documentation for the STACK API.
1190 * Introduced a new method type and API, OSSL_ENCODER, to
1191 represent generic encoders. An implementation is expected to
1192 be able to encode an object associated with a given name (such
1193 as an algorithm name for an asymmetric key) into forms given by
1194 implementation properties.
1196 Encoders are primarily used from inside libcrypto, through
1197 calls to functions like EVP_PKEY_print_private(),
1198 PEM_write_bio_PrivateKey() and similar.
1200 Encoders are specified in such a way that they can be made to
1201 directly handle the provider side portion of an object, if this
1202 provider side part comes from the same provider as the encoder
1203 itself, but can also be made to handle objects in parametrized
1204 form (as an OSSL_PARAM array of data). This allows a provider to
1205 offer generic encoders as a service for any other provider.
1209 * Added a .pragma directive to the syntax of configuration files, to
1210 allow varying behavior in a supported and predictable manner.
1211 Currently added pragma:
1215 This allows dollar signs to be a keyword character unless it's
1216 followed by a opening brace or parenthesis. This is useful for
1217 platforms where dollar signs are commonly used in names, such as
1218 volume names and system directory names on VMS.
1222 * Added functionality to create an EVP_PKEY from user data. This
1223 is effectively the same as creating a RSA, DH or DSA object and
1224 then assigning them to an EVP_PKEY, but directly using algorithm
1225 agnostic EVP functions. A benefit is that this should be future
1226 proof for public key algorithms to come.
1230 * Change the interpretation of the '--api' configuration option to
1231 mean that this is a desired API compatibility level with no
1232 further meaning. The previous interpretation, that this would
1233 also mean to remove all deprecated symbols up to and including
1234 the given version, no requires that 'no-deprecated' is also used
1235 in the configuration.
1237 When building applications, the desired API compatibility level
1238 can be set with the OPENSSL_API_COMPAT macro like before. For
1239 API compatibility version below 3.0, the old style numerical
1240 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1241 For version 3.0 and on, the value is expected to be the decimal
1242 value calculated from the major and minor version like this:
1244 MAJOR * 10000 + MINOR * 100
1248 -DOPENSSL_API_COMPAT=30000 For 3.0
1249 -DOPENSSL_API_COMPAT=30200 For 3.2
1251 To hide declarations that are deprecated up to and including the
1252 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1253 given when building the application as well.
1257 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1258 access to certificate and CRL stores via URIs and OSSL_STORE
1261 This adds the following functions:
1263 - X509_LOOKUP_store()
1264 - X509_STORE_load_file()
1265 - X509_STORE_load_path()
1266 - X509_STORE_load_store()
1267 - SSL_add_store_cert_subjects_to_stack()
1268 - SSL_CTX_set_default_verify_store()
1269 - SSL_CTX_load_verify_file()
1270 - SSL_CTX_load_verify_dir()
1271 - SSL_CTX_load_verify_store()
1275 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1276 The presence of this system service is determined at run-time.
1280 * Added functionality to create an EVP_PKEY context based on data
1281 for methods from providers. This takes an algorithm name and a
1282 property query string and simply stores them, with the intent
1283 that any operation that uses this context will use those strings
1284 to fetch the needed methods implicitly, thereby making the port
1285 of application written for pre-3.0 OpenSSL easier.
1289 * The undocumented function NCONF_WIN32() has been deprecated; for
1290 conversion details see the HISTORY section of doc/man5/config.pod
1294 * Introduced the new functions EVP_DigestSignInit_ex() and
1295 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1296 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1297 pages for further details.
1301 * Over two thousand fixes were made to the documentation, including:
1302 adding missing command flags, better style conformance, documentation
1305 *Rich Salz, Richard Levitte*
1307 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1308 X25519, X448, Ed25519 and Ed448.
1312 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1317 * Deprecated the public definition of `ERR_STATE` as well as the function
1318 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1323 * Added ERR functionality to give callers access to the stored function
1324 names that have replaced the older function code based functions.
1326 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1327 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1328 ERR_peek_error_all() and ERR_peek_last_error_all().
1330 These functions have become deprecated: ERR_get_error_line(),
1331 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1332 ERR_peek_last_error_line_data() and ERR_func_error_string().
1334 Users are recommended to use ERR_get_error_all(), or to pick information
1335 with ERR_peek functions and finish off with getting the error code by using
1340 * Extended testing to be verbose for failing tests only. The make variables
1341 VERBOSE_FAILURE or VF can be used to enable this:
1343 $ make VF=1 test # Unix
1344 $ mms /macro=(VF=1) test ! OpenVMS
1345 $ nmake VF=1 test # Windows
1349 * Added the `-copy_extensions` option to the `x509` command for use with
1350 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1351 all extensions in the request are copied to the certificate or vice versa.
1353 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1355 * Added the `-copy_extensions` option to the `req` command for use with
1356 `-x509`. When given with the `copy` or `copyall` argument,
1357 all extensions in the certification request are copied to the certificate.
1361 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1362 they generate are by default RFC 5280 compliant in the following sense:
1363 There is a subjectKeyIdentifier extension with a hash value of the public key
1364 and for not self-signed certs there is an authorityKeyIdentifier extension
1365 with a keyIdentifier field or issuer information identifying the signing key.
1366 This is done unless some configuration overrides the new default behavior,
1367 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1371 * Added several checks to `X509_verify_cert()` according to requirements in
1372 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1373 (which may be done by using the CLI option `-x509_strict`):
1374 * The basicConstraints of CA certificates must be marked critical.
1375 * CA certificates must explicitly include the keyUsage extension.
1376 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1377 * The issuer name of any certificate must not be empty.
1378 * The subject name of CA certs, certs with keyUsage crlSign,
1379 and certs without subjectAlternativeName must not be empty.
1380 * If a subjectAlternativeName extension is given it must not be empty.
1381 * The signatureAlgorithm field and the cert signature must be consistent.
1382 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1383 must not be marked critical.
1384 * The authorityKeyIdentifier must be given for X.509v3 certs
1385 unless they are self-signed.
1386 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1390 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1391 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1395 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1396 used even when parsing explicit parameters, when loading a encoded key
1397 or calling `EC_GROUP_new_from_ecpkparameters()`/
1398 `EC_GROUP_new_from_ecparameters()`.
1399 This prevents bypass of security hardening and performance gains,
1400 especially for curves with specialized EC_METHODs.
1401 By default, if a key encoded with explicit parameters is loaded and later
1402 encoded, the output is still encoded with explicit parameters, even if
1403 internally a "named" EC_GROUP is used for computation.
1407 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1408 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1409 NULL. After this change, only the cofactor parameter can be NULL. It also
1410 does some minimal sanity checks on the passed order.
1415 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1416 An attack is simple, if the first CMS_recipientInfo is valid but the
1417 second CMS_recipientInfo is chosen ciphertext. If the second
1418 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1419 encryption key will be replaced by garbage, and the message cannot be
1420 decoded, but if the RSA decryption fails, the correct encryption key is
1421 used and the recipient will not notice the attack.
1422 As a work around for this potential attack the length of the decrypted
1423 key must be equal to the cipher default key length, in case the
1424 certifiate is not given and all recipientInfo are tried out.
1425 The old behaviour can be re-enabled in the CMS code by setting the
1426 CMS_DEBUG_DECRYPT flag.
1430 * Early start up entropy quality from the DEVRANDOM seed source has been
1431 improved for older Linux systems. The RAND subsystem will wait for
1432 /dev/random to be producing output before seeding from /dev/urandom.
1433 The seeded state is stored for future library initialisations using
1434 a system global shared memory segment. The shared memory identifier
1435 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1436 the desired value. The default identifier is 114.
1440 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1441 when primes for RSA keys are computed.
1442 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1443 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1444 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1445 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1446 This avoids possible fingerprinting of newly generated RSA modules.
1450 * Correct the extended master secret constant on EBCDIC systems. Without this
1451 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1452 negotiate EMS will fail. Unfortunately this also means that TLS connections
1453 between EBCDIC systems with this fix, and EBCDIC systems without this
1454 fix will fail if they negotiate EMS.
1458 * Changed the library initialisation so that the config file is now loaded
1459 by default. This was already the case for libssl. It now occurs for both
1460 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1461 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1465 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1466 where the former acts as a replacement for `ERR_put_error()`, and the
1467 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1468 `ERR_raise_data()` adds more flexibility by taking a format string and
1469 an arbitrary number of arguments following it, to be processed with
1474 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1475 to check if a named provider is loaded and available. When called, it
1476 will also activate all fallback providers if such are still present.
1480 * Enforce a minimum DH modulus size of 512 bits.
1484 * Changed DH parameters to generate the order q subgroup instead of 2q.
1485 Previously generated DH parameters are still accepted by DH_check
1486 but DH_generate_key works around that by clearing bit 0 of the
1487 private key for those. This avoids leaking bit 0 of the private key.
1491 * Significantly reduce secure memory usage by the randomness pools.
1495 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1500 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1501 algorithms. An implementation of a key exchange algorithm can be obtained
1502 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1503 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1504 the older EVP_PKEY_derive_init() function. See the man pages for the new
1505 functions for further details.
1509 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1513 * Removed the function names from error messages and deprecated the
1516 * Removed NextStep support and the macro OPENSSL_UNISTD
1520 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1521 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1522 Also removed "export var as function" capability; we do not export
1523 variables, only functions.
1527 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1528 an error and 1 indicating success. In previous versions of OpenSSL this
1529 was a void type. If a key was set longer than the maximum possible this
1534 * Support SM2 signing and verification schemes with X509 certificate.
1538 * Use SHA256 as the default digest for TS query in the `ts` app.
1542 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1543 This checks that the salt length is at least 128 bits, the derived key
1544 length is at least 112 bits, and that the iteration count is at least 1000.
1545 For backwards compatibility these checks are disabled by default in the
1546 default provider, but are enabled by default in the fips provider.
1547 To enable or disable these checks use the control
1548 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1552 * Default cipher lists/suites are now available via a function, the
1553 #defines are deprecated.
1557 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1558 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1559 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1563 * Join the directories crypto/x509 and crypto/x509v3
1567 * Added command 'openssl kdf' that uses the EVP_KDF API.
1571 * Added command 'openssl mac' that uses the EVP_MAC API.
1575 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1576 as default directories. Also added the command 'openssl info'
1577 for scripting purposes.
1581 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1582 deprecated. These undocumented functions were never integrated into the EVP
1583 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1584 Bi-directional IGE mode. These modes were never formally standardised and
1585 usage of these functions is believed to be very small. In particular
1586 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1587 is ever used. The security implications are believed to be minimal, but
1588 this issue was never fixed for backwards compatibility reasons. New code
1589 should not use these modes.
1593 * Add prediction resistance to the DRBG reseeding process.
1597 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1598 mandated by IEEE Std 1619-2018.
1602 * Added newline escaping functionality to a filename when using openssl dgst.
1603 This output format is to replicate the output format found in the `*sum`
1604 checksum programs. This aims to preserve backward compatibility.
1606 *Matt Eaton, Richard Levitte, and Paul Dale*
1608 * Removed the heartbeat message in DTLS feature, as it has very
1609 little usage and doesn't seem to fulfill a valuable purpose.
1610 The configuration option is now deprecated.
1614 * Changed the output of 'openssl {digestname} < file' to display the
1615 digest name in its output.
1619 * Added a new generic trace API which provides support for enabling
1620 instrumentation through trace output. This feature is mainly intended
1621 as an aid for developers and is disabled by default. To utilize it,
1622 OpenSSL needs to be configured with the `enable-trace` option.
1624 If the tracing API is enabled, the application can activate trace output
1625 by registering BIOs as trace channels for a number of tracing and debugging
1628 The `openssl` program has been expanded to enable any of the types
1629 available via environment variables defined by the user, and serves as
1630 one possible example on how to use this functionality.
1632 *Richard Levitte & Matthias St. Pierre*
1634 * Added build tests for C++. These are generated files that only do one
1635 thing, to include one public OpenSSL head file each. This tests that
1636 the public header files can be usefully included in a C++ application.
1638 This test isn't enabled by default. It can be enabled with the option
1639 'enable-buildtest-c++'.
1643 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1647 * Add KMAC to EVP_MAC.
1651 * Added property based algorithm implementation selection framework to
1656 * Added SCA hardening for modular field inversion in EC_GROUP through
1657 a new dedicated field_inv() pointer in EC_METHOD.
1658 This also addresses a leakage affecting conversions from projective
1659 to affine coordinates.
1661 *Billy Bob Brumley, Nicola Tuveri*
1663 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1664 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1665 those algorithms that were already supported through the EVP_PKEY API
1666 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1667 and scrypt are now wrappers that call EVP_KDF.
1671 * Build devcrypto engine as a dynamic engine.
1673 *Eneas U de Queiroz*
1675 * Add keyed BLAKE2 to EVP_MAC.
1679 * Fix a bug in the computation of the endpoint-pair shared secret used
1680 by DTLS over SCTP. This breaks interoperability with older versions
1681 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1682 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1683 interoperability with such broken implementations. However, enabling
1684 this switch breaks interoperability with correct implementations.
1686 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1687 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1691 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1695 * Changed the license to the Apache License v2.0.
1699 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1701 - Major releases (indicated by incrementing the MAJOR release number)
1702 may introduce incompatible API/ABI changes.
1703 - Minor releases (indicated by incrementing the MINOR release number)
1704 may introduce new features but retain API/ABI compatibility.
1705 - Patch releases (indicated by incrementing the PATCH number)
1706 are intended for bug fixes and other improvements of existing
1707 features only (like improving performance or adding documentation)
1708 and retain API/ABI compatibility.
1712 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1716 * Remove the 'dist' target and add a tarball building script. The
1717 'dist' target has fallen out of use, and it shouldn't be
1718 necessary to configure just to create a source distribution.
1722 * Recreate the OS390-Unix config target. It no longer relies on a
1723 special script like it did for OpenSSL pre-1.1.0.
1727 * Instead of having the source directories listed in Configure, add
1728 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1733 * Add GMAC to EVP_MAC.
1737 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1741 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1742 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1743 to facilitate the continued use of MACs through raw private keys in
1744 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1748 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1749 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1753 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1754 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1755 are retained for backwards compatibility.
1759 * AES-XTS mode now enforces that its two keys are different to mitigate
1760 the attacked described in "Efficient Instantiations of Tweakable
1761 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1762 Details of this attack can be obtained from:
1763 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1767 * Rename the object files, i.e. give them other names than in previous
1768 versions. Their names now include the name of the final product, as
1769 well as its type mnemonic (bin, lib, shlib).
1773 * Added new option for 'openssl list', '-objects', which will display the
1774 list of built in objects, i.e. OIDs with names.
1778 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1779 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1784 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1785 improves application performance by removing data copies and providing
1786 applications with zero-copy system calls such as sendfile and splice.
1790 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1791 option is set, openssl cleanses (zeroize) plaintext bytes from
1792 internal buffers after delivering them to the application. Note,
1793 the application is still responsible for cleansing other copies
1794 (e.g.: data received by SSL_read(3)).
1798 * `PKCS12_parse` now maintains the order of the parsed certificates
1799 when outputting them via `*ca` (rather than reversing it).
1803 * Deprecated pthread fork support methods. These were unused so no
1804 replacement is required.
1806 - OPENSSL_fork_prepare()
1807 - OPENSSL_fork_parent()
1808 - OPENSSL_fork_child()
1812 * Added support for FFDHE key exchange in TLS 1.3.
1819 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1821 * Fixed a problem with verifying a certificate chain when using the
1822 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1823 the certificates present in a certificate chain. It is not set by default.
1825 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1826 the chain that have explicitly encoded elliptic curve parameters was added
1827 as an additional strict check.
1829 An error in the implementation of this check meant that the result of a
1830 previous check to confirm that certificates in the chain are valid CA
1831 certificates was overwritten. This effectively bypasses the check
1832 that non-CA certificates must not be able to issue other certificates.
1834 If a "purpose" has been configured then there is a subsequent opportunity
1835 for checks that the certificate is a valid CA. All of the named "purpose"
1836 values implemented in libcrypto perform this check. Therefore, where
1837 a purpose is set the certificate chain will still be rejected even when the
1838 strict flag has been used. A purpose is set by default in libssl client and
1839 server certificate verification routines, but it can be overridden or
1840 removed by an application.
1842 In order to be affected, an application must explicitly set the
1843 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1844 for the certificate verification or, in the case of TLS client or server
1845 applications, override the default purpose.
1850 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1851 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1852 renegotiation ClientHello omits the signature_algorithms extension (where it
1853 was present in the initial ClientHello), but includes a
1854 signature_algorithms_cert extension then a NULL pointer dereference will
1855 result, leading to a crash and a denial of service attack.
1857 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1858 (which is the default configuration). OpenSSL TLS clients are not impacted by
1862 *Peter Kästle and Samuel Sapalski*
1864 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1866 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1867 create a unique hash value based on the issuer and serial number data
1868 contained within an X509 certificate. However it was failing to correctly
1869 handle any errors that may occur while parsing the issuer field (which might
1870 occur if the issuer field is maliciously constructed). This may subsequently
1871 result in a NULL pointer deref and a crash leading to a potential denial of
1877 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1878 padding mode to correctly check for rollback attacks. This is considered a
1879 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1884 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1885 functions. Previously they could overflow the output length argument in some
1886 cases where the input length is close to the maximum permissable length for
1887 an integer on the platform. In such cases the return value from the function
1888 call would be 1 (indicating success), but the output length value would be
1889 negative. This could cause applications to behave incorrectly or crash.
1894 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1895 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1896 could be exploited in a side channel attack to recover the password. Since
1897 the attack is local host only this is outside of the current OpenSSL
1898 threat model and therefore no CVE is assigned.
1900 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1905 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1907 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1908 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1909 If an attacker can control both items being compared then this could lead
1910 to a possible denial of service attack. OpenSSL itself uses the
1911 GENERAL_NAME_cmp function for two purposes:
1912 1) Comparing CRL distribution point names between an available CRL and a
1913 CRL distribution point embedded in an X509 certificate
1914 2) When verifying that a timestamp response token signer matches the
1915 timestamp authority name (exposed via the API functions
1916 TS_RESP_verify_response and TS_RESP_verify_token)
1921 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1923 * Certificates with explicit curve parameters are now disallowed in
1924 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1928 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1929 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1930 conversely, silently ignore DTLS protocol version bounds when configuring
1931 TLS-based contexts. The commands can be repeated to set bounds of both
1932 types. The same applies with the corresponding "min_protocol" and
1933 "max_protocol" command-line switches, in case some application uses both TLS
1936 SSL_CTX instances that are created for a fixed protocol version (e.g.
1937 TLSv1_server_method()) also silently ignore version bounds. Previously
1938 attempts to apply bounds to these protocol versions would result in an
1939 error. Now only the "version-flexible" SSL_CTX instances are subject to
1940 limits in configuration files in command-line options.
1944 * Handshake now fails if Extended Master Secret extension is dropped
1949 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1951 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1953 * Fixed segmentation fault in SSL_check_chain()
1954 Server or client applications that call the SSL_check_chain() function
1955 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1956 dereference as a result of incorrect handling of the
1957 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1958 or unrecognised signature algorithm is received from the peer. This could
1959 be exploited by a malicious peer in a Denial of Service attack.
1964 * Added AES consttime code for no-asm configurations
1965 an optional constant time support for AES was added
1966 when building openssl for no-asm.
1967 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1968 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1969 At this time this feature is by default disabled.
1970 It will be enabled by default in 3.0.
1974 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1976 * Revert the change of EOF detection while reading in libssl to avoid
1977 regressions in applications depending on the current way of reporting
1978 the EOF. As the existing method is not fully accurate the change to
1979 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1980 branch and will be present in the 3.0 release.
1984 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1985 when primes for RSA keys are computed.
1986 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1987 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1988 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1989 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1990 This avoids possible fingerprinting of newly generated RSA modules.
1994 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1996 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1997 while reading in libssl then we would report an error back to the
1998 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1999 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2000 therefore give a hint as to what went wrong.
2004 * Check that ed25519 and ed448 are allowed by the security level. Previously
2005 signature algorithms not using an MD were not being checked that they were
2006 allowed by the security level.
2010 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2011 was not quite right. The behaviour was not consistent between resumption
2012 and normal handshakes, and also not quite consistent with historical
2013 behaviour. The behaviour in various scenarios has been clarified and
2014 it has been updated to make it match historical behaviour as closely as
2019 * *[VMS only]* The header files that the VMS compilers include automatically,
2020 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2021 that the C++ compiler doesn't understand. This is a shortcoming in the
2022 compiler, but can be worked around with `__cplusplus` guards.
2024 C++ applications that use OpenSSL libraries must be compiled using the
2025 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2026 functions. Otherwise, only functions with symbols of less than 31
2027 characters can be used, as the linker will not be able to successfully
2028 resolve symbols with longer names.
2032 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2033 The presence of this system service is determined at run-time.
2037 * Added newline escaping functionality to a filename when using openssl dgst.
2038 This output format is to replicate the output format found in the `*sum`
2039 checksum programs. This aims to preserve backward compatibility.
2041 *Matt Eaton, Richard Levitte, and Paul Dale*
2043 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2048 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2050 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2051 number generator (RNG). This was intended to include protection in the
2052 event of a fork() system call in order to ensure that the parent and child
2053 processes did not share the same RNG state. However this protection was not
2054 being used in the default case.
2056 A partial mitigation for this issue is that the output from a high
2057 precision timer is mixed into the RNG state so the likelihood of a parent
2058 and child process sharing state is significantly reduced.
2060 If an application already calls OPENSSL_init_crypto() explicitly using
2061 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2064 *Matthias St. Pierre*
2066 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2067 used even when parsing explicit parameters, when loading a encoded key
2068 or calling `EC_GROUP_new_from_ecpkparameters()`/
2069 `EC_GROUP_new_from_ecparameters()`.
2070 This prevents bypass of security hardening and performance gains,
2071 especially for curves with specialized EC_METHODs.
2072 By default, if a key encoded with explicit parameters is loaded and later
2073 encoded, the output is still encoded with explicit parameters, even if
2074 internally a "named" EC_GROUP is used for computation.
2078 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2079 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2080 NULL. After this change, only the cofactor parameter can be NULL. It also
2081 does some minimal sanity checks on the passed order.
2086 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2087 An attack is simple, if the first CMS_recipientInfo is valid but the
2088 second CMS_recipientInfo is chosen ciphertext. If the second
2089 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2090 encryption key will be replaced by garbage, and the message cannot be
2091 decoded, but if the RSA decryption fails, the correct encryption key is
2092 used and the recipient will not notice the attack.
2093 As a work around for this potential attack the length of the decrypted
2094 key must be equal to the cipher default key length, in case the
2095 certifiate is not given and all recipientInfo are tried out.
2096 The old behaviour can be re-enabled in the CMS code by setting the
2097 CMS_DEBUG_DECRYPT flag.
2102 * Early start up entropy quality from the DEVRANDOM seed source has been
2103 improved for older Linux systems. The RAND subsystem will wait for
2104 /dev/random to be producing output before seeding from /dev/urandom.
2105 The seeded state is stored for future library initialisations using
2106 a system global shared memory segment. The shared memory identifier
2107 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2108 the desired value. The default identifier is 114.
2112 * Correct the extended master secret constant on EBCDIC systems. Without this
2113 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2114 negotiate EMS will fail. Unfortunately this also means that TLS connections
2115 between EBCDIC systems with this fix, and EBCDIC systems without this
2116 fix will fail if they negotiate EMS.
2120 * Use Windows installation paths in the mingw builds
2122 Mingw isn't a POSIX environment per se, which means that Windows
2123 paths should be used for installation.
2128 * Changed DH_check to accept parameters with order q and 2q subgroups.
2129 With order 2q subgroups the bit 0 of the private key is not secret
2130 but DH_generate_key works around that by clearing bit 0 of the
2131 private key for those. This avoids leaking bit 0 of the private key.
2135 * Significantly reduce secure memory usage by the randomness pools.
2139 * Revert the DEVRANDOM_WAIT feature for Linux systems
2141 The DEVRANDOM_WAIT feature added a select() call to wait for the
2142 /dev/random device to become readable before reading from the
2143 /dev/urandom device.
2145 It turned out that this change had negative side effects on
2146 performance which were not acceptable. After some discussion it
2147 was decided to revert this feature and leave it up to the OS
2148 resp. the platform maintainer to ensure a proper initialization
2149 during early boot time.
2151 *Matthias St. Pierre*
2153 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2155 * Add build tests for C++. These are generated files that only do one
2156 thing, to include one public OpenSSL head file each. This tests that
2157 the public header files can be usefully included in a C++ application.
2159 This test isn't enabled by default. It can be enabled with the option
2160 'enable-buildtest-c++'.
2164 * Enable SHA3 pre-hashing for ECDSA and DSA.
2168 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2169 This changes the size when using the `genpkey` command when no size is given.
2170 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2171 generation commands to use 2048 bits by default.
2175 * Reorganize the manual pages to consistently have RETURN VALUES,
2176 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2177 util/fix-doc-nits accordingly.
2179 *Paul Yang, Joshua Lock*
2181 * Add the missing accessor EVP_PKEY_get0_engine()
2185 * Have commands like `s_client` and `s_server` output the signature scheme
2186 along with other cipher suite parameters when debugging.
2190 * Make OPENSSL_config() error agnostic again.
2194 * Do the error handling in RSA decryption constant time.
2198 * Prevent over long nonces in ChaCha20-Poly1305.
2200 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2201 for every encryption operation. RFC 7539 specifies that the nonce value
2202 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2203 and front pads the nonce with 0 bytes if it is less than 12
2204 bytes. However it also incorrectly allows a nonce to be set of up to 16
2205 bytes. In this case only the last 12 bytes are significant and any
2206 additional leading bytes are ignored.
2208 It is a requirement of using this cipher that nonce values are
2209 unique. Messages encrypted using a reused nonce value are susceptible to
2210 serious confidentiality and integrity attacks. If an application changes
2211 the default nonce length to be longer than 12 bytes and then makes a
2212 change to the leading bytes of the nonce expecting the new value to be a
2213 new unique nonce then such an application could inadvertently encrypt
2214 messages with a reused nonce.
2216 Additionally the ignored bytes in a long nonce are not covered by the
2217 integrity guarantee of this cipher. Any application that relies on the
2218 integrity of these ignored leading bytes of a long nonce may be further
2219 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2220 is safe because no such use sets such a long nonce value. However user
2221 applications that use this cipher directly and set a non-default nonce
2222 length to be longer than 12 bytes may be vulnerable.
2224 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2230 * Add DEVRANDOM_WAIT feature for Linux systems
2232 On older Linux systems where the getrandom() system call is not available,
2233 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2234 Contrary to getrandom(), the /dev/urandom device will not block during
2235 early boot when the kernel CSPRNG has not been seeded yet.
2237 To mitigate this known weakness, use select() to wait for /dev/random to
2238 become readable before reading from /dev/urandom.
2240 * Ensure that SM2 only uses SM3 as digest algorithm
2244 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2246 * Change the info callback signals for the start and end of a post-handshake
2247 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2248 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2249 confused by this and assume that a TLSv1.2 renegotiation has started. This
2250 can break KeyUpdate handling. Instead we no longer signal the start and end
2251 of a post handshake message exchange (although the messages themselves are
2252 still signalled). This could break some applications that were expecting
2253 the old signals. However without this KeyUpdate is not usable for many
2258 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2260 * Timing vulnerability in DSA signature generation
2262 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2263 timing side channel attack. An attacker could use variations in the signing
2264 algorithm to recover the private key.
2266 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2271 * Timing vulnerability in ECDSA signature generation
2273 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2274 timing side channel attack. An attacker could use variations in the signing
2275 algorithm to recover the private key.
2277 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2282 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2283 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2284 of two gigabytes and the error handling improved.
2286 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2287 categorized as a normal bug, not a security issue, because the DRBG reseeds
2288 automatically and is fully functional even without additional randomness
2289 provided by the application.
2291 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2293 * Add a new ClientHello callback. Provides a callback interface that gives
2294 the application the ability to adjust the nascent SSL object at the
2295 earliest stage of ClientHello processing, immediately after extensions have
2296 been collected but before they have been processed. In particular, this
2297 callback can adjust the supported TLS versions in response to the contents
2302 * Add SM2 base algorithm support.
2306 * s390x assembly pack: add (improved) hardware-support for the following
2307 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2308 aes-cfb/cfb8, aes-ecb.
2312 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2313 parameter is no longer accepted, as it leads to a corrupt table. NULL
2314 pem_str is reserved for alias entries only.
2318 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2319 step for prime curves. The new implementation is based on formulae from
2320 differential addition-and-doubling in homogeneous projective coordinates
2321 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2322 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2323 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2324 to work in projective coordinates.
2326 *Billy Bob Brumley, Nicola Tuveri*
2328 * Change generating and checking of primes so that the error rate of not
2329 being prime depends on the intended use based on the size of the input.
2330 For larger primes this will result in more rounds of Miller-Rabin.
2331 The maximal error rate for primes with more than 1080 bits is lowered
2334 *Kurt Roeckx, Annie Yousar*
2336 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2340 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2341 moving between systems, and to avoid confusion when a Windows build is
2342 done with mingw vs with MSVC. For POSIX installs, there's still a
2343 symlink or copy named 'tsget' to avoid that confusion as well.
2347 * Revert blinding in ECDSA sign and instead make problematic addition
2348 length-invariant. Switch even to fixed-length Montgomery multiplication.
2352 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2353 step for binary curves. The new implementation is based on formulae from
2354 differential addition-and-doubling in mixed Lopez-Dahab projective
2355 coordinates, modified to independently blind the operands.
2357 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2359 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2360 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2361 EC_METHODs to implement their own specialized "ladder step", to take
2362 advantage of more favorable coordinate systems or more efficient
2363 differential addition-and-doubling algorithms.
2365 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2367 * Modified the random device based seed sources to keep the relevant
2368 file descriptors open rather than reopening them on each access.
2369 This allows such sources to operate in a chroot() jail without
2370 the associated device nodes being available. This behaviour can be
2371 controlled using RAND_keep_random_devices_open().
2375 * Numerous side-channel attack mitigations have been applied. This may have
2376 performance impacts for some algorithms for the benefit of improved
2377 security. Specific changes are noted in this change log by their respective
2382 * AIX shared library support overhaul. Switch to AIX "natural" way of
2383 handling shared libraries, which means collecting shared objects of
2384 different versions and bitnesses in one common archive. This allows to
2385 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2386 doesn't affect the way 3rd party applications are linked, only how
2387 multi-version installation is managed.
2391 * Make ec_group_do_inverse_ord() more robust and available to other
2392 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2393 mitigations are applied to the fallback BN_mod_inverse().
2394 When using this function rather than BN_mod_inverse() directly, new
2395 EC cryptosystem implementations are then safer-by-default.
2399 * Add coordinate blinding for EC_POINT and implement projective
2400 coordinate blinding for generic prime curves as a countermeasure to
2401 chosen point SCA attacks.
2403 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2405 * Add blinding to ECDSA and DSA signatures to protect against side channel
2406 attacks discovered by Keegan Ryan (NCC Group).
2410 * Enforce checking in the `pkeyutl` command to ensure that the input
2411 length does not exceed the maximum supported digest length when performing
2412 a sign, verify or verifyrecover operation.
2416 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2417 I/O in combination with something like select() or poll() will hang. This
2418 can be turned off again using SSL_CTX_clear_mode().
2419 Many applications do not properly handle non-application data records, and
2420 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2421 around the problems in those applications, but can also break some.
2422 It's recommended to read the manpages about SSL_read(), SSL_write(),
2423 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2424 SSL_CTX_set_read_ahead() again.
2428 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2429 now allow empty (zero character) pass phrases.
2433 * Apply blinding to binary field modular inversion and remove patent
2434 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2438 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2439 binary and prime elliptic curves.
2443 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2444 constant time fixed point multiplication.
2448 * Revise elliptic curve scalar multiplication with timing attack
2449 defenses: ec_wNAF_mul redirects to a constant time implementation
2450 when computing fixed point and variable point multiplication (which
2451 in OpenSSL are mostly used with secret scalars in keygen, sign,
2452 ECDH derive operations).
2453 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2456 * Updated CONTRIBUTING
2460 * Updated DRBG / RAND to request nonce and additional low entropy
2461 randomness from the system.
2463 *Matthias St. Pierre*
2465 * Updated 'openssl rehash' to use OpenSSL consistent default.
2469 * Moved the load of the ssl_conf module to libcrypto, which helps
2470 loading engines that libssl uses before libssl is initialised.
2474 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2478 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2480 *Ingo Schwarze, Rich Salz*
2482 * Added output of accepting IP address and port for 'openssl s_server'
2486 * Added a new API for TLSv1.3 ciphersuites:
2487 SSL_CTX_set_ciphersuites()
2488 SSL_set_ciphersuites()
2492 * Memory allocation failures consistently add an error to the error
2497 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2498 in libcrypto when run as setuid/setgid.
2502 * Load any config file by default when libssl is used.
2506 * Added new public header file <openssl/rand_drbg.h> and documentation
2507 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2509 *Matthias St. Pierre*
2511 * QNX support removed (cannot find contributors to get their approval
2512 for the license change).
2516 * TLSv1.3 replay protection for early data has been implemented. See the
2517 SSL_read_early_data() man page for further details.
2521 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2522 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2523 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2524 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2525 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2526 configuration has been separated out. See the ciphers man page or the
2527 SSL_CTX_set_ciphersuites() man page for more information.
2531 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2532 in responder mode now supports the new "-multi" option, which
2533 spawns the specified number of child processes to handle OCSP
2534 requests. The "-timeout" option now also limits the OCSP
2535 responder's patience to wait to receive the full client request
2536 on a newly accepted connection. Child processes are respawned
2537 as needed, and the CA index file is automatically reloaded
2538 when changed. This makes it possible to run the "ocsp" responder
2539 as a long-running service, making the OpenSSL CA somewhat more
2540 feature-complete. In this mode, most diagnostic messages logged
2541 after entering the event loop are logged via syslog(3) rather than
2546 * Added support for X448 and Ed448. Heavily based on original work by
2551 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2552 objects loaded. This adds the functions OSSL_STORE_expect() and
2553 OSSL_STORE_find() as well as needed tools to construct searches and
2554 get the search data out of them.
2558 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2559 version of OpenSSL should review their configuration settings to ensure
2560 that they are still appropriate for TLSv1.3. For further information see:
2561 <https://wiki.openssl.org/index.php/TLS1.3>
2565 * Grand redesign of the OpenSSL random generator
2567 The default RAND method now utilizes an AES-CTR DRBG according to
2568 NIST standard SP 800-90Ar1. The new random generator is essentially
2569 a port of the default random generator from the OpenSSL FIPS 2.0
2570 object module. It is a hybrid deterministic random bit generator
2571 using an AES-CTR bit stream and which seeds and reseeds itself
2572 automatically using trusted system entropy sources.
2574 Some of its new features are:
2575 - Support for multiple DRBG instances with seed chaining.
2576 - The default RAND method makes use of a DRBG.
2577 - There is a public and private DRBG instance.
2578 - The DRBG instances are fork-safe.
2579 - Keep all global DRBG instances on the secure heap if it is enabled.
2580 - The public and private DRBG instance are per thread for lock free
2583 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2585 * Changed Configure so it only says what it does and doesn't dump
2586 so much data. Instead, ./configdata.pm should be used as a script
2587 to display all sorts of configuration data.
2591 * Added processing of "make variables" to Configure.
2595 * Added SHA512/224 and SHA512/256 algorithm support.
2599 * The last traces of Netware support, first removed in 1.1.0, have
2604 * Get rid of Makefile.shared, and in the process, make the processing
2605 of certain files (rc.obj, or the .def/.map/.opt files produced from
2606 the ordinal files) more visible and hopefully easier to trace and
2607 debug (or make silent).
2611 * Make it possible to have environment variable assignments as
2612 arguments to config / Configure.
2616 * Add multi-prime RSA (RFC 8017) support.
2620 * Add SM3 implemented according to GB/T 32905-2016
2621 *Jack Lloyd <jack.lloyd@ribose.com>,*
2622 *Ronald Tse <ronald.tse@ribose.com>,*
2623 *Erick Borsboom <erick.borsboom@ribose.com>*
2625 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2626 as documented in RFC6066.
2627 Based on a patch from Tomasz Moń
2629 *Filipe Raimundo da Silva*
2631 * Add SM4 implemented according to GB/T 32907-2016.
2632 *Jack Lloyd <jack.lloyd@ribose.com>,*
2633 *Ronald Tse <ronald.tse@ribose.com>,*
2634 *Erick Borsboom <erick.borsboom@ribose.com>*
2636 * Reimplement -newreq-nodes and ERR_error_string_n; the
2637 original author does not agree with the license change.
2641 * Add ARIA AEAD TLS support.
2645 * Some macro definitions to support VS6 have been removed. Visual
2646 Studio 6 has not worked since 1.1.0
2650 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2651 without clearing the errors.
2655 * Add "atfork" functions. If building on a system that without
2656 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2657 requirements. The RAND facility now uses/requires this.
2665 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2666 not possible to disable entirely. However, it's still possible to
2667 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2670 To disable, configure with 'no-ui-console'. 'no-ui' is still
2671 possible to use as an alias. Check at compile time with the
2672 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2673 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2677 * Add a STORE module, which implements a uniform and URI based reader of
2678 stores that can contain keys, certificates, CRLs and numerous other
2679 objects. The main API is loosely based on a few stdio functions,
2680 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2681 OSSL_STORE_error and OSSL_STORE_close.
2682 The implementation uses backends called "loaders" to implement arbitrary
2683 URI schemes. There is one built in "loader" for the 'file' scheme.
2687 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2688 then adjusted to work on FreeBSD 8.4 as well.
2689 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2690 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2694 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2695 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2696 error code calls like this:
2698 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2700 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2701 that can be encoded in C. For the foreseeable future, this will only
2704 *Richard Levitte and Tim Hudson*
2706 * Removed BSD cryptodev engine.
2710 * Add a build target 'build_all_generated', to build all generated files
2711 and only that. This can be used to prepare everything that requires
2712 things like perl for a system that lacks perl and then move everything
2713 to that system and do the rest of the build there.
2717 * In the UI interface, make it possible to duplicate the user data. This
2718 can be used by engines that need to retain the data for a longer time
2719 than just the call where this user data is passed.
2723 * Ignore the '-named_curve auto' value for compatibility of applications
2726 *Tomáš Mráz <tmraz@fedoraproject.org>*
2728 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2729 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2730 alerts across multiple records (some of which could be empty). In practice
2731 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2732 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2733 support this at all. Supporting it adds significant complexity to the
2734 record layer, and its removal is unlikely to cause interoperability
2739 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2740 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2741 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2746 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2747 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2749 *Richard Levitte, Andy Polyakov*
2751 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2756 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2757 platform rather than 'mingw'.
2761 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2762 success if they are asked to add an object which already exists
2763 in the store. This change cascades to other functions which load
2764 certificates and CRLs.
2768 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2769 facilitate stack unwinding even from assembly subroutines.
2773 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2774 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2778 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2779 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2780 which is the minimum version we support.
2784 * Certificate time validation (X509_cmp_time) enforces stricter
2785 compliance with RFC 5280. Fractional seconds and timezone offsets
2786 are no longer allowed.
2790 * Add support for ARIA
2794 * s_client will now send the Server Name Indication (SNI) extension by
2795 default unless the new "-noservername" option is used. The server name is
2796 based on the host provided to the "-connect" option unless overridden by
2797 using "-servername".
2801 * Add support for SipHash
2805 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2806 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2807 prevent issues where no progress is being made and the peer continually
2808 sends unrecognised record types, using up resources processing them.
2812 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2813 using the algorithm defined in
2814 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2818 * Heartbeat support has been removed; the ABI is changed for now.
2820 *Richard Levitte, Rich Salz*
2822 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2826 * The RSA "null" method, which was partially supported to avoid patent
2827 issues, has been replaced to always returns NULL.
2834 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2836 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2837 used even when parsing explicit parameters, when loading a encoded key
2838 or calling `EC_GROUP_new_from_ecpkparameters()`/
2839 `EC_GROUP_new_from_ecparameters()`.
2840 This prevents bypass of security hardening and performance gains,
2841 especially for curves with specialized EC_METHODs.
2842 By default, if a key encoded with explicit parameters is loaded and later
2843 encoded, the output is still encoded with explicit parameters, even if
2844 internally a "named" EC_GROUP is used for computation.
2848 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2849 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2850 NULL. After this change, only the cofactor parameter can be NULL. It also
2851 does some minimal sanity checks on the passed order.
2856 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2857 An attack is simple, if the first CMS_recipientInfo is valid but the
2858 second CMS_recipientInfo is chosen ciphertext. If the second
2859 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2860 encryption key will be replaced by garbage, and the message cannot be
2861 decoded, but if the RSA decryption fails, the correct encryption key is
2862 used and the recipient will not notice the attack.
2863 As a work around for this potential attack the length of the decrypted
2864 key must be equal to the cipher default key length, in case the
2865 certifiate is not given and all recipientInfo are tried out.
2866 The old behaviour can be re-enabled in the CMS code by setting the
2867 CMS_DEBUG_DECRYPT flag.
2872 * Use Windows installation paths in the mingw builds
2874 Mingw isn't a POSIX environment per se, which means that Windows
2875 paths should be used for installation.
2880 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2882 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2883 This changes the size when using the `genpkey` command when no size is given.
2884 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2885 generation commands to use 2048 bits by default.
2889 * Prevent over long nonces in ChaCha20-Poly1305.
2891 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2892 for every encryption operation. RFC 7539 specifies that the nonce value
2893 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2894 and front pads the nonce with 0 bytes if it is less than 12
2895 bytes. However it also incorrectly allows a nonce to be set of up to 16
2896 bytes. In this case only the last 12 bytes are significant and any
2897 additional leading bytes are ignored.
2899 It is a requirement of using this cipher that nonce values are
2900 unique. Messages encrypted using a reused nonce value are susceptible to
2901 serious confidentiality and integrity attacks. If an application changes
2902 the default nonce length to be longer than 12 bytes and then makes a
2903 change to the leading bytes of the nonce expecting the new value to be a
2904 new unique nonce then such an application could inadvertently encrypt
2905 messages with a reused nonce.
2907 Additionally the ignored bytes in a long nonce are not covered by the
2908 integrity guarantee of this cipher. Any application that relies on the
2909 integrity of these ignored leading bytes of a long nonce may be further
2910 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2911 is safe because no such use sets such a long nonce value. However user
2912 applications that use this cipher directly and set a non-default nonce
2913 length to be longer than 12 bytes may be vulnerable.
2915 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2921 * Added SCA hardening for modular field inversion in EC_GROUP through
2922 a new dedicated field_inv() pointer in EC_METHOD.
2923 This also addresses a leakage affecting conversions from projective
2924 to affine coordinates.
2926 *Billy Bob Brumley, Nicola Tuveri*
2928 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2929 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2933 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2937 * Remove the 'dist' target and add a tarball building script. The
2938 'dist' target has fallen out of use, and it shouldn't be
2939 necessary to configure just to create a source distribution.
2943 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2945 * Timing vulnerability in DSA signature generation
2947 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2948 timing side channel attack. An attacker could use variations in the signing
2949 algorithm to recover the private key.
2951 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2956 * Timing vulnerability in ECDSA signature generation
2958 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2959 timing side channel attack. An attacker could use variations in the signing
2960 algorithm to recover the private key.
2962 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2967 * Add coordinate blinding for EC_POINT and implement projective
2968 coordinate blinding for generic prime curves as a countermeasure to
2969 chosen point SCA attacks.
2971 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2973 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2975 * Client DoS due to large DH parameter
2977 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2978 malicious server can send a very large prime value to the client. This will
2979 cause the client to spend an unreasonably long period of time generating a
2980 key for this prime resulting in a hang until the client has finished. This
2981 could be exploited in a Denial Of Service attack.
2983 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2988 * Cache timing vulnerability in RSA Key Generation
2990 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2991 a cache timing side channel attack. An attacker with sufficient access to
2992 mount cache timing attacks during the RSA key generation process could
2993 recover the private key.
2995 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2996 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3001 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3002 parameter is no longer accepted, as it leads to a corrupt table. NULL
3003 pem_str is reserved for alias entries only.
3007 * Revert blinding in ECDSA sign and instead make problematic addition
3008 length-invariant. Switch even to fixed-length Montgomery multiplication.
3012 * Change generating and checking of primes so that the error rate of not
3013 being prime depends on the intended use based on the size of the input.
3014 For larger primes this will result in more rounds of Miller-Rabin.
3015 The maximal error rate for primes with more than 1080 bits is lowered
3018 *Kurt Roeckx, Annie Yousar*
3020 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3024 * Add blinding to ECDSA and DSA signatures to protect against side channel
3025 attacks discovered by Keegan Ryan (NCC Group).
3029 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3030 now allow empty (zero character) pass phrases.
3034 * Certificate time validation (X509_cmp_time) enforces stricter
3035 compliance with RFC 5280. Fractional seconds and timezone offsets
3036 are no longer allowed.
3040 * Fixed a text canonicalisation bug in CMS
3042 Where a CMS detached signature is used with text content the text goes
3043 through a canonicalisation process first prior to signing or verifying a
3044 signature. This process strips trailing space at the end of lines, converts
3045 line terminators to CRLF and removes additional trailing line terminators
3046 at the end of a file. A bug in the canonicalisation process meant that
3047 some characters, such as form-feed, were incorrectly treated as whitespace
3048 and removed. This is contrary to the specification (RFC5485). This fix
3049 could mean that detached text data signed with an earlier version of
3050 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3051 signed with a fixed OpenSSL may fail to verify with an earlier version of
3052 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3053 and use the "-binary" flag (for the "cms" command line application) or set
3054 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3058 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3060 * Constructed ASN.1 types with a recursive definition could exceed the stack
3062 Constructed ASN.1 types with a recursive definition (such as can be found
3063 in PKCS7) could eventually exceed the stack given malicious input with
3064 excessive recursion. This could result in a Denial Of Service attack. There
3065 are no such structures used within SSL/TLS that come from untrusted sources
3066 so this is considered safe.
3068 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3074 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3076 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3077 effectively reduced to only comparing the least significant bit of each
3078 byte. This allows an attacker to forge messages that would be considered as
3079 authenticated in an amount of tries lower than that guaranteed by the
3080 security claims of the scheme. The module can only be compiled by the
3081 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3083 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3089 * Add a build target 'build_all_generated', to build all generated files
3090 and only that. This can be used to prepare everything that requires
3091 things like perl for a system that lacks perl and then move everything
3092 to that system and do the rest of the build there.
3096 * Backport SSL_OP_NO_RENGOTIATION
3098 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3099 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3100 changes this is no longer possible in 1.1.0. Therefore the new
3101 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3102 1.1.0 to provide equivalent functionality.
3104 Note that if an application built against 1.1.0h headers (or above) is run
3105 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3106 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3110 * Removed the OS390-Unix config target. It relied on a script that doesn't
3115 * rsaz_1024_mul_avx2 overflow bug on x86_64
3117 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3118 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3119 Analysis suggests that attacks against RSA and DSA as a result of this
3120 defect would be very difficult to perform and are not believed likely.
3121 Attacks against DH1024 are considered just feasible, because most of the
3122 work necessary to deduce information about a private key may be performed
3123 offline. The amount of resources required for such an attack would be
3124 significant. However, for an attack on TLS to be meaningful, the server
3125 would have to share the DH1024 private key among multiple clients, which is
3126 no longer an option since CVE-2016-0701.
3128 This only affects processors that support the AVX2 but not ADX extensions
3129 like Intel Haswell (4th generation).
3131 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3132 was originally found via the OSS-Fuzz project.
3137 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3139 * bn_sqrx8x_internal carry bug on x86_64
3141 There is a carry propagating bug in the x86_64 Montgomery squaring
3142 procedure. No EC algorithms are affected. Analysis suggests that attacks
3143 against RSA and DSA as a result of this defect would be very difficult to
3144 perform and are not believed likely. Attacks against DH are considered just
3145 feasible (although very difficult) because most of the work necessary to
3146 deduce information about a private key may be performed offline. The amount
3147 of resources required for such an attack would be very significant and
3148 likely only accessible to a limited number of attackers. An attacker would
3149 additionally need online access to an unpatched system using the target
3150 private key in a scenario with persistent DH parameters and a private
3151 key that is shared between multiple clients.
3153 This only affects processors that support the BMI1, BMI2 and ADX extensions
3154 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3156 This issue was reported to OpenSSL by the OSS-Fuzz project.
3161 * Malformed X.509 IPAddressFamily could cause OOB read
3163 If an X.509 certificate has a malformed IPAddressFamily extension,
3164 OpenSSL could do a one-byte buffer overread. The most likely result
3165 would be an erroneous display of the certificate in text format.
3167 This issue was reported to OpenSSL by the OSS-Fuzz project.
3172 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3174 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3175 platform rather than 'mingw'.
3179 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3180 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3181 which is the minimum version we support.
3185 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3187 * Encrypt-Then-Mac renegotiation crash
3189 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3190 negotiated where it was not in the original handshake (or vice-versa) then
3191 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3192 and servers are affected.
3194 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3199 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3201 * Truncated packet could crash via OOB read
3203 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3204 cipher is being used, then a truncated packet can cause that host to
3205 perform an out-of-bounds read, usually resulting in a crash.
3207 This issue was reported to OpenSSL by Robert Święcki of Google.
3212 * Bad (EC)DHE parameters cause a client crash
3214 If a malicious server supplies bad parameters for a DHE or ECDHE key
3215 exchange then this can result in the client attempting to dereference a
3216 NULL pointer leading to a client crash. This could be exploited in a Denial
3219 This issue was reported to OpenSSL by Guido Vranken.
3224 * BN_mod_exp may produce incorrect results on x86_64
3226 There is a carry propagating bug in the x86_64 Montgomery squaring
3227 procedure. No EC algorithms are affected. Analysis suggests that attacks
3228 against RSA and DSA as a result of this defect would be very difficult to
3229 perform and are not believed likely. Attacks against DH are considered just
3230 feasible (although very difficult) because most of the work necessary to
3231 deduce information about a private key may be performed offline. The amount
3232 of resources required for such an attack would be very significant and
3233 likely only accessible to a limited number of attackers. An attacker would
3234 additionally need online access to an unpatched system using the target
3235 private key in a scenario with persistent DH parameters and a private
3236 key that is shared between multiple clients. For example this can occur by
3237 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3238 similar to CVE-2015-3193 but must be treated as a separate problem.
3240 This issue was reported to OpenSSL by the OSS-Fuzz project.
3245 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3247 * ChaCha20/Poly1305 heap-buffer-overflow
3249 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3250 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3251 crash. This issue is not considered to be exploitable beyond a DoS.
3253 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3258 * CMS Null dereference
3260 Applications parsing invalid CMS structures can crash with a NULL pointer
3261 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3262 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3263 structure callback if an attempt is made to free certain invalid encodings.
3264 Only CHOICE structures using a callback which do not handle NULL value are
3267 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3272 * Montgomery multiplication may produce incorrect results
3274 There is a carry propagating bug in the Broadwell-specific Montgomery
3275 multiplication procedure that handles input lengths divisible by, but
3276 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3277 and DH private keys are impossible. This is because the subroutine in
3278 question is not used in operations with the private key itself and an input
3279 of the attacker's direct choice. Otherwise the bug can manifest itself as
3280 transient authentication and key negotiation failures or reproducible
3281 erroneous outcome of public-key operations with specially crafted input.
3282 Among EC algorithms only Brainpool P-512 curves are affected and one
3283 presumably can attack ECDH key negotiation. Impact was not analyzed in
3284 detail, because pre-requisites for attack are considered unlikely. Namely
3285 multiple clients have to choose the curve in question and the server has to
3286 share the private key among them, neither of which is default behaviour.
3287 Even then only clients that chose the curve will be affected.
3289 This issue was publicly reported as transient failures and was not
3290 initially recognized as a security issue. Thanks to Richard Morgan for
3291 providing reproducible case.
3296 * Removed automatic addition of RPATH in shared libraries and executables,
3297 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3301 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3303 * Fix Use After Free for large message sizes
3305 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3306 message larger than approx 16k is received then the underlying buffer to
3307 store the incoming message is reallocated and moved. Unfortunately a
3308 dangling pointer to the old location is left which results in an attempt to
3309 write to the previously freed location. This is likely to result in a
3310 crash, however it could potentially lead to execution of arbitrary code.
3312 This issue only affects OpenSSL 1.1.0a.
3314 This issue was reported to OpenSSL by Robert Święcki.
3319 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3321 * OCSP Status Request extension unbounded memory growth
3323 A malicious client can send an excessively large OCSP Status Request
3324 extension. If that client continually requests renegotiation, sending a
3325 large OCSP Status Request extension each time, then there will be unbounded
3326 memory growth on the server. This will eventually lead to a Denial Of
3327 Service attack through memory exhaustion. Servers with a default
3328 configuration are vulnerable even if they do not support OCSP. Builds using
3329 the "no-ocsp" build time option are not affected.
3331 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3336 * SSL_peek() hang on empty record
3338 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3339 sends an empty record. This could be exploited by a malicious peer in a
3340 Denial Of Service attack.
3342 This issue was reported to OpenSSL by Alex Gaynor.
3347 * Excessive allocation of memory in tls_get_message_header() and
3348 dtls1_preprocess_fragment()
3350 A (D)TLS message includes 3 bytes for its length in the header for the
3351 message. This would allow for messages up to 16Mb in length. Messages of
3352 this length are excessive and OpenSSL includes a check to ensure that a
3353 peer is sending reasonably sized messages in order to avoid too much memory
3354 being consumed to service a connection. A flaw in the logic of version
3355 1.1.0 means that memory for the message is allocated too early, prior to
3356 the excessive message length check. Due to way memory is allocated in
3357 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3358 to service a connection. This could lead to a Denial of Service through
3359 memory exhaustion. However, the excessive message length check still takes
3360 place, and this would cause the connection to immediately fail. Assuming
3361 that the application calls SSL_free() on the failed connection in a timely
3362 manner then the 21Mb of allocated memory will then be immediately freed
3363 again. Therefore the excessive memory allocation will be transitory in
3364 nature. This then means that there is only a security impact if:
3366 1) The application does not call SSL_free() in a timely manner in the event
3367 that the connection fails
3369 2) The application is working in a constrained environment where there is
3370 very little free memory
3372 3) The attacker initiates multiple connection attempts such that there are
3373 multiple connections in a state where memory has been allocated for the
3374 connection; SSL_free() has not yet been called; and there is insufficient
3375 memory to service the multiple requests.
3377 Except in the instance of (1) above any Denial Of Service is likely to be
3378 transitory because as soon as the connection fails the memory is
3379 subsequently freed again in the SSL_free() call. However there is an
3380 increased risk during this period of application crashes due to the lack of
3381 memory - which would then mean a more serious Denial of Service.
3383 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3384 (CVE-2016-6307 and CVE-2016-6308)
3388 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3389 had to be removed. Primary reason is that vendor assembler can't
3390 assemble our modules with -KPIC flag. As result it, assembly
3391 support, was not even available as option. But its lack means
3392 lack of side-channel resistant code, which is incompatible with
3393 security by todays standards. Fortunately gcc is readily available
3394 prepackaged option, which we firmly point at...
3398 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3400 * Windows command-line tool supports UTF-8 opt-in option for arguments
3401 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3402 (to any value) allows Windows user to access PKCS#12 file generated
3403 with Windows CryptoAPI and protected with non-ASCII password, as well
3404 as files generated under UTF-8 locale on Linux also protected with
3409 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3410 have been disabled by default and removed from DEFAULT, just like RC4.
3411 See the RC4 item below to re-enable both.
3415 * The method for finding the storage location for the Windows RAND seed file
3416 has changed. First we check %RANDFILE%. If that is not set then we check
3417 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3418 all else fails we fall back to C:\.
3422 * The EVP_EncryptUpdate() function has had its return type changed from void
3423 to int. A return of 0 indicates and error while a return of 1 indicates
3428 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3429 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3430 off the constant time implementation for RSA, DSA and DH have been made
3431 no-ops and deprecated.
3435 * Windows RAND implementation was simplified to only get entropy by
3436 calling CryptGenRandom(). Various other RAND-related tickets
3439 *Joseph Wylie Yandle, Rich Salz*
3441 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3442 and `OPENSSL_LH_`, respectively. The old names are available
3443 with API compatibility. They new names are now completely documented.
3447 * Unify TYPE_up_ref(obj) methods signature.
3448 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3449 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3450 int (instead of void) like all others TYPE_up_ref() methods.
3451 So now these methods also check the return value of CRYPTO_atomic_add(),
3452 and the validity of object reference counter.
3454 *fdasilvayy@gmail.com*
3456 * With Windows Visual Studio builds, the .pdb files are installed
3457 alongside the installed libraries and executables. For a static
3458 library installation, ossl_static.pdb is the associate compiler
3459 generated .pdb file to be used when linking programs.
3463 * Remove openssl.spec. Packaging files belong with the packagers.
3467 * Automatic Darwin/OSX configuration has had a refresh, it will now
3468 recognise x86_64 architectures automatically. You can still decide
3469 to build for a different bitness with the environment variable
3470 KERNEL_BITS (can be 32 or 64), for example:
3472 KERNEL_BITS=32 ./config
3476 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3477 256 bit AES and HMAC with SHA256.
3481 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3485 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3489 * To enable users to have their own config files and build file templates,
3490 Configure looks in the directory indicated by the environment variable
3491 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3492 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3493 name and is used as is.
3497 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3498 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3499 X509_CERT_FILE_CTX was removed.
3503 * "shared" builds are now the default. To create only static libraries use
3504 the "no-shared" Configure option.
3508 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3509 All of these option have not worked for some while and are fundamental
3514 * Make various cleanup routines no-ops and mark them as deprecated. Most
3515 global cleanup functions are no longer required because they are handled
3516 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3517 Explicitly de-initing can cause problems (e.g. where a library that uses
3518 OpenSSL de-inits, but an application is still using it). The affected
3519 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3520 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3521 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3522 COMP_zlib_cleanup().
3526 * --strict-warnings no longer enables runtime debugging options
3527 such as REF_DEBUG. Instead, debug options are automatically
3528 enabled with '--debug' builds.
3530 *Andy Polyakov, Emilia Käsper*
3532 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3533 have been moved out of the public header files. New functions for managing
3534 these have been added.
3538 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3539 objects have been moved out of the public header files. New
3540 functions for managing these have been added.
3544 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3545 have been moved out of the public header files. New functions for managing
3546 these have been added.
3550 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3551 moved out of the public header files. New functions for managing these
3556 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3560 * Removed the mk1mf build scripts.
3564 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3565 it is always safe to #include a header now.
3569 * Removed the aged BC-32 config and all its supporting scripts
3573 * Removed support for Ultrix, Netware, and OS/2.
3577 * Add support for HKDF.
3579 *Alessandro Ghedini*
3581 * Add support for blake2b and blake2s
3585 * Added support for "pipelining". Ciphers that have the
3586 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3587 encryptions/decryptions simultaneously. There are currently no built-in
3588 ciphers with this property but the expectation is that engines will be able
3589 to offer it to significantly improve throughput. Support has been extended
3590 into libssl so that multiple records for a single connection can be
3591 processed in one go (for >=TLS 1.1).
3595 * Added the AFALG engine. This is an async capable engine which is able to
3596 offload work to the Linux kernel. In this initial version it only supports
3597 AES128-CBC. The kernel must be version 4.1.0 or greater.
3601 * OpenSSL now uses a new threading API. It is no longer necessary to
3602 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3603 are two supported threading models: pthreads and windows threads. It is
3604 also possible to configure OpenSSL at compile time for "no-threads". The
3605 old threading API should no longer be used. The functions have been
3606 replaced with "no-op" compatibility macros.
3608 *Alessandro Ghedini, Matt Caswell*
3610 * Modify behavior of ALPN to invoke callback after SNI/servername
3611 callback, such that updates to the SSL_CTX affect ALPN.
3615 * Add SSL_CIPHER queries for authentication and key-exchange.
3619 * Changes to the DEFAULT cipherlist:
3620 - Prefer (EC)DHE handshakes over plain RSA.
3621 - Prefer AEAD ciphers over legacy ciphers.
3622 - Prefer ECDSA over RSA when both certificates are available.
3623 - Prefer TLSv1.2 ciphers/PRF.
3624 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3629 * Change the ECC default curve list to be this, in order: x25519,
3630 secp256r1, secp521r1, secp384r1.
3634 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3635 disabled by default. They can be re-enabled using the
3636 enable-weak-ssl-ciphers option to Configure.
3640 * If the server has ALPN configured, but supports no protocols that the
3641 client advertises, send a fatal "no_application_protocol" alert.
3642 This behaviour is SHALL in RFC 7301, though it isn't universally
3643 implemented by other servers.
3647 * Add X25519 support.
3648 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3649 for public and private key encoding using the format documented in
3650 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3651 key generation and key derivation.
3653 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3658 * Deprecate SRP_VBASE_get_by_user.
3659 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3660 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3661 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3662 seed, even if the seed is configured.
3664 Users should use SRP_VBASE_get1_by_user instead. Note that in
3665 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3666 also that even though configuring the SRP seed attempts to hide
3667 invalid usernames by continuing the handshake with fake
3668 credentials, this behaviour is not constant time and no strong
3669 guarantees are made that the handshake is indistinguishable from
3670 that of a valid user.
3674 * Configuration change; it's now possible to build dynamic engines
3675 without having to build shared libraries and vice versa. This
3676 only applies to the engines in `engines/`, those in `crypto/engine/`
3677 will always be built into libcrypto (i.e. "static").
3679 Building dynamic engines is enabled by default; to disable, use
3680 the configuration option "disable-dynamic-engine".
3682 The only requirements for building dynamic engines are the
3683 presence of the DSO module and building with position independent
3684 code, so they will also automatically be disabled if configuring
3685 with "disable-dso" or "disable-pic".
3687 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3688 are also taken away from openssl/opensslconf.h, as they are
3693 * Configuration change; if there is a known flag to compile
3694 position independent code, it will always be applied on the
3695 libcrypto and libssl object files, and never on the application
3696 object files. This means other libraries that use routines from
3697 libcrypto / libssl can be made into shared libraries regardless
3698 of how OpenSSL was configured.
3700 If this isn't desirable, the configuration options "disable-pic"
3701 or "no-pic" can be used to disable the use of PIC. This will
3702 also disable building shared libraries and dynamic engines.
3706 * Removed JPAKE code. It was experimental and has no wide use.
3710 * The INSTALL_PREFIX Makefile variable has been renamed to
3711 DESTDIR. That makes for less confusion on what this variable
3712 is for. Also, the configuration option --install_prefix is
3717 * Heartbeat for TLS has been removed and is disabled by default
3718 for DTLS; configure with enable-heartbeats. Code that uses the
3719 old #define's might need to be updated.
3721 *Emilia Käsper, Rich Salz*
3723 * Rename REF_CHECK to REF_DEBUG.
3727 * New "unified" build system
3729 The "unified" build system is aimed to be a common system for all
3730 platforms we support. With it comes new support for VMS.
3732 This system builds supports building in a different directory tree
3733 than the source tree. It produces one Makefile (for unix family
3734 or lookalikes), or one descrip.mms (for VMS).
3736 The source of information to make the Makefile / descrip.mms is
3737 small files called 'build.info', holding the necessary
3738 information for each directory with source to compile, and a
3739 template in Configurations, like unix-Makefile.tmpl or
3742 With this change, the library names were also renamed on Windows
3743 and on VMS. They now have names that are closer to the standard
3744 on Unix, and include the major version number, and in certain
3745 cases, the architecture they are built for. See "Notes on shared
3746 libraries" in INSTALL.
3748 We rely heavily on the perl module Text::Template.
3752 * Added support for auto-initialisation and de-initialisation of the library.
3753 OpenSSL no longer requires explicit init or deinit routines to be called,
3754 except in certain circumstances. See the OPENSSL_init_crypto() and
3755 OPENSSL_init_ssl() man pages for further information.
3759 * The arguments to the DTLSv1_listen function have changed. Specifically the
3760 "peer" argument is now expected to be a BIO_ADDR object.
3762 * Rewrite of BIO networking library. The BIO library lacked consistent
3763 support of IPv6, and adding it required some more extensive
3764 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3765 which hold all types of addresses and chains of address information.
3766 It also introduces a new API, with functions like BIO_socket,
3767 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3768 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3769 have been adapted accordingly.
3773 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3778 * CRIME protection: disable compression by default, even if OpenSSL is
3779 compiled with zlib enabled. Applications can still enable compression
3780 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3781 using the SSL_CONF library to configure compression.
3785 * The signature of the session callback configured with
3786 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3787 was explicitly marked as `const unsigned char*` instead of
3792 * Always DPURIFY. Remove the use of uninitialized memory in the
3793 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3797 * Removed many obsolete configuration items, including
3798 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3799 MD2_CHAR, MD2_INT, MD2_LONG
3801 IDEA_SHORT, IDEA_LONG
3802 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3804 *Rich Salz, with advice from Andy Polyakov*
3806 * Many BN internals have been moved to an internal header file.
3808 *Rich Salz with help from Andy Polyakov*
3810 * Configuration and writing out the results from it has changed.
3811 Files such as Makefile include/openssl/opensslconf.h and are now
3812 produced through general templates, such as Makefile.in and
3813 crypto/opensslconf.h.in and some help from the perl module
3816 Also, the center of configuration information is no longer
3817 Makefile. Instead, Configure produces a perl module in
3818 configdata.pm which holds most of the config data (in the hash
3819 table %config), the target data that comes from the target
3820 configuration in one of the `Configurations/*.conf` files (in
3825 * To clarify their intended purposes, the Configure options
3826 --prefix and --openssldir change their semantics, and become more
3827 straightforward and less interdependent.
3829 --prefix shall be used exclusively to give the location INSTALLTOP
3830 where programs, scripts, libraries, include files and manuals are
3831 going to be installed. The default is now /usr/local.
3833 --openssldir shall be used exclusively to give the default
3834 location OPENSSLDIR where certificates, private keys, CRLs are
3835 managed. This is also where the default openssl.cnf gets
3837 If the directory given with this option is a relative path, the
3838 values of both the --prefix value and the --openssldir value will
3839 be combined to become OPENSSLDIR.
3840 The default for --openssldir is INSTALLTOP/ssl.
3842 Anyone who uses --openssldir to specify where OpenSSL is to be
3843 installed MUST change to use --prefix instead.
3847 * The GOST engine was out of date and therefore it has been removed. An up
3848 to date GOST engine is now being maintained in an external repository.
3849 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3850 support for GOST ciphersuites (these are only activated if a GOST engine
3855 * EGD is no longer supported by default; use enable-egd when
3858 *Ben Kaduk and Rich Salz*
3860 * The distribution now has Makefile.in files, which are used to
3861 create Makefile's when Configure is run. *Configure must be run
3862 before trying to build now.*
3866 * The return value for SSL_CIPHER_description() for error conditions
3871 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3873 Obtaining and performing DNSSEC validation of TLSA records is
3874 the application's responsibility. The application provides
3875 the TLSA records of its choice to OpenSSL, and these are then
3876 used to authenticate the peer.
3878 The TLSA records need not even come from DNS. They can, for
3879 example, be used to implement local end-entity certificate or
3880 trust-anchor "pinning", where the "pin" data takes the form
3881 of TLSA records, which can augment or replace verification
3882 based on the usual WebPKI public certification authorities.
3886 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3887 continues to support deprecated interfaces in default builds.
3888 However, applications are strongly advised to compile their
3889 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3890 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3891 or the 1.1.0 releases.
3893 In environments in which all applications have been ported to
3894 not use any deprecated interfaces OpenSSL's Configure script
3895 should be used with the --api=1.1.0 option to entirely remove
3896 support for the deprecated features from the library and
3897 unconditionally disable them in the installed headers.
3898 Essentially the same effect can be achieved with the "no-deprecated"
3899 argument to Configure, except that this will always restrict
3900 the build to just the latest API, rather than a fixed API
3903 As applications are ported to future revisions of the API,
3904 they should update their compile-time OPENSSL_API_COMPAT define
3905 accordingly, but in most cases should be able to continue to
3906 compile with later releases.
3908 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3909 0x10000000L and 0x00908000L, respectively. However those
3910 versions did not support the OPENSSL_API_COMPAT feature, and
3911 so applications are not typically tested for explicit support
3912 of just the undeprecated features of either release.
3916 * Add support for setting the minimum and maximum supported protocol.
3917 It can bet set via the SSL_set_min_proto_version() and
3918 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3919 MaxProtocol. It's recommended to use the new APIs to disable
3920 protocols instead of disabling individual protocols using
3921 SSL_set_options() or SSL_CONF's Protocol. This change also
3922 removes support for disabling TLS 1.2 in the OpenSSL TLS
3923 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3927 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3931 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3932 and integrates ECDSA and ECDH functionality into EC. Implementations can
3933 now redirect key generation and no longer need to convert to or from
3936 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3937 include the ec.h header file instead.
3941 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3942 ciphers who are no longer supported and drops support the ephemeral RSA key
3943 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3947 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3948 opaque. For HMAC_CTX, the following constructors and destructors
3951 HMAC_CTX *HMAC_CTX_new(void);
3952 void HMAC_CTX_free(HMAC_CTX *ctx);
3954 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3955 destroy such methods has been added. See EVP_MD_meth_new(3) and
3956 EVP_CIPHER_meth_new(3) for documentation.
3959 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3960 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3961 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3962 an already created structure.
3963 2) For consistency with the majority of our object creators and
3964 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3965 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3966 for deprecated builds.
3970 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3971 cryptographic operations to be performed asynchronously as long as an
3972 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3973 further details. Libssl has also had this capability integrated with the
3974 introduction of the new mode SSL_MODE_ASYNC and associated error
3975 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3976 pages. This work was developed in partnership with Intel Corp.
3980 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3981 always enabled now. If you want to disable the support you should
3982 exclude it using the list of supported ciphers. This also means that the
3983 "-no_ecdhe" option has been removed from s_server.
3987 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3988 SSL_{CTX_}set1_curves() which can set a list.
3992 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3993 curve you want to support using SSL_{CTX_}set1_curves().
3997 * State machine rewrite. The state machine code has been significantly
3998 refactored in order to remove much duplication of code and solve issues
3999 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4000 further details). This change does have some associated API changes.
4001 Notably the SSL_state() function has been removed and replaced by
4002 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4003 SSL_set_state() has been removed altogether. The previous handshake states
4004 defined in ssl.h and ssl3.h have also been removed.
4008 * All instances of the string "ssleay" in the public API were replaced
4009 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4010 Some error codes related to internal RSA_eay API's were renamed.
4014 * The demo files in crypto/threads were moved to demo/threads.
4018 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4021 *Matt Caswell, Rich Salz*
4023 * New ASN.1 embed macro.
4025 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4026 structure is not allocated: it is part of the parent. That is instead of
4034 This reduces memory fragmentation and make it impossible to accidentally
4035 set a mandatory field to NULL.
4037 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4038 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4039 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4044 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4048 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4049 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4050 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4051 DES and RC4 ciphersuites.
4055 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4056 This changes the decoding behaviour for some invalid messages,
4057 though the change is mostly in the more lenient direction, and
4058 legacy behaviour is preserved as much as possible.
4062 * Fix no-stdio build.
4063 *David Woodhouse <David.Woodhouse@intel.com> and also*
4064 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4066 * New testing framework
4067 The testing framework has been largely rewritten and is now using
4068 perl and the perl modules Test::Harness and an extended variant of
4069 Test::More called OpenSSL::Test to do its work. All test scripts in
4070 test/ have been rewritten into test recipes, and all direct calls to
4071 executables in test/Makefile have become individual recipes using the
4072 simplified testing OpenSSL::Test::Simple.
4074 For documentation on our testing modules, do:
4076 perldoc test/testlib/OpenSSL/Test/Simple.pm
4077 perldoc test/testlib/OpenSSL/Test.pm
4081 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4082 are used; the latter aborts on memory leaks (usually checked on exit).
4083 Some undocumented "set malloc, etc., hooks" functions were removed
4084 and others were changed. All are now documented.
4088 * In DSA_generate_parameters_ex, if the provided seed is too short,
4091 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4093 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4094 from RFC4279, RFC4785, RFC5487, RFC5489.
4096 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4097 original RSA_PSK patch.
4101 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4102 era flag was never set throughout the codebase (only read). Also removed
4103 SSL3_FLAGS_POP_BUFFER which was only used if
4104 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4108 * Changed the default name options in the "ca", "crl", "req" and "x509"
4109 to be "oneline" instead of "compat".
4113 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4114 not aware of clients that still exhibit this bug, and the workaround
4115 hasn't been working properly for a while.
4119 * The return type of BIO_number_read() and BIO_number_written() as well as
4120 the corresponding num_read and num_write members in the BIO structure has
4121 changed from unsigned long to uint64_t. On platforms where an unsigned
4122 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4127 * Given the pervasive nature of TLS extensions it is inadvisable to run
4128 OpenSSL without support for them. It also means that maintaining
4129 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4130 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4134 * Removed support for the two export grade static DH ciphersuites
4135 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4136 were newly added (along with a number of other static DH ciphersuites) to
4137 1.0.2. However the two export ones have *never* worked since they were
4138 introduced. It seems strange in any case to be adding new export
4139 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4143 * Version negotiation has been rewritten. In particular SSLv23_method(),
4144 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4145 and turned into macros which simply call the new preferred function names
4146 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4147 should use the new names instead. Also as part of this change the ssl23.h
4148 header file has been removed.
4152 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4153 code and the associated standard is no longer considered fit-for-purpose.
4157 * RT2547 was closed. When generating a private key, try to make the
4158 output file readable only by the owner. This behavior change might
4159 be noticeable when interacting with other software.
4161 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4166 * Added HTTP GET support to the ocsp command.
4170 * Changed default digest for the dgst and enc commands from MD5 to
4175 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4179 * Added support for TLS extended master secret from
4180 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4181 initial patch which was a great help during development.
4185 * All libssl internal structures have been removed from the public header
4186 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4187 now redundant). Users should not attempt to access internal structures
4188 directly. Instead they should use the provided API functions.
4192 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4193 Access to deprecated functions can be re-enabled by running config with
4194 "enable-deprecated". In addition applications wishing to use deprecated
4195 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4196 will, by default, disable some transitive includes that previously existed
4197 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4201 * Added support for OCB mode. OpenSSL has been granted a patent license
4202 compatible with the OpenSSL license for use of OCB. Details are available
4203 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4204 for OCB can be removed by calling config with no-ocb.
4208 * SSLv2 support has been removed. It still supports receiving a SSLv2
4209 compatible client hello.
4213 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4214 done while fixing the error code for the key-too-small case.
4216 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4218 * CA.sh has been removed; use CA.pl instead.
4222 * Removed old DES API.
4226 * Remove various unsupported platforms:
4232 Sinix/ReliantUNIX RM400
4237 16-bit platforms such as WIN16
4241 * Clean up OPENSSL_NO_xxx #define's
4242 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4243 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4244 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4245 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4246 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4247 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4248 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4249 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4250 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4251 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4255 * Cleaned up dead code
4256 Remove all but one '#ifdef undef' which is to be looked at.
4260 * Clean up calling of xxx_free routines.
4261 Just like free(), fix most of the xxx_free routines to accept
4262 NULL. Remove the non-null checks from callers. Save much code.
4266 * Add secure heap for storage of private keys (when possible).
4267 Add BIO_s_secmem(), CBIGNUM, etc.
4268 Contributed by Akamai Technologies under our Corporate CLA.
4272 * Experimental support for a new, fast, unbiased prime candidate generator,
4273 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4275 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4277 * New output format NSS in the sess_id command line tool. This allows
4278 exporting the session id and the master key in NSS keylog format.
4280 *Martin Kaiser <martin@kaiser.cx>*
4282 * Harmonize version and its documentation. -f flag is used to display
4285 *mancha <mancha1@zoho.com>*
4287 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4288 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4290 *mancha <mancha1@zoho.com>*
4292 * Fix some double frees. These are not thought to be exploitable.
4294 *mancha <mancha1@zoho.com>*
4296 * A missing bounds check in the handling of the TLS heartbeat extension
4297 can be used to reveal up to 64k of memory to a connected client or
4300 Thanks for Neel Mehta of Google Security for discovering this bug and to
4301 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4302 preparing the fix ([CVE-2014-0160])
4304 *Adam Langley, Bodo Moeller*
4306 * Fix for the attack described in the paper "Recovering OpenSSL
4307 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4308 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4309 <http://eprint.iacr.org/2014/140>
4311 Thanks to Yuval Yarom and Naomi Benger for discovering this
4312 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4314 *Yuval Yarom and Naomi Benger*
4316 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4317 this fixes a limitation in previous versions of OpenSSL.
4321 * Experimental encrypt-then-mac support.
4323 Experimental support for encrypt then mac from
4324 draft-gutmann-tls-encrypt-then-mac-02.txt
4326 To enable it set the appropriate extension number (0x42 for the test
4327 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4329 For non-compliant peers (i.e. just about everything) this should have no
4332 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4336 * Add EVP support for key wrapping algorithms, to avoid problems with
4337 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4338 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4339 algorithms and include tests cases.
4343 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4348 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4349 MGF1 digest and OAEP label.
4353 * Make openssl verify return errors.
4355 *Chris Palmer <palmer@google.com> and Ben Laurie*
4357 * New function ASN1_TIME_diff to calculate the difference between two
4358 ASN1_TIME structures or one structure and the current time.
4362 * Update fips_test_suite to support multiple command line options. New
4363 test to induce all self test errors in sequence and check expected
4368 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4369 sign or verify all in one operation.
4373 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4374 test programs and fips_test_suite. Includes functionality to parse
4375 the minimal script output of fipsalgest.pl directly.
4379 * Add authorisation parameter to FIPS_module_mode_set().
4383 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4387 * Use separate DRBG fields for internal and external flags. New function
4388 FIPS_drbg_health_check() to perform on demand health checking. Add
4389 generation tests to fips_test_suite with reduced health check interval to
4390 demonstrate periodic health checking. Add "nodh" option to
4391 fips_test_suite to skip very slow DH test.
4395 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4400 * More extensive health check for DRBG checking many more failure modes.
4401 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4402 combination: call this in fips_test_suite.
4406 * Add support for canonical generation of DSA parameter 'g'. See
4409 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4410 POST to handle HMAC cases.
4414 * Add functions FIPS_module_version() and FIPS_module_version_text()
4415 to return numerical and string versions of the FIPS module number.
4419 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4420 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4421 outside the validated module in the FIPS capable OpenSSL.
4425 * Minor change to DRBG entropy callback semantics. In some cases
4426 there is no multiple of the block length between min_len and
4427 max_len. Allow the callback to return more than max_len bytes
4428 of entropy but discard any extra: it is the callback's responsibility
4429 to ensure that the extra data discarded does not impact the
4430 requested amount of entropy.
4434 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4435 information in FIPS186-3, SP800-57 and SP800-131A.
4439 * CCM support via EVP. Interface is very similar to GCM case except we
4440 must supply all data in one chunk (i.e. no update, final) and the
4441 message length must be supplied if AAD is used. Add algorithm test
4446 * Initial version of POST overhaul. Add POST callback to allow the status
4447 of POST to be monitored and/or failures induced. Modify fips_test_suite
4448 to use callback. Always run all selftests even if one fails.
4452 * XTS support including algorithm test driver in the fips_gcmtest program.
4453 Note: this does increase the maximum key length from 32 to 64 bytes but
4454 there should be no binary compatibility issues as existing applications
4455 will never use XTS mode.
4459 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4460 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4461 performs algorithm blocking for unapproved PRNG types. Also do not
4462 set PRNG type in FIPS_mode_set(): leave this to the application.
4463 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4464 the standard OpenSSL PRNG: set additional data to a date time vector.
4468 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4469 This shouldn't present any incompatibility problems because applications
4470 shouldn't be using these directly and any that are will need to rethink
4471 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4475 * Extensive self tests and health checking required by SP800-90 DRBG.
4476 Remove strength parameter from FIPS_drbg_instantiate and always
4477 instantiate at maximum supported strength.
4481 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4485 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4489 * New function DH_compute_key_padded() to compute a DH key and pad with
4490 leading zeroes if needed: this complies with SP800-56A et al.
4494 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4495 anything, incomplete, subject to change and largely untested at present.
4499 * Modify fipscanisteronly build option to only build the necessary object
4500 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4504 * Add experimental option FIPSSYMS to give all symbols in
4505 fipscanister.o and FIPS or fips prefix. This will avoid
4506 conflicts with future versions of OpenSSL. Add perl script
4507 util/fipsas.pl to preprocess assembly language source files
4508 and rename any affected symbols.
4512 * Add selftest checks and algorithm block of non-fips algorithms in
4513 FIPS mode. Remove DES2 from selftests.
4517 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4518 return internal method without any ENGINE dependencies. Add new
4519 tiny fips sign and verify functions.
4523 * New build option no-ec2m to disable characteristic 2 code.
4527 * New build option "fipscanisteronly". This only builds fipscanister.o
4528 and (currently) associated fips utilities. Uses the file Makefile.fips
4529 instead of Makefile.org as the prototype.
4533 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4534 Update fips_gcmtest to use IV generator.
4538 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4539 setting output buffer to NULL. The `*Final` function must be
4540 called although it will not retrieve any additional data. The tag
4541 can be set or retrieved with a ctrl. The IV length is by default 12
4542 bytes (96 bits) but can be set to an alternative value. If the IV
4543 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4548 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4549 underlying do_cipher function handles all cipher semantics itself
4550 including padding and finalisation. This is useful if (for example)
4551 an ENGINE cipher handles block padding itself. The behaviour of
4552 do_cipher is subtly changed if this flag is set: the return value
4553 is the number of characters written to the output buffer (zero is
4554 no longer an error code) or a negative error code. Also if the
4555 input buffer is NULL and length 0 finalisation should be performed.
4559 * If a candidate issuer certificate is already part of the constructed
4560 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4564 * Improve forward-security support: add functions
4566 void SSL_CTX_set_not_resumable_session_callback(
4567 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4568 void SSL_set_not_resumable_session_callback(
4569 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4571 for use by SSL/TLS servers; the callback function will be called whenever a
4572 new session is created, and gets to decide whether the session may be
4573 cached to make it resumable (return 0) or not (return 1). (As by the
4574 SSL/TLS protocol specifications, the session_id sent by the server will be
4575 empty to indicate that the session is not resumable; also, the server will
4576 not generate RFC 4507 (RFC 5077) session tickets.)
4578 A simple reasonable callback implementation is to return is_forward_secure.
4579 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4580 by the SSL/TLS server library, indicating whether it can provide forward
4583 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4585 * New -verify_name option in command line utilities to set verification
4590 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4591 Add CMAC pkey methods.
4595 * Experimental renegotiation in s_server -www mode. If the client
4596 browses /reneg connection is renegotiated. If /renegcert it is
4597 renegotiated requesting a certificate.
4601 * Add an "external" session cache for debugging purposes to s_server. This
4602 should help trace issues which normally are only apparent in deployed
4603 multi-process servers.
4607 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4608 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4609 BIO_set_cipher() and some obscure PEM functions were changed so they
4610 can now return an error. The RAND changes required a change to the
4611 RAND_METHOD structure.
4615 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4616 a gcc attribute to warn if the result of a function is ignored. This
4617 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4618 whose return value is often ignored.
4622 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4623 These allow SCTs (signed certificate timestamps) to be requested and
4624 validated when establishing a connection.
4626 *Rob Percival <robpercival@google.com>*
4631 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4633 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4634 used even when parsing explicit parameters, when loading a encoded key
4635 or calling `EC_GROUP_new_from_ecpkparameters()`/
4636 `EC_GROUP_new_from_ecparameters()`.
4637 This prevents bypass of security hardening and performance gains,
4638 especially for curves with specialized EC_METHODs.
4639 By default, if a key encoded with explicit parameters is loaded and later
4640 encoded, the output is still encoded with explicit parameters, even if
4641 internally a "named" EC_GROUP is used for computation.
4645 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4646 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4647 NULL. After this change, only the cofactor parameter can be NULL. It also
4648 does some minimal sanity checks on the passed order.
4653 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4654 An attack is simple, if the first CMS_recipientInfo is valid but the
4655 second CMS_recipientInfo is chosen ciphertext. If the second
4656 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4657 encryption key will be replaced by garbage, and the message cannot be
4658 decoded, but if the RSA decryption fails, the correct encryption key is
4659 used and the recipient will not notice the attack.
4660 As a work around for this potential attack the length of the decrypted
4661 key must be equal to the cipher default key length, in case the
4662 certifiate is not given and all recipientInfo are tried out.
4663 The old behaviour can be re-enabled in the CMS code by setting the
4664 CMS_DEBUG_DECRYPT flag.
4669 * Document issue with installation paths in diverse Windows builds
4671 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4672 binaries and run-time config file.
4677 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4679 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4680 This changes the size when using the `genpkey` command when no size is given.
4681 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4682 generation commands to use 2048 bits by default.
4686 * Add FIPS support for Android Arm 64-bit
4688 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4689 Module in Version 2.0.10. For some reason, the corresponding target
4690 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4691 built with FIPS support on Android Arm 64-bit. This omission has been
4694 *Matthias St. Pierre*
4696 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4698 * 0-byte record padding oracle
4700 If an application encounters a fatal protocol error and then calls
4701 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4702 then OpenSSL can respond differently to the calling application if a 0 byte
4703 record is received with invalid padding compared to if a 0 byte record is
4704 received with an invalid MAC. If the application then behaves differently
4705 based on that in a way that is detectable to the remote peer, then this
4706 amounts to a padding oracle that could be used to decrypt data.
4708 In order for this to be exploitable "non-stitched" ciphersuites must be in
4709 use. Stitched ciphersuites are optimised implementations of certain
4710 commonly used ciphersuites. Also the application must call SSL_shutdown()
4711 twice even if a protocol error has occurred (applications should not do
4712 this but some do anyway).
4714 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4715 Aviram, with additional investigation by Steven Collison and Andrew
4716 Hourselt. It was reported to OpenSSL on 10th December 2018.
4721 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4725 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4727 * Microarchitecture timing vulnerability in ECC scalar multiplication
4729 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4730 shown to be vulnerable to a microarchitecture timing side channel attack.
4731 An attacker with sufficient access to mount local timing attacks during
4732 ECDSA signature generation could recover the private key.
4734 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4735 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4741 * Timing vulnerability in DSA signature generation
4743 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4744 timing side channel attack. An attacker could use variations in the signing
4745 algorithm to recover the private key.
4747 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4752 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4753 Module, accidentally introduced while backporting security fixes from the
4754 development branch and hindering the use of ECC in FIPS mode.
4758 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4760 * Client DoS due to large DH parameter
4762 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4763 malicious server can send a very large prime value to the client. This will
4764 cause the client to spend an unreasonably long period of time generating a
4765 key for this prime resulting in a hang until the client has finished. This
4766 could be exploited in a Denial Of Service attack.
4768 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4773 * Cache timing vulnerability in RSA Key Generation
4775 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4776 a cache timing side channel attack. An attacker with sufficient access to
4777 mount cache timing attacks during the RSA key generation process could
4778 recover the private key.
4780 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4781 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4786 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4787 parameter is no longer accepted, as it leads to a corrupt table. NULL
4788 pem_str is reserved for alias entries only.
4792 * Revert blinding in ECDSA sign and instead make problematic addition
4793 length-invariant. Switch even to fixed-length Montgomery multiplication.
4797 * Change generating and checking of primes so that the error rate of not
4798 being prime depends on the intended use based on the size of the input.
4799 For larger primes this will result in more rounds of Miller-Rabin.
4800 The maximal error rate for primes with more than 1080 bits is lowered
4803 *Kurt Roeckx, Annie Yousar*
4805 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4809 * Add blinding to ECDSA and DSA signatures to protect against side channel
4810 attacks discovered by Keegan Ryan (NCC Group).
4814 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4815 now allow empty (zero character) pass phrases.
4819 * Certificate time validation (X509_cmp_time) enforces stricter
4820 compliance with RFC 5280. Fractional seconds and timezone offsets
4821 are no longer allowed.
4825 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4827 * Constructed ASN.1 types with a recursive definition could exceed the stack
4829 Constructed ASN.1 types with a recursive definition (such as can be found
4830 in PKCS7) could eventually exceed the stack given malicious input with
4831 excessive recursion. This could result in a Denial Of Service attack. There
4832 are no such structures used within SSL/TLS that come from untrusted sources
4833 so this is considered safe.
4835 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4841 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4843 * Read/write after SSL object in error state
4845 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4846 mechanism. The intent was that if a fatal error occurred during a handshake
4847 then OpenSSL would move into the error state and would immediately fail if
4848 you attempted to continue the handshake. This works as designed for the
4849 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4850 SSL_connect()), however due to a bug it does not work correctly if
4851 SSL_read() or SSL_write() is called directly. In that scenario, if the
4852 handshake fails then a fatal error will be returned in the initial function
4853 call. If SSL_read()/SSL_write() is subsequently called by the application
4854 for the same SSL object then it will succeed and the data is passed without
4855 being decrypted/encrypted directly from the SSL/TLS record layer.
4857 In order to exploit this issue an application bug would have to be present
4858 that resulted in a call to SSL_read()/SSL_write() being issued after having
4859 already received a fatal error.
4861 This issue was reported to OpenSSL by David Benjamin (Google).
4866 * rsaz_1024_mul_avx2 overflow bug on x86_64
4868 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4869 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4870 Analysis suggests that attacks against RSA and DSA as a result of this
4871 defect would be very difficult to perform and are not believed likely.
4872 Attacks against DH1024 are considered just feasible, because most of the
4873 work necessary to deduce information about a private key may be performed
4874 offline. The amount of resources required for such an attack would be
4875 significant. However, for an attack on TLS to be meaningful, the server
4876 would have to share the DH1024 private key among multiple clients, which is
4877 no longer an option since CVE-2016-0701.
4879 This only affects processors that support the AVX2 but not ADX extensions
4880 like Intel Haswell (4th generation).
4882 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4883 was originally found via the OSS-Fuzz project.
4888 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4890 * bn_sqrx8x_internal carry bug on x86_64
4892 There is a carry propagating bug in the x86_64 Montgomery squaring
4893 procedure. No EC algorithms are affected. Analysis suggests that attacks
4894 against RSA and DSA as a result of this defect would be very difficult to
4895 perform and are not believed likely. Attacks against DH are considered just
4896 feasible (although very difficult) because most of the work necessary to
4897 deduce information about a private key may be performed offline. The amount
4898 of resources required for such an attack would be very significant and
4899 likely only accessible to a limited number of attackers. An attacker would
4900 additionally need online access to an unpatched system using the target
4901 private key in a scenario with persistent DH parameters and a private
4902 key that is shared between multiple clients.
4904 This only affects processors that support the BMI1, BMI2 and ADX extensions
4905 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4907 This issue was reported to OpenSSL by the OSS-Fuzz project.
4912 * Malformed X.509 IPAddressFamily could cause OOB read
4914 If an X.509 certificate has a malformed IPAddressFamily extension,
4915 OpenSSL could do a one-byte buffer overread. The most likely result
4916 would be an erroneous display of the certificate in text format.
4918 This issue was reported to OpenSSL by the OSS-Fuzz project.
4923 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4925 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4926 platform rather than 'mingw'.
4930 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4932 * Truncated packet could crash via OOB read
4934 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4935 cipher is being used, then a truncated packet can cause that host to
4936 perform an out-of-bounds read, usually resulting in a crash.
4938 This issue was reported to OpenSSL by Robert Święcki of Google.
4943 * BN_mod_exp may produce incorrect results on x86_64
4945 There is a carry propagating bug in the x86_64 Montgomery squaring
4946 procedure. No EC algorithms are affected. Analysis suggests that attacks
4947 against RSA and DSA as a result of this defect would be very difficult to
4948 perform and are not believed likely. Attacks against DH are considered just
4949 feasible (although very difficult) because most of the work necessary to
4950 deduce information about a private key may be performed offline. The amount
4951 of resources required for such an attack would be very significant and
4952 likely only accessible to a limited number of attackers. An attacker would
4953 additionally need online access to an unpatched system using the target
4954 private key in a scenario with persistent DH parameters and a private
4955 key that is shared between multiple clients. For example this can occur by
4956 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4957 similar to CVE-2015-3193 but must be treated as a separate problem.
4959 This issue was reported to OpenSSL by the OSS-Fuzz project.
4964 * Montgomery multiplication may produce incorrect results
4966 There is a carry propagating bug in the Broadwell-specific Montgomery
4967 multiplication procedure that handles input lengths divisible by, but
4968 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4969 and DH private keys are impossible. This is because the subroutine in
4970 question is not used in operations with the private key itself and an input
4971 of the attacker's direct choice. Otherwise the bug can manifest itself as
4972 transient authentication and key negotiation failures or reproducible
4973 erroneous outcome of public-key operations with specially crafted input.
4974 Among EC algorithms only Brainpool P-512 curves are affected and one
4975 presumably can attack ECDH key negotiation. Impact was not analyzed in
4976 detail, because pre-requisites for attack are considered unlikely. Namely
4977 multiple clients have to choose the curve in question and the server has to
4978 share the private key among them, neither of which is default behaviour.
4979 Even then only clients that chose the curve will be affected.
4981 This issue was publicly reported as transient failures and was not
4982 initially recognized as a security issue. Thanks to Richard Morgan for
4983 providing reproducible case.
4988 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4989 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4990 prevent issues where no progress is being made and the peer continually
4991 sends unrecognised record types, using up resources processing them.
4995 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4997 * Missing CRL sanity check
4999 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5000 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5001 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5003 This issue only affects the OpenSSL 1.0.2i
5008 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5010 * OCSP Status Request extension unbounded memory growth
5012 A malicious client can send an excessively large OCSP Status Request
5013 extension. If that client continually requests renegotiation, sending a
5014 large OCSP Status Request extension each time, then there will be unbounded
5015 memory growth on the server. This will eventually lead to a Denial Of
5016 Service attack through memory exhaustion. Servers with a default
5017 configuration are vulnerable even if they do not support OCSP. Builds using
5018 the "no-ocsp" build time option are not affected.
5020 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5025 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5028 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5034 * OOB write in MDC2_Update()
5036 An overflow can occur in MDC2_Update() either if called directly or
5037 through the EVP_DigestUpdate() function using MDC2. If an attacker
5038 is able to supply very large amounts of input data after a previous
5039 call to EVP_EncryptUpdate() with a partial block then a length check
5040 can overflow resulting in a heap corruption.
5042 The amount of data needed is comparable to SIZE_MAX which is impractical
5045 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5050 * Malformed SHA512 ticket DoS
5052 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5053 DoS attack where a malformed ticket will result in an OOB read which will
5056 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5057 a custom server callback and ticket lookup mechanism.
5059 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5064 * OOB write in BN_bn2dec()
5066 The function BN_bn2dec() does not check the return value of BN_div_word().
5067 This can cause an OOB write if an application uses this function with an
5068 overly large BIGNUM. This could be a problem if an overly large certificate
5069 or CRL is printed out from an untrusted source. TLS is not affected because
5070 record limits will reject an oversized certificate before it is parsed.
5072 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5077 * OOB read in TS_OBJ_print_bio()
5079 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5080 the total length the OID text representation would use and not the amount
5081 of data written. This will result in OOB reads when large OIDs are
5084 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5089 * Pointer arithmetic undefined behaviour
5091 Avoid some undefined pointer arithmetic
5093 A common idiom in the codebase is to check limits in the following manner:
5096 Where "p" points to some malloc'd data of SIZE bytes and
5099 "len" here could be from some externally supplied data (e.g. from a TLS
5102 The rules of C pointer arithmetic are such that "p + len" is only well
5103 defined where len <= SIZE. Therefore the above idiom is actually
5104 undefined behaviour.
5106 For example this could cause problems if some malloc implementation
5107 provides an address for "p" such that "p + len" actually overflows for
5108 values of len that are too big and therefore p + len < limit.
5110 This issue was reported to OpenSSL by Guido Vranken
5115 * Constant time flag not preserved in DSA signing
5117 Operations in the DSA signing algorithm should run in constant time in
5118 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5119 implementation means that a non-constant time codepath is followed for
5120 certain operations. This has been demonstrated through a cache-timing
5121 attack to be sufficient for an attacker to recover the private DSA key.
5123 This issue was reported by César Pereida (Aalto University), Billy Brumley
5124 (Tampere University of Technology), and Yuval Yarom (The University of
5125 Adelaide and NICTA).
5130 * DTLS buffered message DoS
5132 In a DTLS connection where handshake messages are delivered out-of-order
5133 those messages that OpenSSL is not yet ready to process will be buffered
5134 for later use. Under certain circumstances, a flaw in the logic means that
5135 those messages do not get removed from the buffer even though the handshake
5136 has been completed. An attacker could force up to approx. 15 messages to
5137 remain in the buffer when they are no longer required. These messages will
5138 be cleared when the DTLS connection is closed. The default maximum size for
5139 a message is 100k. Therefore the attacker could force an additional 1500k
5140 to be consumed per connection. By opening many simulataneous connections an
5141 attacker could cause a DoS attack through memory exhaustion.
5143 This issue was reported to OpenSSL by Quan Luo.
5148 * DTLS replay protection DoS
5150 A flaw in the DTLS replay attack protection mechanism means that records
5151 that arrive for future epochs update the replay protection "window" before
5152 the MAC for the record has been validated. This could be exploited by an
5153 attacker by sending a record for the next epoch (which does not have to
5154 decrypt or have a valid MAC), with a very large sequence number. This means
5155 that all subsequent legitimate packets are dropped causing a denial of
5156 service for a specific DTLS connection.
5158 This issue was reported to OpenSSL by the OCAP audit team.
5163 * Certificate message OOB reads
5165 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5166 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5167 theoretical DoS risk but this has not been observed in practice on common
5170 The messages affected are client certificate, client certificate request
5171 and server certificate. As a result the attack can only be performed
5172 against a client or a server which enables client authentication.
5174 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5179 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5181 * Prevent padding oracle in AES-NI CBC MAC check
5183 A MITM attacker can use a padding oracle attack to decrypt traffic
5184 when the connection uses an AES CBC cipher and the server support
5187 This issue was introduced as part of the fix for Lucky 13 padding
5188 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5189 constant time by making sure that always the same bytes are read and
5190 compared against either the MAC or padding bytes. But it no longer
5191 checked that there was enough data to have both the MAC and padding
5194 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5199 * Fix EVP_EncodeUpdate overflow
5201 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5202 Base64 encoding of binary data. If an attacker is able to supply very large
5203 amounts of input data then a length check can overflow resulting in a heap
5206 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5207 the `PEM_write_bio*` family of functions. These are mainly used within the
5208 OpenSSL command line applications, so any application which processes data
5209 from an untrusted source and outputs it as a PEM file should be considered
5210 vulnerable to this issue. User applications that call these APIs directly
5211 with large amounts of untrusted data may also be vulnerable.
5213 This issue was reported by Guido Vranken.
5218 * Fix EVP_EncryptUpdate overflow
5220 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5221 is able to supply very large amounts of input data after a previous call to
5222 EVP_EncryptUpdate() with a partial block then a length check can overflow
5223 resulting in a heap corruption. Following an analysis of all OpenSSL
5224 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5225 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5226 the first called function after an EVP_EncryptInit(), and therefore that
5227 specific call must be safe. The second form is where the length passed to
5228 EVP_EncryptUpdate() can be seen from the code to be some small value and
5229 therefore there is no possibility of an overflow. Since all instances are
5230 one of these two forms, it is believed that there can be no overflows in
5231 internal code due to this problem. It should be noted that
5232 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5233 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5234 of these calls have also been analysed too and it is believed there are no
5235 instances in internal usage where an overflow could occur.
5237 This issue was reported by Guido Vranken.
5242 * Prevent ASN.1 BIO excessive memory allocation
5244 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5245 a short invalid encoding can cause allocation of large amounts of memory
5246 potentially consuming excessive resources or exhausting memory.
5248 Any application parsing untrusted data through d2i BIO functions is
5249 affected. The memory based functions such as d2i_X509() are *not* affected.
5250 Since the memory based functions are used by the TLS library, TLS
5251 applications are not affected.
5253 This issue was reported by Brian Carpenter.
5260 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5261 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5262 in arbitrary stack data being returned in the buffer.
5264 This issue was reported by Guido Vranken.
5269 * Modify behavior of ALPN to invoke callback after SNI/servername
5270 callback, such that updates to the SSL_CTX affect ALPN.
5274 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5279 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5280 methods are enabled and ssl2 is disabled the methods return NULL.
5284 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5286 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5287 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5288 provide any "EXPORT" or "LOW" strength ciphers.
5292 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5293 is by default disabled at build-time. Builds that are not configured with
5294 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5295 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5296 will need to explicitly call either of:
5298 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5300 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5302 as appropriate. Even if either of those is used, or the application
5303 explicitly uses the version-specific SSLv2_method() or its client and
5304 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5305 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5306 ciphers, and SSLv2 56-bit DES are no longer available.
5311 * Fix a double-free in DSA code
5313 A double free bug was discovered when OpenSSL parses malformed DSA private
5314 keys and could lead to a DoS attack or memory corruption for applications
5315 that receive DSA private keys from untrusted sources. This scenario is
5318 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5324 * Disable SRP fake user seed to address a server memory leak.
5326 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5328 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5329 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5330 was changed to ignore the "fake user" SRP seed, even if the seed
5333 Users should use SRP_VBASE_get1_by_user instead. Note that in
5334 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5335 also that even though configuring the SRP seed attempts to hide
5336 invalid usernames by continuing the handshake with fake
5337 credentials, this behaviour is not constant time and no strong
5338 guarantees are made that the handshake is indistinguishable from
5339 that of a valid user.
5344 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5346 In the BN_hex2bn function the number of hex digits is calculated using an
5347 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5348 large values of `i` this can result in `bn_expand` not allocating any
5349 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5350 field as NULL leading to a subsequent NULL ptr deref. For very large values
5351 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5352 In this case memory is allocated to the internal BIGNUM data field, but it
5353 is insufficiently sized leading to heap corruption. A similar issue exists
5354 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5355 is ever called by user applications with very large untrusted hex/dec data.
5356 This is anticipated to be a rare occurrence.
5358 All OpenSSL internal usage of these functions use data that is not expected
5359 to be untrusted, e.g. config file data or application command line
5360 arguments. If user developed applications generate config file data based
5361 on untrusted data then it is possible that this could also lead to security
5362 consequences. This is also anticipated to be rare.
5364 This issue was reported to OpenSSL by Guido Vranken.
5369 * Fix memory issues in `BIO_*printf` functions
5371 The internal `fmtstr` function used in processing a "%s" format string in
5372 the `BIO_*printf` functions could overflow while calculating the length of a
5373 string and cause an OOB read when printing very long strings.
5375 Additionally the internal `doapr_outch` function can attempt to write to an
5376 OOB memory location (at an offset from the NULL pointer) in the event of a
5377 memory allocation failure. In 1.0.2 and below this could be caused where
5378 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5379 could be in processing a very long "%s" format string. Memory leaks can
5382 The first issue may mask the second issue dependent on compiler behaviour.
5383 These problems could enable attacks where large amounts of untrusted data
5384 is passed to the `BIO_*printf` functions. If applications use these functions
5385 in this way then they could be vulnerable. OpenSSL itself uses these
5386 functions when printing out human-readable dumps of ASN.1 data. Therefore
5387 applications that print this data could be vulnerable if the data is from
5388 untrusted sources. OpenSSL command line applications could also be
5389 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5390 as command line arguments.
5392 Libssl is not considered directly vulnerable. Additionally certificates etc
5393 received via remote connections via libssl are also unlikely to be able to
5394 trigger these issues because of message size limits enforced within libssl.
5396 This issue was reported to OpenSSL Guido Vranken.
5401 * Side channel attack on modular exponentiation
5403 A side-channel attack was found which makes use of cache-bank conflicts on
5404 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5405 of RSA keys. The ability to exploit this issue is limited as it relies on
5406 an attacker who has control of code in a thread running on the same
5407 hyper-threaded core as the victim thread which is performing decryptions.
5409 This issue was reported to OpenSSL by Yuval Yarom, The University of
5410 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5411 Nadia Heninger, University of Pennsylvania with more information at
5412 <http://cachebleed.info>.
5417 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5418 if no keysize is specified with default_bits. This fixes an
5419 omission in an earlier change that changed all RSA/DSA key generation
5420 commands to use 2048 bits by default.
5424 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5426 * DH small subgroups
5428 Historically OpenSSL only ever generated DH parameters based on "safe"
5429 primes. More recently (in version 1.0.2) support was provided for
5430 generating X9.42 style parameter files such as those required for RFC 5114
5431 support. The primes used in such files may not be "safe". Where an
5432 application is using DH configured with parameters based on primes that are
5433 not "safe" then an attacker could use this fact to find a peer's private
5434 DH exponent. This attack requires that the attacker complete multiple
5435 handshakes in which the peer uses the same private DH exponent. For example
5436 this could be used to discover a TLS server's private DH exponent if it's
5437 reusing the private DH exponent or it's using a static DH ciphersuite.
5439 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5440 TLS. It is not on by default. If the option is not set then the server
5441 reuses the same private DH exponent for the life of the server process and
5442 would be vulnerable to this attack. It is believed that many popular
5443 applications do set this option and would therefore not be at risk.
5445 The fix for this issue adds an additional check where a "q" parameter is
5446 available (as is the case in X9.42 based parameters). This detects the
5447 only known attack, and is the only possible defense for static DH
5448 ciphersuites. This could have some performance impact.
5450 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5451 default and cannot be disabled. This could have some performance impact.
5453 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5458 * SSLv2 doesn't block disabled ciphers
5460 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5461 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5462 been disabled, provided that the SSLv2 protocol was not also disabled via
5465 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5466 and Sebastian Schinzel.
5471 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5473 * BN_mod_exp may produce incorrect results on x86_64
5475 There is a carry propagating bug in the x86_64 Montgomery squaring
5476 procedure. No EC algorithms are affected. Analysis suggests that attacks
5477 against RSA and DSA as a result of this defect would be very difficult to
5478 perform and are not believed likely. Attacks against DH are considered just
5479 feasible (although very difficult) because most of the work necessary to
5480 deduce information about a private key may be performed offline. The amount
5481 of resources required for such an attack would be very significant and
5482 likely only accessible to a limited number of attackers. An attacker would
5483 additionally need online access to an unpatched system using the target
5484 private key in a scenario with persistent DH parameters and a private
5485 key that is shared between multiple clients. For example this can occur by
5486 default in OpenSSL DHE based SSL/TLS ciphersuites.
5488 This issue was reported to OpenSSL by Hanno Böck.
5493 * Certificate verify crash with missing PSS parameter
5495 The signature verification routines will crash with a NULL pointer
5496 dereference if presented with an ASN.1 signature using the RSA PSS
5497 algorithm and absent mask generation function parameter. Since these
5498 routines are used to verify certificate signature algorithms this can be
5499 used to crash any certificate verification operation and exploited in a
5500 DoS attack. Any application which performs certificate verification is
5501 vulnerable including OpenSSL clients and servers which enable client
5504 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5509 * X509_ATTRIBUTE memory leak
5511 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5512 memory. This structure is used by the PKCS#7 and CMS routines so any
5513 application which reads PKCS#7 or CMS data from untrusted sources is
5514 affected. SSL/TLS is not affected.
5516 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5522 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5523 This changes the decoding behaviour for some invalid messages,
5524 though the change is mostly in the more lenient direction, and
5525 legacy behaviour is preserved as much as possible.
5529 * In DSA_generate_parameters_ex, if the provided seed is too short,
5532 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5534 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5536 * Alternate chains certificate forgery
5538 During certificate verification, OpenSSL will attempt to find an
5539 alternative certificate chain if the first attempt to build such a chain
5540 fails. An error in the implementation of this logic can mean that an
5541 attacker could cause certain checks on untrusted certificates to be
5542 bypassed, such as the CA flag, enabling them to use a valid leaf
5543 certificate to act as a CA and "issue" an invalid certificate.
5545 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5550 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5552 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5553 incompatibility in the handling of HMAC. The previous ABI has now been
5558 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5560 * Malformed ECParameters causes infinite loop
5562 When processing an ECParameters structure OpenSSL enters an infinite loop
5563 if the curve specified is over a specially malformed binary polynomial
5566 This can be used to perform denial of service against any
5567 system which processes public keys, certificate requests or
5568 certificates. This includes TLS clients and TLS servers with
5569 client authentication enabled.
5571 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5576 * Exploitable out-of-bounds read in X509_cmp_time
5578 X509_cmp_time does not properly check the length of the ASN1_TIME
5579 string and can read a few bytes out of bounds. In addition,
5580 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5583 An attacker can use this to craft malformed certificates and CRLs of
5584 various sizes and potentially cause a segmentation fault, resulting in
5585 a DoS on applications that verify certificates or CRLs. TLS clients
5586 that verify CRLs are affected. TLS clients and servers with client
5587 authentication enabled may be affected if they use custom verification
5590 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5591 independently by Hanno Böck.
5596 * PKCS7 crash with missing EnvelopedContent
5598 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5599 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5600 with missing content and trigger a NULL pointer dereference on parsing.
5602 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5603 structures from untrusted sources are affected. OpenSSL clients and
5604 servers are not affected.
5606 This issue was reported to OpenSSL by Michal Zalewski (Google).
5611 * CMS verify infinite loop with unknown hash function
5613 When verifying a signedData message the CMS code can enter an infinite loop
5614 if presented with an unknown hash function OID. This can be used to perform
5615 denial of service against any system which verifies signedData messages using
5617 This issue was reported to OpenSSL by Johannes Bauer.
5622 * Race condition handling NewSessionTicket
5624 If a NewSessionTicket is received by a multi-threaded client when attempting to
5625 reuse a previous ticket then a race condition can occur potentially leading to
5626 a double free of the ticket data.
5631 * Only support 256-bit or stronger elliptic curves with the
5632 'ecdh_auto' setting (server) or by default (client). Of supported
5633 curves, prefer P-256 (both).
5637 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5639 * ClientHello sigalgs DoS fix
5641 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5642 invalid signature algorithms extension a NULL pointer dereference will
5643 occur. This can be exploited in a DoS attack against the server.
5645 This issue was was reported to OpenSSL by David Ramos of Stanford
5649 *Stephen Henson and Matt Caswell*
5651 * Multiblock corrupted pointer fix
5653 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5654 feature only applies on 64 bit x86 architecture platforms that support AES
5655 NI instructions. A defect in the implementation of "multiblock" can cause
5656 OpenSSL's internal write buffer to become incorrectly set to NULL when
5657 using non-blocking IO. Typically, when the user application is using a
5658 socket BIO for writing, this will only result in a failed connection.
5659 However if some other BIO is used then it is likely that a segmentation
5660 fault will be triggered, thus enabling a potential DoS attack.
5662 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5667 * Segmentation fault in DTLSv1_listen fix
5669 The DTLSv1_listen function is intended to be stateless and processes the
5670 initial ClientHello from many peers. It is common for user code to loop
5671 over the call to DTLSv1_listen until a valid ClientHello is received with
5672 an associated cookie. A defect in the implementation of DTLSv1_listen means
5673 that state is preserved in the SSL object from one invocation to the next
5674 that can lead to a segmentation fault. Errors processing the initial
5675 ClientHello can trigger this scenario. An example of such an error could be
5676 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5679 This issue was reported to OpenSSL by Per Allansson.
5684 * Segmentation fault in ASN1_TYPE_cmp fix
5686 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5687 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5688 certificate signature algorithm consistency this can be used to crash any
5689 certificate verification operation and exploited in a DoS attack. Any
5690 application which performs certificate verification is vulnerable including
5691 OpenSSL clients and servers which enable client authentication.
5696 * Segmentation fault for invalid PSS parameters fix
5698 The signature verification routines will crash with a NULL pointer
5699 dereference if presented with an ASN.1 signature using the RSA PSS
5700 algorithm and invalid parameters. Since these routines are used to verify
5701 certificate signature algorithms this can be used to crash any
5702 certificate verification operation and exploited in a DoS attack. Any
5703 application which performs certificate verification is vulnerable including
5704 OpenSSL clients and servers which enable client authentication.
5706 This issue was was reported to OpenSSL by Brian Carpenter.
5711 * ASN.1 structure reuse memory corruption fix
5713 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5714 memory corruption via an invalid write. Such reuse is and has been
5715 strongly discouraged and is believed to be rare.
5717 Applications that parse structures containing CHOICE or ANY DEFINED BY
5718 components may be affected. Certificate parsing (d2i_X509 and related
5719 functions) are however not affected. OpenSSL clients and servers are
5725 * PKCS7 NULL pointer dereferences fix
5727 The PKCS#7 parsing code does not handle missing outer ContentInfo
5728 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5729 missing content and trigger a NULL pointer dereference on parsing.
5731 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5732 otherwise parse PKCS#7 structures from untrusted sources are
5733 affected. OpenSSL clients and servers are not affected.
5735 This issue was reported to OpenSSL by Michal Zalewski (Google).
5740 * DoS via reachable assert in SSLv2 servers fix
5742 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5743 servers that both support SSLv2 and enable export cipher suites by sending
5744 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5746 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5747 (OpenSSL development team).
5752 * Empty CKE with client auth and DHE fix
5754 If client auth is used then a server can seg fault in the event of a DHE
5755 ciphersuite being selected and a zero length ClientKeyExchange message
5756 being sent by the client. This could be exploited in a DoS attack.
5761 * Handshake with unseeded PRNG fix
5763 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5764 with an unseeded PRNG. The conditions are:
5765 - The client is on a platform where the PRNG has not been seeded
5766 automatically, and the user has not seeded manually
5767 - A protocol specific client method version has been used (i.e. not
5768 SSL_client_methodv23)
5769 - A ciphersuite is used that does not require additional random data from
5770 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5772 If the handshake succeeds then the client random that has been used will
5773 have been generated from a PRNG with insufficient entropy and therefore the
5774 output may be predictable.
5776 For example using the following command with an unseeded openssl will
5777 succeed on an unpatched platform:
5779 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5784 * Use After Free following d2i_ECPrivatekey error fix
5786 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5787 could cause a use after free condition. This, in turn, could cause a double
5788 free in several private key parsing functions (such as d2i_PrivateKey
5789 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5790 for applications that receive EC private keys from untrusted
5791 sources. This scenario is considered rare.
5793 This issue was discovered by the BoringSSL project and fixed in their
5799 * X509_to_X509_REQ NULL pointer deref fix
5801 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5802 the certificate key is invalid. This function is rarely used in practice.
5804 This issue was discovered by Brian Carpenter.
5809 * Removed the export ciphers from the DEFAULT ciphers
5813 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5815 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5816 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5817 So far those who have to target multiple platforms would compromise
5818 and argue that binary targeting say ARMv5 would still execute on
5819 ARMv8. "Universal" build resolves this compromise by providing
5820 near-optimal performance even on newer platforms.
5824 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5825 (other platforms pending).
5827 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5829 * Add support for the SignedCertificateTimestampList certificate and
5830 OCSP response extensions from RFC6962.
5834 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5835 for corner cases. (Certain input points at infinity could lead to
5836 bogus results, with non-infinity inputs mapped to infinity too.)
5840 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5841 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5842 common cases are optimized and there still is room for further
5843 improvements. Vector Permutation AES for Altivec is also added.
5847 * Add support for little-endian ppc64 Linux target.
5849 *Marcelo Cerri (IBM)*
5851 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5852 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5853 are optimized and there still is room for further improvements.
5854 Both 32- and 64-bit modes are supported.
5856 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5858 * Improved ARMv7 NEON support.
5862 * Support for SPARC Architecture 2011 crypto extensions, first
5863 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5864 SHA256/512, MD5, GHASH and modular exponentiation.
5866 *Andy Polyakov, David Miller*
5868 * Accelerated modular exponentiation for Intel processors, a.k.a.
5871 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5873 * Support for new and upcoming Intel processors, including AVX2,
5874 BMI and SHA ISA extensions. This includes additional "stitched"
5875 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5878 This work was sponsored by Intel Corp.
5882 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5883 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5884 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5888 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5889 this fixes a limitation in previous versions of OpenSSL.
5893 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5894 MGF1 digest and OAEP label.
5898 * Add EVP support for key wrapping algorithms, to avoid problems with
5899 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5900 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5901 algorithms and include tests cases.
5905 * Add functions to allocate and set the fields of an ECDSA_METHOD
5908 *Douglas E. Engert, Steve Henson*
5910 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5911 difference in days and seconds between two tm or ASN1_TIME structures.
5915 * Add -rev test option to s_server to just reverse order of characters
5916 received by client and send back to server. Also prints an abbreviated
5917 summary of the connection parameters.
5921 * New option -brief for s_client and s_server to print out a brief summary
5922 of connection parameters.
5926 * Add callbacks for arbitrary TLS extensions.
5928 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5930 * New option -crl_download in several openssl utilities to download CRLs
5931 from CRLDP extension in certificates.
5935 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5939 * New function X509_CRL_diff to generate a delta CRL from the difference
5940 of two full CRLs. Add support to "crl" utility.
5944 * New functions to set lookup_crls function and to retrieve
5945 X509_STORE from X509_STORE_CTX.
5949 * Print out deprecated issuer and subject unique ID fields in
5954 * Extend OCSP I/O functions so they can be used for simple general purpose
5955 HTTP as well as OCSP. New wrapper function which can be used to download
5956 CRLs using the OCSP API.
5960 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5964 * `SSL_CONF*` functions. These provide a common framework for application
5965 configuration using configuration files or command lines.
5969 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5970 message callback and prints the results. Needs compile time option
5971 "enable-ssl-trace". New options to s_client and s_server to enable
5976 * New ctrl and macro to retrieve supported points extensions.
5977 Print out extension in s_server and s_client.
5981 * New functions to retrieve certificate signature and signature
5986 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5991 * New Suite B modes for TLS code. These use and enforce the requirements
5992 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5993 only use Suite B curves. The Suite B modes can be set by using the
5994 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5998 * New chain verification flags for Suite B levels of security. Check
5999 algorithms are acceptable when flags are set in X509_verify_cert.
6003 * Make tls1_check_chain return a set of flags indicating checks passed
6004 by a certificate chain. Add additional tests to handle client
6005 certificates: checks for matching certificate type and issuer name
6010 * If an attempt is made to use a signature algorithm not in the peer
6011 preference list abort the handshake. If client has no suitable
6012 signature algorithms in response to a certificate request do not
6013 use the certificate.
6017 * If server EC tmp key is not in client preference list abort handshake.
6021 * Add support for certificate stores in CERT structure. This makes it
6022 possible to have different stores per SSL structure or one store in
6023 the parent SSL_CTX. Include distinct stores for certificate chain
6024 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6025 to build and store a certificate chain in CERT structure: returning
6026 an error if the chain cannot be built: this will allow applications
6027 to test if a chain is correctly configured.
6029 Note: if the CERT based stores are not set then the parent SSL_CTX
6030 store is used to retain compatibility with existing behaviour.
6034 * New function ssl_set_client_disabled to set a ciphersuite disabled
6035 mask based on the current session, check mask when sending client
6036 hello and checking the requested ciphersuite.
6040 * New ctrls to retrieve and set certificate types in a certificate
6041 request message. Print out received values in s_client. If certificate
6042 types is not set with custom values set sensible values based on
6043 supported signature algorithms.
6047 * Support for distinct client and server supported signature algorithms.
6051 * Add certificate callback. If set this is called whenever a certificate
6052 is required by client or server. An application can decide which
6053 certificate chain to present based on arbitrary criteria: for example
6054 supported signature algorithms. Add very simple example to s_server.
6055 This fixes many of the problems and restrictions of the existing client
6056 certificate callback: for example you can now clear an existing
6057 certificate and specify the whole chain.
6061 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6062 the certificate can be used for (if anything). Set valid_flags field
6063 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6064 to have similar checks in it.
6066 Add new "cert_flags" field to CERT structure and include a "strict mode".
6067 This enforces some TLS certificate requirements (such as only permitting
6068 certificate signature algorithms contained in the supported algorithms
6069 extension) which some implementations ignore: this option should be used
6070 with caution as it could cause interoperability issues.
6074 * Update and tidy signature algorithm extension processing. Work out
6075 shared signature algorithms based on preferences and peer algorithms
6076 and print them out in s_client and s_server. Abort handshake if no
6077 shared signature algorithms.
6081 * Add new functions to allow customised supported signature algorithms
6082 for SSL and SSL_CTX structures. Add options to s_client and s_server
6087 * New function SSL_certs_clear() to delete all references to certificates
6088 from an SSL structure. Before this once a certificate had been added
6089 it couldn't be removed.
6093 * Integrate hostname, email address and IP address checking with certificate
6094 verification. New verify options supporting checking in openssl utility.
6098 * Fixes and wildcard matching support to hostname and email checking
6099 functions. Add manual page.
6101 *Florian Weimer (Red Hat Product Security Team)*
6103 * New functions to check a hostname email or IP address against a
6104 certificate. Add options x509 utility to print results of checks against
6109 * Fix OCSP checking.
6111 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6113 * Initial experimental support for explicitly trusted non-root CAs.
6114 OpenSSL still tries to build a complete chain to a root but if an
6115 intermediate CA has a trust setting included that is used. The first
6116 setting is used: whether to trust (e.g., -addtrust option to the x509
6121 * Add -trusted_first option which attempts to find certificates in the
6122 trusted store even if an untrusted chain is also supplied.
6126 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6127 platform support for Linux and Android.
6131 * Support for linux-x32, ILP32 environment in x86_64 framework.
6135 * Experimental multi-implementation support for FIPS capable OpenSSL.
6136 When in FIPS mode the approved implementations are used as normal,
6137 when not in FIPS mode the internal unapproved versions are used instead.
6138 This means that the FIPS capable OpenSSL isn't forced to use the
6139 (often lower performance) FIPS implementations outside FIPS mode.
6143 * Transparently support X9.42 DH parameters when calling
6144 PEM_read_bio_DHparameters. This means existing applications can handle
6145 the new parameter format automatically.
6149 * Initial experimental support for X9.42 DH parameter format: mainly
6150 to support use of 'q' parameter for RFC5114 parameters.
6154 * Add DH parameters from RFC5114 including test data to dhtest.
6158 * Support for automatic EC temporary key parameter selection. If enabled
6159 the most preferred EC parameters are automatically used instead of
6160 hardcoded fixed parameters. Now a server just has to call:
6161 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6162 support ECDH and use the most appropriate parameters.
6166 * Enhance and tidy EC curve and point format TLS extension code. Use
6167 static structures instead of allocation if default values are used.
6168 New ctrls to set curves we wish to support and to retrieve shared curves.
6169 Print out shared curves in s_server. New options to s_server and s_client
6170 to set list of supported curves.
6174 * New ctrls to retrieve supported signature algorithms and
6175 supported curve values as an array of NIDs. Extend openssl utility
6176 to print out received values.
6180 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6181 between NIDs and the more common NIST names such as "P-256". Enhance
6182 ecparam utility and ECC method to recognise the NIST names for curves.
6186 * Enhance SSL/TLS certificate chain handling to support different
6187 chains for each certificate instead of one chain in the parent SSL_CTX.
6191 * Support for fixed DH ciphersuite client authentication: where both
6192 server and client use DH certificates with common parameters.
6196 * Support for fixed DH ciphersuites: those requiring DH server
6201 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6203 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6204 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6205 X509_CINF_get_signature were reverted post internal team review.
6210 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6212 * OCSP Status Request extension unbounded memory growth
6214 A malicious client can send an excessively large OCSP Status Request
6215 extension. If that client continually requests renegotiation, sending a
6216 large OCSP Status Request extension each time, then there will be unbounded
6217 memory growth on the server. This will eventually lead to a Denial Of
6218 Service attack through memory exhaustion. Servers with a default
6219 configuration are vulnerable even if they do not support OCSP. Builds using
6220 the "no-ocsp" build time option are not affected.
6222 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6227 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6230 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6236 * OOB write in MDC2_Update()
6238 An overflow can occur in MDC2_Update() either if called directly or
6239 through the EVP_DigestUpdate() function using MDC2. If an attacker
6240 is able to supply very large amounts of input data after a previous
6241 call to EVP_EncryptUpdate() with a partial block then a length check
6242 can overflow resulting in a heap corruption.
6244 The amount of data needed is comparable to SIZE_MAX which is impractical
6247 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6252 * Malformed SHA512 ticket DoS
6254 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6255 DoS attack where a malformed ticket will result in an OOB read which will
6258 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6259 a custom server callback and ticket lookup mechanism.
6261 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6266 * OOB write in BN_bn2dec()
6268 The function BN_bn2dec() does not check the return value of BN_div_word().
6269 This can cause an OOB write if an application uses this function with an
6270 overly large BIGNUM. This could be a problem if an overly large certificate
6271 or CRL is printed out from an untrusted source. TLS is not affected because
6272 record limits will reject an oversized certificate before it is parsed.
6274 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6279 * OOB read in TS_OBJ_print_bio()
6281 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6282 the total length the OID text representation would use and not the amount
6283 of data written. This will result in OOB reads when large OIDs are
6286 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6291 * Pointer arithmetic undefined behaviour
6293 Avoid some undefined pointer arithmetic
6295 A common idiom in the codebase is to check limits in the following manner:
6298 Where "p" points to some malloc'd data of SIZE bytes and
6301 "len" here could be from some externally supplied data (e.g. from a TLS
6304 The rules of C pointer arithmetic are such that "p + len" is only well
6305 defined where len <= SIZE. Therefore the above idiom is actually
6306 undefined behaviour.
6308 For example this could cause problems if some malloc implementation
6309 provides an address for "p" such that "p + len" actually overflows for
6310 values of len that are too big and therefore p + len < limit.
6312 This issue was reported to OpenSSL by Guido Vranken
6317 * Constant time flag not preserved in DSA signing
6319 Operations in the DSA signing algorithm should run in constant time in
6320 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6321 implementation means that a non-constant time codepath is followed for
6322 certain operations. This has been demonstrated through a cache-timing
6323 attack to be sufficient for an attacker to recover the private DSA key.
6325 This issue was reported by César Pereida (Aalto University), Billy Brumley
6326 (Tampere University of Technology), and Yuval Yarom (The University of
6327 Adelaide and NICTA).
6332 * DTLS buffered message DoS
6334 In a DTLS connection where handshake messages are delivered out-of-order
6335 those messages that OpenSSL is not yet ready to process will be buffered
6336 for later use. Under certain circumstances, a flaw in the logic means that
6337 those messages do not get removed from the buffer even though the handshake
6338 has been completed. An attacker could force up to approx. 15 messages to
6339 remain in the buffer when they are no longer required. These messages will
6340 be cleared when the DTLS connection is closed. The default maximum size for
6341 a message is 100k. Therefore the attacker could force an additional 1500k
6342 to be consumed per connection. By opening many simulataneous connections an
6343 attacker could cause a DoS attack through memory exhaustion.
6345 This issue was reported to OpenSSL by Quan Luo.
6350 * DTLS replay protection DoS
6352 A flaw in the DTLS replay attack protection mechanism means that records
6353 that arrive for future epochs update the replay protection "window" before
6354 the MAC for the record has been validated. This could be exploited by an
6355 attacker by sending a record for the next epoch (which does not have to
6356 decrypt or have a valid MAC), with a very large sequence number. This means
6357 that all subsequent legitimate packets are dropped causing a denial of
6358 service for a specific DTLS connection.
6360 This issue was reported to OpenSSL by the OCAP audit team.
6365 * Certificate message OOB reads
6367 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6368 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6369 theoretical DoS risk but this has not been observed in practice on common
6372 The messages affected are client certificate, client certificate request
6373 and server certificate. As a result the attack can only be performed
6374 against a client or a server which enables client authentication.
6376 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6381 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6383 * Prevent padding oracle in AES-NI CBC MAC check
6385 A MITM attacker can use a padding oracle attack to decrypt traffic
6386 when the connection uses an AES CBC cipher and the server support
6389 This issue was introduced as part of the fix for Lucky 13 padding
6390 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6391 constant time by making sure that always the same bytes are read and
6392 compared against either the MAC or padding bytes. But it no longer
6393 checked that there was enough data to have both the MAC and padding
6396 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6401 * Fix EVP_EncodeUpdate overflow
6403 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6404 Base64 encoding of binary data. If an attacker is able to supply very large
6405 amounts of input data then a length check can overflow resulting in a heap
6408 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6409 the `PEM_write_bio*` family of functions. These are mainly used within the
6410 OpenSSL command line applications, so any application which processes data
6411 from an untrusted source and outputs it as a PEM file should be considered
6412 vulnerable to this issue. User applications that call these APIs directly
6413 with large amounts of untrusted data may also be vulnerable.
6415 This issue was reported by Guido Vranken.
6420 * Fix EVP_EncryptUpdate overflow
6422 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6423 is able to supply very large amounts of input data after a previous call to
6424 EVP_EncryptUpdate() with a partial block then a length check can overflow
6425 resulting in a heap corruption. Following an analysis of all OpenSSL
6426 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6427 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6428 the first called function after an EVP_EncryptInit(), and therefore that
6429 specific call must be safe. The second form is where the length passed to
6430 EVP_EncryptUpdate() can be seen from the code to be some small value and
6431 therefore there is no possibility of an overflow. Since all instances are
6432 one of these two forms, it is believed that there can be no overflows in
6433 internal code due to this problem. It should be noted that
6434 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6435 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6436 of these calls have also been analysed too and it is believed there are no
6437 instances in internal usage where an overflow could occur.
6439 This issue was reported by Guido Vranken.
6444 * Prevent ASN.1 BIO excessive memory allocation
6446 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6447 a short invalid encoding can casuse allocation of large amounts of memory
6448 potentially consuming excessive resources or exhausting memory.
6450 Any application parsing untrusted data through d2i BIO functions is
6451 affected. The memory based functions such as d2i_X509() are *not* affected.
6452 Since the memory based functions are used by the TLS library, TLS
6453 applications are not affected.
6455 This issue was reported by Brian Carpenter.
6462 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6463 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6464 in arbitrary stack data being returned in the buffer.
6466 This issue was reported by Guido Vranken.
6471 * Modify behavior of ALPN to invoke callback after SNI/servername
6472 callback, such that updates to the SSL_CTX affect ALPN.
6476 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6481 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6482 methods are enabled and ssl2 is disabled the methods return NULL.
6486 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6488 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6489 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6490 provide any "EXPORT" or "LOW" strength ciphers.
6494 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6495 is by default disabled at build-time. Builds that are not configured with
6496 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6497 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6498 will need to explicitly call either of:
6500 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6502 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6504 as appropriate. Even if either of those is used, or the application
6505 explicitly uses the version-specific SSLv2_method() or its client and
6506 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6507 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6508 ciphers, and SSLv2 56-bit DES are no longer available.
6513 * Fix a double-free in DSA code
6515 A double free bug was discovered when OpenSSL parses malformed DSA private
6516 keys and could lead to a DoS attack or memory corruption for applications
6517 that receive DSA private keys from untrusted sources. This scenario is
6520 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6526 * Disable SRP fake user seed to address a server memory leak.
6528 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6530 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6531 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6532 was changed to ignore the "fake user" SRP seed, even if the seed
6535 Users should use SRP_VBASE_get1_by_user instead. Note that in
6536 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6537 also that even though configuring the SRP seed attempts to hide
6538 invalid usernames by continuing the handshake with fake
6539 credentials, this behaviour is not constant time and no strong
6540 guarantees are made that the handshake is indistinguishable from
6541 that of a valid user.
6546 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6548 In the BN_hex2bn function the number of hex digits is calculated using an
6549 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6550 large values of `i` this can result in `bn_expand` not allocating any
6551 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6552 field as NULL leading to a subsequent NULL ptr deref. For very large values
6553 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6554 In this case memory is allocated to the internal BIGNUM data field, but it
6555 is insufficiently sized leading to heap corruption. A similar issue exists
6556 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6557 is ever called by user applications with very large untrusted hex/dec data.
6558 This is anticipated to be a rare occurrence.
6560 All OpenSSL internal usage of these functions use data that is not expected
6561 to be untrusted, e.g. config file data or application command line
6562 arguments. If user developed applications generate config file data based
6563 on untrusted data then it is possible that this could also lead to security
6564 consequences. This is also anticipated to be rare.
6566 This issue was reported to OpenSSL by Guido Vranken.
6571 * Fix memory issues in `BIO_*printf` functions
6573 The internal `fmtstr` function used in processing a "%s" format string in
6574 the `BIO_*printf` functions could overflow while calculating the length of a
6575 string and cause an OOB read when printing very long strings.
6577 Additionally the internal `doapr_outch` function can attempt to write to an
6578 OOB memory location (at an offset from the NULL pointer) in the event of a
6579 memory allocation failure. In 1.0.2 and below this could be caused where
6580 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6581 could be in processing a very long "%s" format string. Memory leaks can
6584 The first issue may mask the second issue dependent on compiler behaviour.
6585 These problems could enable attacks where large amounts of untrusted data
6586 is passed to the `BIO_*printf` functions. If applications use these functions
6587 in this way then they could be vulnerable. OpenSSL itself uses these
6588 functions when printing out human-readable dumps of ASN.1 data. Therefore
6589 applications that print this data could be vulnerable if the data is from
6590 untrusted sources. OpenSSL command line applications could also be
6591 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6592 as command line arguments.
6594 Libssl is not considered directly vulnerable. Additionally certificates etc
6595 received via remote connections via libssl are also unlikely to be able to
6596 trigger these issues because of message size limits enforced within libssl.
6598 This issue was reported to OpenSSL Guido Vranken.
6603 * Side channel attack on modular exponentiation
6605 A side-channel attack was found which makes use of cache-bank conflicts on
6606 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6607 of RSA keys. The ability to exploit this issue is limited as it relies on
6608 an attacker who has control of code in a thread running on the same
6609 hyper-threaded core as the victim thread which is performing decryptions.
6611 This issue was reported to OpenSSL by Yuval Yarom, The University of
6612 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6613 Nadia Heninger, University of Pennsylvania with more information at
6614 <http://cachebleed.info>.
6619 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6620 if no keysize is specified with default_bits. This fixes an
6621 omission in an earlier change that changed all RSA/DSA key generation
6622 commands to use 2048 bits by default.
6626 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6628 * Protection for DH small subgroup attacks
6630 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6631 switched on by default and cannot be disabled. This could have some
6636 * SSLv2 doesn't block disabled ciphers
6638 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6639 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6640 been disabled, provided that the SSLv2 protocol was not also disabled via
6643 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6644 and Sebastian Schinzel.
6649 * Reject DH handshakes with parameters shorter than 1024 bits.
6653 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6655 * Certificate verify crash with missing PSS parameter
6657 The signature verification routines will crash with a NULL pointer
6658 dereference if presented with an ASN.1 signature using the RSA PSS
6659 algorithm and absent mask generation function parameter. Since these
6660 routines are used to verify certificate signature algorithms this can be
6661 used to crash any certificate verification operation and exploited in a
6662 DoS attack. Any application which performs certificate verification is
6663 vulnerable including OpenSSL clients and servers which enable client
6666 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6671 * X509_ATTRIBUTE memory leak
6673 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6674 memory. This structure is used by the PKCS#7 and CMS routines so any
6675 application which reads PKCS#7 or CMS data from untrusted sources is
6676 affected. SSL/TLS is not affected.
6678 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6684 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6685 This changes the decoding behaviour for some invalid messages,
6686 though the change is mostly in the more lenient direction, and
6687 legacy behaviour is preserved as much as possible.
6691 * In DSA_generate_parameters_ex, if the provided seed is too short,
6692 use a random seed, as already documented.
6694 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6696 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6698 * Alternate chains certificate forgery
6700 During certificate verfification, OpenSSL will attempt to find an
6701 alternative certificate chain if the first attempt to build such a chain
6702 fails. An error in the implementation of this logic can mean that an
6703 attacker could cause certain checks on untrusted certificates to be
6704 bypassed, such as the CA flag, enabling them to use a valid leaf
6705 certificate to act as a CA and "issue" an invalid certificate.
6707 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6713 * Race condition handling PSK identify hint
6715 If PSK identity hints are received by a multi-threaded client then
6716 the values are wrongly updated in the parent SSL_CTX structure. This can
6717 result in a race condition potentially leading to a double free of the
6723 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6725 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6726 incompatibility in the handling of HMAC. The previous ABI has now been
6729 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6731 * Malformed ECParameters causes infinite loop
6733 When processing an ECParameters structure OpenSSL enters an infinite loop
6734 if the curve specified is over a specially malformed binary polynomial
6737 This can be used to perform denial of service against any
6738 system which processes public keys, certificate requests or
6739 certificates. This includes TLS clients and TLS servers with
6740 client authentication enabled.
6742 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6747 * Exploitable out-of-bounds read in X509_cmp_time
6749 X509_cmp_time does not properly check the length of the ASN1_TIME
6750 string and can read a few bytes out of bounds. In addition,
6751 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6754 An attacker can use this to craft malformed certificates and CRLs of
6755 various sizes and potentially cause a segmentation fault, resulting in
6756 a DoS on applications that verify certificates or CRLs. TLS clients
6757 that verify CRLs are affected. TLS clients and servers with client
6758 authentication enabled may be affected if they use custom verification
6761 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6762 independently by Hanno Böck.
6767 * PKCS7 crash with missing EnvelopedContent
6769 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6770 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6771 with missing content and trigger a NULL pointer dereference on parsing.
6773 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6774 structures from untrusted sources are affected. OpenSSL clients and
6775 servers are not affected.
6777 This issue was reported to OpenSSL by Michal Zalewski (Google).
6782 * CMS verify infinite loop with unknown hash function
6784 When verifying a signedData message the CMS code can enter an infinite loop
6785 if presented with an unknown hash function OID. This can be used to perform
6786 denial of service against any system which verifies signedData messages using
6788 This issue was reported to OpenSSL by Johannes Bauer.
6793 * Race condition handling NewSessionTicket
6795 If a NewSessionTicket is received by a multi-threaded client when attempting to
6796 reuse a previous ticket then a race condition can occur potentially leading to
6797 a double free of the ticket data.
6802 * Reject DH handshakes with parameters shorter than 768 bits.
6804 *Kurt Roeckx and Emilia Kasper*
6806 * dhparam: generate 2048-bit parameters by default.
6808 *Kurt Roeckx and Emilia Kasper*
6810 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6812 * Segmentation fault in ASN1_TYPE_cmp fix
6814 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6815 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6816 certificate signature algorithm consistency this can be used to crash any
6817 certificate verification operation and exploited in a DoS attack. Any
6818 application which performs certificate verification is vulnerable including
6819 OpenSSL clients and servers which enable client authentication.
6824 * ASN.1 structure reuse memory corruption fix
6826 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6827 memory corruption via an invalid write. Such reuse is and has been
6828 strongly discouraged and is believed to be rare.
6830 Applications that parse structures containing CHOICE or ANY DEFINED BY
6831 components may be affected. Certificate parsing (d2i_X509 and related
6832 functions) are however not affected. OpenSSL clients and servers are
6838 * PKCS7 NULL pointer dereferences fix
6840 The PKCS#7 parsing code does not handle missing outer ContentInfo
6841 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6842 missing content and trigger a NULL pointer dereference on parsing.
6844 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6845 otherwise parse PKCS#7 structures from untrusted sources are
6846 affected. OpenSSL clients and servers are not affected.
6848 This issue was reported to OpenSSL by Michal Zalewski (Google).
6853 * DoS via reachable assert in SSLv2 servers fix
6855 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6856 servers that both support SSLv2 and enable export cipher suites by sending
6857 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6859 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6860 (OpenSSL development team).
6865 * Use After Free following d2i_ECPrivatekey error fix
6867 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6868 could cause a use after free condition. This, in turn, could cause a double
6869 free in several private key parsing functions (such as d2i_PrivateKey
6870 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6871 for applications that receive EC private keys from untrusted
6872 sources. This scenario is considered rare.
6874 This issue was discovered by the BoringSSL project and fixed in their
6880 * X509_to_X509_REQ NULL pointer deref fix
6882 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6883 the certificate key is invalid. This function is rarely used in practice.
6885 This issue was discovered by Brian Carpenter.
6890 * Removed the export ciphers from the DEFAULT ciphers
6894 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6896 * Build fixes for the Windows and OpenVMS platforms
6898 *Matt Caswell and Richard Levitte*
6900 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6902 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6903 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6904 dereference. This could lead to a Denial Of Service attack. Thanks to
6905 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6910 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6911 dtls1_buffer_record function under certain conditions. In particular this
6912 could occur if an attacker sent repeated DTLS records with the same
6913 sequence number but for the next epoch. The memory leak could be exploited
6914 by an attacker in a Denial of Service attack through memory exhaustion.
6915 Thanks to Chris Mueller for reporting this issue.
6920 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6921 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6922 method would be set to NULL which could later result in a NULL pointer
6923 dereference. Thanks to Frank Schmirler for reporting this issue.
6928 * Abort handshake if server key exchange message is omitted for ephemeral
6931 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6932 reporting this issue.
6937 * Remove non-export ephemeral RSA code on client and server. This code
6938 violated the TLS standard by allowing the use of temporary RSA keys in
6939 non-export ciphersuites and could be used by a server to effectively
6940 downgrade the RSA key length used to a value smaller than the server
6941 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6942 INRIA or reporting this issue.
6947 * Fixed issue where DH client certificates are accepted without verification.
6948 An OpenSSL server will accept a DH certificate for client authentication
6949 without the certificate verify message. This effectively allows a client to
6950 authenticate without the use of a private key. This only affects servers
6951 which trust a client certificate authority which issues certificates
6952 containing DH keys: these are extremely rare and hardly ever encountered.
6953 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6959 * Ensure that the session ID context of an SSL is updated when its
6960 SSL_CTX is updated via SSL_set_SSL_CTX.
6962 The session ID context is typically set from the parent SSL_CTX,
6963 and can vary with the CTX.
6967 * Fix various certificate fingerprint issues.
6969 By using non-DER or invalid encodings outside the signed portion of a
6970 certificate the fingerprint can be changed without breaking the signature.
6971 Although no details of the signed portion of the certificate can be changed
6972 this can cause problems with some applications: e.g. those using the
6973 certificate fingerprint for blacklists.
6975 1. Reject signatures with non zero unused bits.
6977 If the BIT STRING containing the signature has non zero unused bits reject
6978 the signature. All current signature algorithms require zero unused bits.
6980 2. Check certificate algorithm consistency.
6982 Check the AlgorithmIdentifier inside TBS matches the one in the
6983 certificate signature. NB: this will result in signature failure
6984 errors for some broken certificates.
6986 Thanks to Konrad Kraszewski from Google for reporting this issue.
6988 3. Check DSA/ECDSA signatures use DER.
6990 Re-encode DSA/ECDSA signatures and compare with the original received
6991 signature. Return an error if there is a mismatch.
6993 This will reject various cases including garbage after signature
6994 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6995 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6996 (negative or with leading zeroes).
6998 Further analysis was conducted and fixes were developed by Stephen Henson
6999 of the OpenSSL core team.
7005 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7006 results on some platforms, including x86_64. This bug occurs at random
7007 with a very low probability, and is not known to be exploitable in any
7008 way, though its exact impact is difficult to determine. Thanks to Pieter
7009 Wuille (Blockstream) who reported this issue and also suggested an initial
7010 fix. Further analysis was conducted by the OpenSSL development team and
7011 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7012 the OpenSSL core team.
7017 * Do not resume sessions on the server if the negotiated protocol
7018 version does not match the session's version. Resuming with a different
7019 version, while not strictly forbidden by the RFC, is of questionable
7020 sanity and breaks all known clients.
7022 *David Benjamin, Emilia Käsper*
7024 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7025 early CCS messages during renegotiation. (Note that because
7026 renegotiation is encrypted, this early CCS was not exploitable.)
7030 * Tighten client-side session ticket handling during renegotiation:
7031 ensure that the client only accepts a session ticket if the server sends
7032 the extension anew in the ServerHello. Previously, a TLS client would
7033 reuse the old extension state and thus accept a session ticket if one was
7034 announced in the initial ServerHello.
7036 Similarly, ensure that the client requires a session ticket if one
7037 was advertised in the ServerHello. Previously, a TLS client would
7038 ignore a missing NewSessionTicket message.
7042 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7046 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7047 sends a carefully crafted handshake message, to cause OpenSSL to fail
7048 to free up to 64k of memory causing a memory leak. This could be
7049 exploited in a Denial Of Service attack. This issue affects OpenSSL
7050 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7051 whether SRTP is used or configured. Implementations of OpenSSL that
7052 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7054 The fix was developed by the OpenSSL team.
7059 * Session Ticket Memory Leak.
7061 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7062 integrity of that ticket is first verified. In the event of a session
7063 ticket integrity check failing, OpenSSL will fail to free memory
7064 causing a memory leak. By sending a large number of invalid session
7065 tickets an attacker could exploit this issue in a Denial Of Service
7071 * Build option no-ssl3 is incomplete.
7073 When OpenSSL is configured with "no-ssl3" as a build option, servers
7074 could accept and complete a SSL 3.0 handshake, and clients could be
7075 configured to send them.
7078 *Akamai and the OpenSSL team*
7080 * Add support for TLS_FALLBACK_SCSV.
7081 Client applications doing fallback retries should call
7082 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7085 *Adam Langley, Bodo Moeller*
7087 * Add additional DigestInfo checks.
7089 Re-encode DigestInto in DER and check against the original when
7090 verifying RSA signature: this will reject any improperly encoded
7091 DigestInfo structures.
7093 Note: this is a precautionary measure and no attacks are currently known.
7097 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7099 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7100 SRP code can be overrun an internal buffer. Add sanity check that
7101 g, A, B < N to SRP code.
7103 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7104 Group for discovering this issue.
7109 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7110 TLS 1.0 instead of higher protocol versions when the ClientHello message
7111 is badly fragmented. This allows a man-in-the-middle attacker to force a
7112 downgrade to TLS 1.0 even if both the server and the client support a
7113 higher protocol version, by modifying the client's TLS records.
7115 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7116 researching this issue.
7121 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7122 to a denial of service attack. A malicious server can crash the client
7123 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7124 ciphersuite and sending carefully crafted handshake messages.
7126 Thanks to Felix Gröbert (Google) for discovering and researching this
7132 * By sending carefully crafted DTLS packets an attacker could cause openssl
7133 to leak memory. This can be exploited through a Denial of Service attack.
7134 Thanks to Adam Langley for discovering and researching this issue.
7139 * An attacker can force openssl to consume large amounts of memory whilst
7140 processing DTLS handshake messages. This can be exploited through a
7141 Denial of Service attack.
7142 Thanks to Adam Langley for discovering and researching this issue.
7147 * An attacker can force an error condition which causes openssl to crash
7148 whilst processing DTLS packets due to memory being freed twice. This
7149 can be exploited through a Denial of Service attack.
7150 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7156 * If a multithreaded client connects to a malicious server using a resumed
7157 session and the server sends an ec point format extension it could write
7158 up to 255 bytes to freed memory.
7160 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7166 * A malicious server can crash an OpenSSL client with a null pointer
7167 dereference (read) by specifying an SRP ciphersuite even though it was not
7168 properly negotiated with the client. This can be exploited through a
7169 Denial of Service attack.
7171 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7172 discovering and researching this issue.
7177 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7178 X509_name_oneline, X509_name_print_ex et al. to leak some information
7179 from the stack. Applications may be affected if they echo pretty printing
7180 output to the attacker.
7182 Thanks to Ivan Fratric (Google) for discovering this issue.
7185 *Emilia Käsper, and Steve Henson*
7187 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7188 for corner cases. (Certain input points at infinity could lead to
7189 bogus results, with non-infinity inputs mapped to infinity too.)
7193 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7195 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7196 handshake can force the use of weak keying material in OpenSSL
7197 SSL/TLS clients and servers.
7199 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7200 researching this issue. ([CVE-2014-0224])
7202 *KIKUCHI Masashi, Steve Henson*
7204 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7205 OpenSSL DTLS client the code can be made to recurse eventually crashing
7208 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7211 *Imre Rad, Steve Henson*
7213 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7214 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7215 client or server. This is potentially exploitable to run arbitrary
7216 code on a vulnerable client or server.
7218 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7220 *Jüri Aedla, Steve Henson*
7222 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7223 are subject to a denial of service attack.
7225 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7226 this issue. ([CVE-2014-3470])
7228 *Felix Gröbert, Ivan Fratric, Steve Henson*
7230 * Harmonize version and its documentation. -f flag is used to display
7233 *mancha <mancha1@zoho.com>*
7235 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7236 in i2d_ECPrivateKey.
7238 *mancha <mancha1@zoho.com>*
7240 * Fix some double frees. These are not thought to be exploitable.
7242 *mancha <mancha1@zoho.com>*
7244 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7246 * A missing bounds check in the handling of the TLS heartbeat extension
7247 can be used to reveal up to 64k of memory to a connected client or
7250 Thanks for Neel Mehta of Google Security for discovering this bug and to
7251 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7252 preparing the fix ([CVE-2014-0160])
7254 *Adam Langley, Bodo Moeller*
7256 * Fix for the attack described in the paper "Recovering OpenSSL
7257 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7258 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7259 <http://eprint.iacr.org/2014/140>
7261 Thanks to Yuval Yarom and Naomi Benger for discovering this
7262 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7264 *Yuval Yarom and Naomi Benger*
7266 * TLS pad extension: draft-agl-tls-padding-03
7268 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7269 TLS client Hello record length value would otherwise be > 255 and
7270 less that 512 pad with a dummy extension containing zeroes so it
7271 is at least 512 bytes long.
7273 *Adam Langley, Steve Henson*
7275 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7277 * Fix for TLS record tampering bug. A carefully crafted invalid
7278 handshake could crash OpenSSL with a NULL pointer exception.
7279 Thanks to Anton Johansson for reporting this issues.
7282 * Keep original DTLS digest and encryption contexts in retransmission
7283 structures so we can use the previous session parameters if they need
7284 to be resent. ([CVE-2013-6450])
7288 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7289 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7290 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7291 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7292 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7293 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7295 *Rob Stradling, Adam Langley*
7297 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7299 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7300 supporting platforms or when small records were transferred.
7302 *Andy Polyakov, Steve Henson*
7304 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7306 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7308 This addresses the flaw in CBC record processing discovered by
7309 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7310 at: <http://www.isg.rhul.ac.uk/tls/>
7312 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7313 Security Group at Royal Holloway, University of London
7314 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7315 Emilia Käsper for the initial patch.
7318 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7320 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7321 ciphersuites which can be exploited in a denial of service attack.
7322 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7323 and detecting this bug and to Wolfgang Ettlinger
7324 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7329 * Return an error when checking OCSP signatures when key is NULL.
7330 This fixes a DoS attack. ([CVE-2013-0166])
7334 * Make openssl verify return errors.
7336 *Chris Palmer <palmer@google.com> and Ben Laurie*
7338 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7339 the right response is stapled. Also change SSL_get_certificate()
7340 so it returns the certificate actually sent.
7341 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7343 *Rob Stradling <rob.stradling@comodo.com>*
7345 * Fix possible deadlock when decoding public keys.
7349 * Don't use TLS 1.0 record version number in initial client hello
7354 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7356 * Sanity check record length before skipping explicit IV in TLS
7357 1.2, 1.1 and DTLS to fix DoS attack.
7359 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7360 fuzzing as a service testing platform.
7365 * Initialise tkeylen properly when encrypting CMS messages.
7366 Thanks to Solar Designer of Openwall for reporting this issue.
7370 * In FIPS mode don't try to use composite ciphers as they are not
7375 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7377 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7378 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7379 mean any application compiled against OpenSSL 1.0.0 headers setting
7380 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7381 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7382 0x10000000L Any application which was previously compiled against
7383 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7384 will need to be recompiled as a result. Letting be results in
7385 inability to disable specifically TLS 1.1 and in client context,
7386 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7390 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7391 disable just protocol X, but all protocols above X *if* there are
7392 protocols *below* X still enabled. In more practical terms it means
7393 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7394 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7395 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7400 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7402 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7403 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7404 in CRYPTO_realloc_clean.
7406 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7407 issue and to Adam Langley <agl@chromium.org> for fixing it.
7410 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7412 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7416 * Workarounds for some broken servers that "hang" if a client hello
7417 record length exceeds 255 bytes.
7419 1. Do not use record version number > TLS 1.0 in initial client
7420 hello: some (but not all) hanging servers will now work.
7421 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7422 the number of ciphers sent in the client hello. This should be
7423 set to an even number, such as 50, for example by passing:
7424 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7425 Most broken servers should now work.
7426 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7427 TLS 1.2 client support entirely.
7431 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7435 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7437 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7438 STRING form instead of a DigestInfo.
7442 * The format used for MDC2 RSA signatures is inconsistent between EVP
7443 and the RSA_sign/RSA_verify functions. This was made more apparent when
7444 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7445 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7446 the correct format in RSA_verify so both forms transparently work.
7450 * Some servers which support TLS 1.0 can choke if we initially indicate
7451 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7452 encrypted premaster secret. As a workaround use the maximum permitted
7453 client version in client hello, this should keep such servers happy
7454 and still work with previous versions of OpenSSL.
7458 * Add support for TLS/DTLS heartbeats.
7460 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7462 * Add support for SCTP.
7464 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7466 * Improved PRNG seeding for VOS.
7468 *Paul Green <Paul.Green@stratus.com>*
7470 * Extensive assembler packs updates, most notably:
7472 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7473 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7474 - x86_64: bit-sliced AES implementation;
7475 - ARM: NEON support, contemporary platforms optimizations;
7476 - s390x: z196 support;
7477 - `*`: GHASH and GF(2^m) multiplication implementations;
7481 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7482 (removal of unnecessary code)
7484 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7486 * Add TLS key material exporter from RFC 5705.
7490 * Add DTLS-SRTP negotiation from RFC 5764.
7494 * Add Next Protocol Negotiation,
7495 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7496 disabled with a no-npn flag to config or Configure. Code donated
7499 *Adam Langley <agl@google.com> and Ben Laurie*
7501 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7502 NIST-P256, NIST-P521, with constant-time single point multiplication on
7503 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7504 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7505 Code made available under Apache License version 2.0.
7507 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7508 line to include this in your build of OpenSSL, and run "make depend" (or
7509 "make update"). This enables the following EC_METHODs:
7511 EC_GFp_nistp224_method()
7512 EC_GFp_nistp256_method()
7513 EC_GFp_nistp521_method()
7515 EC_GROUP_new_by_curve_name() will automatically use these (while
7516 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7519 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7521 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7522 all platforms. Move ssize_t definition from e_os.h to the public
7523 header file e_os2.h as it now appears in public header file cms.h
7527 * New -sigopt option to the ca, req and x509 utilities. Additional
7528 signature parameters can be passed using this option and in
7533 * Add RSA PSS signing function. This will generate and set the
7534 appropriate AlgorithmIdentifiers for PSS based on those in the
7535 corresponding EVP_MD_CTX structure. No application support yet.
7539 * Support for companion algorithm specific ASN1 signing routines.
7540 New function ASN1_item_sign_ctx() signs a pre-initialised
7541 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7542 the appropriate parameters.
7546 * Add new algorithm specific ASN1 verification initialisation function
7547 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7548 handling will be the same no matter what EVP_PKEY_METHOD is used.
7549 Add a PSS handler to support verification of PSS signatures: checked
7550 against a number of sample certificates.
7554 * Add signature printing for PSS. Add PSS OIDs.
7556 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7558 * Add algorithm specific signature printing. An individual ASN1 method
7559 can now print out signatures instead of the standard hex dump.
7561 More complex signatures (e.g. PSS) can print out more meaningful
7562 information. Include DSA version that prints out the signature
7567 * Password based recipient info support for CMS library: implementing
7572 * Split password based encryption into PBES2 and PBKDF2 functions. This
7573 neatly separates the code into cipher and PBE sections and is required
7574 for some algorithms that split PBES2 into separate pieces (such as
7575 password based CMS).
7579 * Session-handling fixes:
7580 - Fix handling of connections that are resuming with a session ID,
7581 but also support Session Tickets.
7582 - Fix a bug that suppressed issuing of a new ticket if the client
7583 presented a ticket with an expired session.
7584 - Try to set the ticket lifetime hint to something reasonable.
7585 - Make tickets shorter by excluding irrelevant information.
7586 - On the client side, don't ignore renewed tickets.
7588 *Adam Langley, Bodo Moeller (Google)*
7590 * Fix PSK session representation.
7594 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7596 This work was sponsored by Intel.
7600 * Add GCM support to TLS library. Some custom code is needed to split
7601 the IV between the fixed (from PRF) and explicit (from TLS record)
7602 portions. This adds all GCM ciphersuites supported by RFC5288 and
7603 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7604 add a special AESGCM string for GCM only.
7608 * Expand range of ctrls for AES GCM. Permit setting invocation
7609 field on decrypt and retrieval of invocation field only on encrypt.
7613 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7614 As required by RFC5289 these ciphersuites cannot be used if for
7615 versions of TLS earlier than 1.2.
7619 * For FIPS capable OpenSSL interpret a NULL default public key method
7620 as unset and return the appropriate default but do *not* set the default.
7621 This means we can return the appropriate method in applications that
7622 switch between FIPS and non-FIPS modes.
7626 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7627 ENGINE is used then we cannot handle that in the FIPS module so we
7628 keep original code iff non-FIPS operations are allowed.
7632 * Add -attime option to openssl utilities.
7634 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7636 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7640 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7641 FIPS EC methods unconditionally for now.
7645 * New build option no-ec2m to disable characteristic 2 code.
7649 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7650 all cases can be covered as some introduce binary incompatibilities.
7654 * Redirect RSA operations to FIPS module including keygen,
7655 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7659 * Add similar low-level API blocking to ciphers.
7663 * low-level digest APIs are not approved in FIPS mode: any attempt
7664 to use these will cause a fatal error. Applications that *really* want
7665 to use them can use the `private_*` version instead.
7669 * Redirect cipher operations to FIPS module for FIPS builds.
7673 * Redirect digest operations to FIPS module for FIPS builds.
7677 * Update build system to add "fips" flag which will link in fipscanister.o
7678 for static and shared library builds embedding a signature if needed.
7682 * Output TLS supported curves in preference order instead of numerical
7683 order. This is currently hardcoded for the highest order curves first.
7684 This should be configurable so applications can judge speed vs strength.
7688 * Add TLS v1.2 server support for client authentication.
7692 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7697 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7698 FIPS modules versions.
7702 * Add TLS v1.2 client side support for client authentication. Keep cache
7703 of handshake records longer as we don't know the hash algorithm to use
7704 until after the certificate request message is received.
7708 * Initial TLS v1.2 client support. Add a default signature algorithms
7709 extension including all the algorithms we support. Parse new signature
7710 format in client key exchange. Relax some ECC signing restrictions for
7711 TLS v1.2 as indicated in RFC5246.
7715 * Add server support for TLS v1.2 signature algorithms extension. Switch
7716 to new signature format when needed using client digest preference.
7717 All server ciphersuites should now work correctly in TLS v1.2. No client
7718 support yet and no support for client certificates.
7722 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7723 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7724 ciphersuites. At present only RSA key exchange ciphersuites work with
7725 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7726 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7727 and version checking.
7731 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7732 with this defined it will not be affected by any changes to ssl internal
7733 structures. Add several utility functions to allow openssl application
7734 to work with OPENSSL_NO_SSL_INTERN defined.
7738 * A long standing patch to add support for SRP from EdelWeb (Peter
7739 Sylvester and Christophe Renou) was integrated.
7740 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7741 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7744 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7748 * Permit abbreviated handshakes when renegotiating using the function
7749 SSL_renegotiate_abbreviated().
7751 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7753 * Add call to ENGINE_register_all_complete() to
7754 ENGINE_load_builtin_engines(), so some implementations get used
7755 automatically instead of needing explicit application support.
7759 * Add support for TLS key exporter as described in RFC5705.
7761 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7763 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7764 a few changes are required:
7766 Add SSL_OP_NO_TLSv1_1 flag.
7767 Add TLSv1_1 methods.
7768 Update version checking logic to handle version 1.1.
7769 Add explicit IV handling (ported from DTLS code).
7770 Add command line options to s_client/s_server.
7777 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7779 * X509_ATTRIBUTE memory leak
7781 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7782 memory. This structure is used by the PKCS#7 and CMS routines so any
7783 application which reads PKCS#7 or CMS data from untrusted sources is
7784 affected. SSL/TLS is not affected.
7786 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7792 * Race condition handling PSK identify hint
7794 If PSK identity hints are received by a multi-threaded client then
7795 the values are wrongly updated in the parent SSL_CTX structure. This can
7796 result in a race condition potentially leading to a double free of the
7802 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7804 * Malformed ECParameters causes infinite loop
7806 When processing an ECParameters structure OpenSSL enters an infinite loop
7807 if the curve specified is over a specially malformed binary polynomial
7810 This can be used to perform denial of service against any
7811 system which processes public keys, certificate requests or
7812 certificates. This includes TLS clients and TLS servers with
7813 client authentication enabled.
7815 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7820 * Exploitable out-of-bounds read in X509_cmp_time
7822 X509_cmp_time does not properly check the length of the ASN1_TIME
7823 string and can read a few bytes out of bounds. In addition,
7824 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7827 An attacker can use this to craft malformed certificates and CRLs of
7828 various sizes and potentially cause a segmentation fault, resulting in
7829 a DoS on applications that verify certificates or CRLs. TLS clients
7830 that verify CRLs are affected. TLS clients and servers with client
7831 authentication enabled may be affected if they use custom verification
7834 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7835 independently by Hanno Böck.
7840 * PKCS7 crash with missing EnvelopedContent
7842 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7843 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7844 with missing content and trigger a NULL pointer dereference on parsing.
7846 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7847 structures from untrusted sources are affected. OpenSSL clients and
7848 servers are not affected.
7850 This issue was reported to OpenSSL by Michal Zalewski (Google).
7855 * CMS verify infinite loop with unknown hash function
7857 When verifying a signedData message the CMS code can enter an infinite loop
7858 if presented with an unknown hash function OID. This can be used to perform
7859 denial of service against any system which verifies signedData messages using
7861 This issue was reported to OpenSSL by Johannes Bauer.
7866 * Race condition handling NewSessionTicket
7868 If a NewSessionTicket is received by a multi-threaded client when attempting to
7869 reuse a previous ticket then a race condition can occur potentially leading to
7870 a double free of the ticket data.
7875 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7877 * Segmentation fault in ASN1_TYPE_cmp fix
7879 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7880 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7881 certificate signature algorithm consistency this can be used to crash any
7882 certificate verification operation and exploited in a DoS attack. Any
7883 application which performs certificate verification is vulnerable including
7884 OpenSSL clients and servers which enable client authentication.
7889 * ASN.1 structure reuse memory corruption fix
7891 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7892 memory corruption via an invalid write. Such reuse is and has been
7893 strongly discouraged and is believed to be rare.
7895 Applications that parse structures containing CHOICE or ANY DEFINED BY
7896 components may be affected. Certificate parsing (d2i_X509 and related
7897 functions) are however not affected. OpenSSL clients and servers are
7903 * PKCS7 NULL pointer dereferences fix
7905 The PKCS#7 parsing code does not handle missing outer ContentInfo
7906 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7907 missing content and trigger a NULL pointer dereference on parsing.
7909 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7910 otherwise parse PKCS#7 structures from untrusted sources are
7911 affected. OpenSSL clients and servers are not affected.
7913 This issue was reported to OpenSSL by Michal Zalewski (Google).
7918 * DoS via reachable assert in SSLv2 servers fix
7920 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7921 servers that both support SSLv2 and enable export cipher suites by sending
7922 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7924 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7925 (OpenSSL development team).
7930 * Use After Free following d2i_ECPrivatekey error fix
7932 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7933 could cause a use after free condition. This, in turn, could cause a double
7934 free in several private key parsing functions (such as d2i_PrivateKey
7935 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7936 for applications that receive EC private keys from untrusted
7937 sources. This scenario is considered rare.
7939 This issue was discovered by the BoringSSL project and fixed in their
7945 * X509_to_X509_REQ NULL pointer deref fix
7947 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7948 the certificate key is invalid. This function is rarely used in practice.
7950 This issue was discovered by Brian Carpenter.
7955 * Removed the export ciphers from the DEFAULT ciphers
7959 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7961 * Build fixes for the Windows and OpenVMS platforms
7963 *Matt Caswell and Richard Levitte*
7965 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7967 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7968 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7969 dereference. This could lead to a Denial Of Service attack. Thanks to
7970 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7975 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7976 dtls1_buffer_record function under certain conditions. In particular this
7977 could occur if an attacker sent repeated DTLS records with the same
7978 sequence number but for the next epoch. The memory leak could be exploited
7979 by an attacker in a Denial of Service attack through memory exhaustion.
7980 Thanks to Chris Mueller for reporting this issue.
7985 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7986 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7987 method would be set to NULL which could later result in a NULL pointer
7988 dereference. Thanks to Frank Schmirler for reporting this issue.
7993 * Abort handshake if server key exchange message is omitted for ephemeral
7996 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7997 reporting this issue.
8002 * Remove non-export ephemeral RSA code on client and server. This code
8003 violated the TLS standard by allowing the use of temporary RSA keys in
8004 non-export ciphersuites and could be used by a server to effectively
8005 downgrade the RSA key length used to a value smaller than the server
8006 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8007 INRIA or reporting this issue.
8012 * Fixed issue where DH client certificates are accepted without verification.
8013 An OpenSSL server will accept a DH certificate for client authentication
8014 without the certificate verify message. This effectively allows a client to
8015 authenticate without the use of a private key. This only affects servers
8016 which trust a client certificate authority which issues certificates
8017 containing DH keys: these are extremely rare and hardly ever encountered.
8018 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8024 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8025 results on some platforms, including x86_64. This bug occurs at random
8026 with a very low probability, and is not known to be exploitable in any
8027 way, though its exact impact is difficult to determine. Thanks to Pieter
8028 Wuille (Blockstream) who reported this issue and also suggested an initial
8029 fix. Further analysis was conducted by the OpenSSL development team and
8030 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8031 the OpenSSL core team.
8036 * Fix various certificate fingerprint issues.
8038 By using non-DER or invalid encodings outside the signed portion of a
8039 certificate the fingerprint can be changed without breaking the signature.
8040 Although no details of the signed portion of the certificate can be changed
8041 this can cause problems with some applications: e.g. those using the
8042 certificate fingerprint for blacklists.
8044 1. Reject signatures with non zero unused bits.
8046 If the BIT STRING containing the signature has non zero unused bits reject
8047 the signature. All current signature algorithms require zero unused bits.
8049 2. Check certificate algorithm consistency.
8051 Check the AlgorithmIdentifier inside TBS matches the one in the
8052 certificate signature. NB: this will result in signature failure
8053 errors for some broken certificates.
8055 Thanks to Konrad Kraszewski from Google for reporting this issue.
8057 3. Check DSA/ECDSA signatures use DER.
8059 Reencode DSA/ECDSA signatures and compare with the original received
8060 signature. Return an error if there is a mismatch.
8062 This will reject various cases including garbage after signature
8063 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8064 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8065 (negative or with leading zeroes).
8067 Further analysis was conducted and fixes were developed by Stephen Henson
8068 of the OpenSSL core team.
8074 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8076 * Session Ticket Memory Leak.
8078 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8079 integrity of that ticket is first verified. In the event of a session
8080 ticket integrity check failing, OpenSSL will fail to free memory
8081 causing a memory leak. By sending a large number of invalid session
8082 tickets an attacker could exploit this issue in a Denial Of Service
8088 * Build option no-ssl3 is incomplete.
8090 When OpenSSL is configured with "no-ssl3" as a build option, servers
8091 could accept and complete a SSL 3.0 handshake, and clients could be
8092 configured to send them.
8095 *Akamai and the OpenSSL team*
8097 * Add support for TLS_FALLBACK_SCSV.
8098 Client applications doing fallback retries should call
8099 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8102 *Adam Langley, Bodo Moeller*
8104 * Add additional DigestInfo checks.
8106 Reencode DigestInto in DER and check against the original when
8107 verifying RSA signature: this will reject any improperly encoded
8108 DigestInfo structures.
8110 Note: this is a precautionary measure and no attacks are currently known.
8114 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8116 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8117 to a denial of service attack. A malicious server can crash the client
8118 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8119 ciphersuite and sending carefully crafted handshake messages.
8121 Thanks to Felix Gröbert (Google) for discovering and researching this
8127 * By sending carefully crafted DTLS packets an attacker could cause openssl
8128 to leak memory. This can be exploited through a Denial of Service attack.
8129 Thanks to Adam Langley for discovering and researching this issue.
8134 * An attacker can force openssl to consume large amounts of memory whilst
8135 processing DTLS handshake messages. This can be exploited through a
8136 Denial of Service attack.
8137 Thanks to Adam Langley for discovering and researching this issue.
8142 * An attacker can force an error condition which causes openssl to crash
8143 whilst processing DTLS packets due to memory being freed twice. This
8144 can be exploited through a Denial of Service attack.
8145 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8151 * If a multithreaded client connects to a malicious server using a resumed
8152 session and the server sends an ec point format extension it could write
8153 up to 255 bytes to freed memory.
8155 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8161 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8162 X509_name_oneline, X509_name_print_ex et al. to leak some information
8163 from the stack. Applications may be affected if they echo pretty printing
8164 output to the attacker.
8166 Thanks to Ivan Fratric (Google) for discovering this issue.
8169 *Emilia Käsper, and Steve Henson*
8171 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8172 for corner cases. (Certain input points at infinity could lead to
8173 bogus results, with non-infinity inputs mapped to infinity too.)
8177 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8179 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8180 handshake can force the use of weak keying material in OpenSSL
8181 SSL/TLS clients and servers.
8183 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8184 researching this issue. ([CVE-2014-0224])
8186 *KIKUCHI Masashi, Steve Henson*
8188 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8189 OpenSSL DTLS client the code can be made to recurse eventually crashing
8192 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8195 *Imre Rad, Steve Henson*
8197 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8198 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8199 client or server. This is potentially exploitable to run arbitrary
8200 code on a vulnerable client or server.
8202 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8204 *Jüri Aedla, Steve Henson*
8206 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8207 are subject to a denial of service attack.
8209 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8210 this issue. ([CVE-2014-3470])
8212 *Felix Gröbert, Ivan Fratric, Steve Henson*
8214 * Harmonize version and its documentation. -f flag is used to display
8217 *mancha <mancha1@zoho.com>*
8219 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8220 in i2d_ECPrivateKey.
8222 *mancha <mancha1@zoho.com>*
8224 * Fix some double frees. These are not thought to be exploitable.
8226 *mancha <mancha1@zoho.com>*
8228 * Fix for the attack described in the paper "Recovering OpenSSL
8229 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8230 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8231 <http://eprint.iacr.org/2014/140>
8233 Thanks to Yuval Yarom and Naomi Benger for discovering this
8234 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8236 *Yuval Yarom and Naomi Benger*
8238 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8240 * Keep original DTLS digest and encryption contexts in retransmission
8241 structures so we can use the previous session parameters if they need
8242 to be resent. ([CVE-2013-6450])
8246 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8247 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8248 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8249 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8250 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8251 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8253 *Rob Stradling, Adam Langley*
8255 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8257 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8259 This addresses the flaw in CBC record processing discovered by
8260 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8261 at: <http://www.isg.rhul.ac.uk/tls/>
8263 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8264 Security Group at Royal Holloway, University of London
8265 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8266 Emilia Käsper for the initial patch.
8269 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8271 * Return an error when checking OCSP signatures when key is NULL.
8272 This fixes a DoS attack. ([CVE-2013-0166])
8276 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8277 the right response is stapled. Also change SSL_get_certificate()
8278 so it returns the certificate actually sent.
8279 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8280 (This is a backport)
8282 *Rob Stradling <rob.stradling@comodo.com>*
8284 * Fix possible deadlock when decoding public keys.
8288 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8290 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8293 * Sanity check record length before skipping explicit IV in DTLS
8296 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8297 fuzzing as a service testing platform.
8302 * Initialise tkeylen properly when encrypting CMS messages.
8303 Thanks to Solar Designer of Openwall for reporting this issue.
8307 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8309 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8310 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8311 in CRYPTO_realloc_clean.
8313 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8314 issue and to Adam Langley <agl@chromium.org> for fixing it.
8317 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8319 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8321 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8322 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8323 content decryption and always return the same error. Note: this attack
8324 needs on average 2^20 messages so it only affects automated senders. The
8325 old behaviour can be re-enabled in the CMS code by setting the
8326 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8327 an MMA defence is not necessary.
8328 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8329 this issue. ([CVE-2012-0884])
8333 * Fix CVE-2011-4619: make sure we really are receiving a
8334 client hello before rejecting multiple SGC restarts. Thanks to
8335 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8339 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8341 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8342 Thanks to Antonio Martin, Enterprise Secure Access Research and
8343 Development, Cisco Systems, Inc. for discovering this bug and
8344 preparing a fix. ([CVE-2012-0050])
8348 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8350 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8351 of the Vaudenay padding oracle attack on CBC mode encryption
8352 which enables an efficient plaintext recovery attack against
8353 the OpenSSL implementation of DTLS. Their attack exploits timing
8354 differences arising during decryption processing. A research
8355 paper describing this attack can be found at:
8356 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8357 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8358 Security Group at Royal Holloway, University of London
8359 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8360 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8361 for preparing the fix. ([CVE-2011-4108])
8363 *Robin Seggelmann, Michael Tuexen*
8365 * Clear bytes used for block padding of SSL 3.0 records.
8368 *Adam Langley (Google)*
8370 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8371 Kadianakis <desnacked@gmail.com> for discovering this issue and
8372 Adam Langley for preparing the fix. ([CVE-2011-4619])
8374 *Adam Langley (Google)*
8376 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8378 *Andrey Kulikov <amdeich@gmail.com>*
8380 * Prevent malformed RFC3779 data triggering an assertion failure.
8381 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8382 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8384 *Rob Austein <sra@hactrn.net>*
8386 * Improved PRNG seeding for VOS.
8388 *Paul Green <Paul.Green@stratus.com>*
8390 * Fix ssl_ciph.c set-up race.
8392 *Adam Langley (Google)*
8394 * Fix spurious failures in ecdsatest.c.
8396 *Emilia Käsper (Google)*
8398 * Fix the BIO_f_buffer() implementation (which was mixing different
8399 interpretations of the `..._len` fields).
8401 *Adam Langley (Google)*
8403 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8404 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8405 threads won't reuse the same blinding coefficients.
8407 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8408 lock to call BN_BLINDING_invert_ex, and avoids one use of
8409 BN_BLINDING_update for each BN_BLINDING structure (previously,
8410 the last update always remained unused).
8412 *Emilia Käsper (Google)*
8414 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8416 *Bob Buckholz (Google)*
8418 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8420 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8421 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8423 *Kaspar Brand <ossl@velox.ch>*
8425 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8426 for multi-threaded use of ECDH. ([CVE-2011-3210])
8428 *Adam Langley (Google)*
8430 * Fix x509_name_ex_d2i memory leak on bad inputs.
8434 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8435 signature public key algorithm by using OID xref utilities instead.
8436 Before this you could only use some ECC ciphersuites with SHA1 only.
8440 * Add protection against ECDSA timing attacks as mentioned in the paper
8441 by Billy Bob Brumley and Nicola Tuveri, see:
8442 <http://eprint.iacr.org/2011/232.pdf>
8444 *Billy Bob Brumley and Nicola Tuveri*
8446 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8448 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8450 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8452 * Fix bug in string printing code: if *any* escaping is enabled we must
8453 escape the escape character (backslash) or the resulting string is
8458 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8460 * Disable code workaround for ancient and obsolete Netscape browsers
8461 and servers: an attacker can use it in a ciphersuite downgrade attack.
8462 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8466 * Fixed J-PAKE implementation error, originally discovered by
8467 Sebastien Martini, further info and confirmation from Stefan
8468 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8472 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8474 * Fix extension code to avoid race conditions which can result in a buffer
8475 overrun vulnerability: resumed sessions must not be modified as they can
8476 be shared by multiple threads. CVE-2010-3864
8480 * Fix WIN32 build system to correctly link an ENGINE directory into
8485 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8487 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8490 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8492 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8494 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8495 context. The operation can be customised via the ctrl mechanism in
8496 case ENGINEs want to include additional functionality.
8500 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8504 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8505 output hashes compatible with older versions of OpenSSL.
8507 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8509 * Fix compression algorithm handling: if resuming a session use the
8510 compression algorithm of the resumed session instead of determining
8511 it from client hello again. Don't allow server to change algorithm.
8515 * Add load_crls() function to commands tidying load_certs() too. Add option
8516 to verify utility to allow additional CRLs to be included.
8520 * Update OCSP request code to permit adding custom headers to the request:
8521 some responders need this.
8525 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8528 *Julia Lawall <julia@diku.dk>*
8530 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8531 needlessly dereferenced structures, used obsolete functions and
8532 didn't handle all updated verify codes correctly.
8536 * Disable MD2 in the default configuration.
8540 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8541 indicate the initial BIO being pushed or popped. This makes it possible
8542 to determine whether the BIO is the one explicitly called or as a result
8543 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8544 it handles reference counts correctly and doesn't zero out the I/O bio
8545 when it is not being explicitly popped. WARNING: applications which
8546 included workarounds for the old buggy behaviour will need to be modified
8547 or they could free up already freed BIOs.
8551 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8552 renaming to all platforms (within the 0.9.8 branch, this was
8553 done conditionally on Netware platforms to avoid a name clash).
8555 *Guenter <lists@gknw.net>*
8557 * Add ECDHE and PSK support to DTLS.
8559 *Michael Tuexen <tuexen@fh-muenster.de>*
8561 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8566 * Add "missing" function EVP_MD_flags() (without this the only way to
8567 retrieve a digest flags is by accessing the structure directly. Update
8568 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8569 or cipher is registered as in the "from" argument. Print out all
8570 registered digests in the dgst usage message instead of manually
8571 attempting to work them out.
8575 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8576 this allows the use of compression and extensions. Change default cipher
8577 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8578 by default unless an application cipher string requests it.
8582 * Alter match criteria in PKCS12_parse(). It used to try to use local
8583 key ids to find matching certificates and keys but some PKCS#12 files
8584 don't follow the (somewhat unwritten) rules and this strategy fails.
8585 Now just gather all certificates together and the first private key
8586 then look for the first certificate that matches the key.
8590 * Support use of registered digest and cipher names for dgst and cipher
8591 commands instead of having to add each one as a special case. So now
8598 openssl dgst -sha256 foo
8600 and this works for ENGINE based algorithms too.
8604 * Update Gost ENGINE to support parameter files.
8606 *Victor B. Wagner <vitus@cryptocom.ru>*
8608 * Support GeneralizedTime in ca utility.
8610 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8612 * Enhance the hash format used for certificate directory links. The new
8613 form uses the canonical encoding (meaning equivalent names will work
8614 even if they aren't identical) and uses SHA1 instead of MD5. This form
8615 is incompatible with the older format and as a result c_rehash should
8616 be used to rebuild symbolic links.
8620 * Make PKCS#8 the default write format for private keys, replacing the
8621 traditional format. This form is standardised, more secure and doesn't
8622 include an implicit MD5 dependency.
8626 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8627 committed to OpenSSL should pass this lot as a minimum.
8631 * Add session ticket override functionality for use by EAP-FAST.
8633 *Jouni Malinen <j@w1.fi>*
8635 * Modify HMAC functions to return a value. Since these can be implemented
8636 in an ENGINE errors can occur.
8640 * Type-checked OBJ_bsearch_ex.
8644 * Type-checked OBJ_bsearch. Also some constification necessitated
8645 by type-checking. Still to come: TXT_DB, bsearch(?),
8646 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8651 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8652 seconds to a tm structure directly, instead of going through OS
8653 specific date routines. This avoids any issues with OS routines such
8654 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8655 and X509_time_adj_ex() to cover the extended range. The existing
8656 X509_time_adj() is still usable and will no longer have any date issues.
8660 * Delta CRL support. New use deltas option which will attempt to locate
8661 and search any appropriate delta CRLs available.
8663 This work was sponsored by Google.
8667 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8668 code and add additional score elements. Validate alternate CRL paths
8669 as part of the CRL checking and indicate a new error "CRL path validation
8670 error" in this case. Applications wanting additional details can use
8671 the verify callback and check the new "parent" field. If this is not
8672 NULL CRL path validation is taking place. Existing applications won't
8673 see this because it requires extended CRL support which is off by
8676 This work was sponsored by Google.
8680 * Support for freshest CRL extension.
8682 This work was sponsored by Google.
8686 * Initial indirect CRL support. Currently only supported in the CRLs
8687 passed directly and not via lookup. Process certificate issuer
8688 CRL entry extension and lookup CRL entries by bother issuer name
8689 and serial number. Check and process CRL issuer entry in IDP extension.
8691 This work was sponsored by Google.
8695 * Add support for distinct certificate and CRL paths. The CRL issuer
8696 certificate is validated separately in this case. Only enabled if
8697 an extended CRL support flag is set: this flag will enable additional
8698 CRL functionality in future.
8700 This work was sponsored by Google.
8704 * Add support for policy mappings extension.
8706 This work was sponsored by Google.
8710 * Fixes to pathlength constraint, self issued certificate handling,
8711 policy processing to align with RFC3280 and PKITS tests.
8713 This work was sponsored by Google.
8717 * Support for name constraints certificate extension. DN, email, DNS
8718 and URI types are currently supported.
8720 This work was sponsored by Google.
8724 * To cater for systems that provide a pointer-based thread ID rather
8725 than numeric, deprecate the current numeric thread ID mechanism and
8726 replace it with a structure and associated callback type. This
8727 mechanism allows a numeric "hash" to be extracted from a thread ID in
8728 either case, and on platforms where pointers are larger than 'long',
8729 mixing is done to help ensure the numeric 'hash' is usable even if it
8730 can't be guaranteed unique. The default mechanism is to use "&errno"
8731 as a pointer-based thread ID to distinguish between threads.
8733 Applications that want to provide their own thread IDs should now use
8734 CRYPTO_THREADID_set_callback() to register a callback that will call
8735 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8737 Note that ERR_remove_state() is now deprecated, because it is tied
8738 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8739 to free the current thread's error state should be replaced by
8740 ERR_remove_thread_state(NULL).
8742 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8743 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8744 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8745 application was previously providing a numeric thread callback that
8746 was inappropriate for distinguishing threads, then uniqueness might
8747 have been obtained with &errno that happened immediately in the
8748 intermediate development versions of OpenSSL; this is no longer the
8749 case, the numeric thread callback will now override the automatic use
8752 *Geoff Thorpe, with help from Bodo Moeller*
8754 * Initial support for different CRL issuing certificates. This covers a
8755 simple case where the self issued certificates in the chain exist and
8756 the real CRL issuer is higher in the existing chain.
8758 This work was sponsored by Google.
8762 * Removed effectively defunct crypto/store from the build.
8766 * Revamp of STACK to provide stronger type-checking. Still to come:
8767 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8768 ASN1_STRING, CONF_VALUE.
8772 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8773 RAM on SSL connections. This option can save about 34k per idle SSL.
8777 * Revamp of LHASH to provide stronger type-checking. Still to come:
8778 STACK, TXT_DB, bsearch, qsort.
8782 * Initial support for Cryptographic Message Syntax (aka CMS) based
8783 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8784 support for data, signedData, compressedData, digestedData and
8785 encryptedData, envelopedData types included. Scripts to check against
8786 RFC4134 examples draft and interop and consistency checks of many
8787 content types and variants.
8791 * Add options to enc utility to support use of zlib compression BIO.
8795 * Extend mk1mf to support importing of options and assembly language
8796 files from Configure script, currently only included in VC-WIN32.
8797 The assembly language rules can now optionally generate the source
8798 files from the associated perl scripts.
8802 * Implement remaining functionality needed to support GOST ciphersuites.
8803 Interop testing has been performed using CryptoPro implementations.
8805 *Victor B. Wagner <vitus@cryptocom.ru>*
8807 * s390x assembler pack.
8811 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8816 * Implement Opaque PRF Input TLS extension as specified in
8817 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8818 official specification yet and no extension type assignment by
8819 IANA exists, this extension (for now) will have to be explicitly
8820 enabled when building OpenSSL by providing the extension number
8821 to use. For example, specify an option
8823 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8825 to the "config" or "Configure" script to enable the extension,
8826 assuming extension number 0x9527 (which is a completely arbitrary
8827 and unofficial assignment based on the MD5 hash of the Internet
8828 Draft). Note that by doing so, you potentially lose
8829 interoperability with other TLS implementations since these might
8830 be using the same extension number for other purposes.
8832 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8833 opaque PRF input value to use in the handshake. This will create
8834 an internal copy of the length-'len' string at 'src', and will
8835 return non-zero for success.
8837 To get more control and flexibility, provide a callback function
8840 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8841 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8845 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8848 Callback function 'cb' will be called in handshakes, and is
8849 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8850 Argument 'arg' is for application purposes (the value as given to
8851 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8852 be provided to the callback function). The callback function
8853 has to return non-zero to report success: usually 1 to use opaque
8854 PRF input just if possible, or 2 to enforce use of the opaque PRF
8855 input. In the latter case, the library will abort the handshake
8856 if opaque PRF input is not successfully negotiated.
8858 Arguments 'peerinput' and 'len' given to the callback function
8859 will always be NULL and 0 in the case of a client. A server will
8860 see the client's opaque PRF input through these variables if
8861 available (NULL and 0 otherwise). Note that if the server
8862 provides an opaque PRF input, the length must be the same as the
8863 length of the client's opaque PRF input.
8865 Note that the callback function will only be called when creating
8866 a new session (session resumption can resume whatever was
8867 previously negotiated), and will not be called in SSL 2.0
8868 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8869 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8870 for applications that need to enforce opaque PRF input.
8874 * Update ssl code to support digests other than SHA1+MD5 for handshake
8877 *Victor B. Wagner <vitus@cryptocom.ru>*
8879 * Add RFC4507 support to OpenSSL. This includes the corrections in
8880 RFC4507bis. The encrypted ticket format is an encrypted encoded
8881 SSL_SESSION structure, that way new session features are automatically
8884 If a client application caches session in an SSL_SESSION structure
8885 support is transparent because tickets are now stored in the encoded
8888 The SSL_CTX structure automatically generates keys for ticket
8889 protection in servers so again support should be possible
8890 with no application modification.
8892 If a client or server wishes to disable RFC4507 support then the option
8893 SSL_OP_NO_TICKET can be set.
8895 Add a TLS extension debugging callback to allow the contents of any client
8896 or server extensions to be examined.
8898 This work was sponsored by Google.
8902 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8903 OpenSSL should now compile cleanly on gcc 4.2
8905 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8907 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8908 support including streaming MAC support: this is required for GOST
8909 ciphersuite support.
8911 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8913 * Add option -stream to use PKCS#7 streaming in smime utility. New
8914 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8915 to output in BER and PEM format.
8919 * Experimental support for use of HMAC via EVP_PKEY interface. This
8920 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8921 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8922 ENGINE support for HMAC keys which are unextractable. New -mac and
8923 -macopt options to dgst utility.
8927 * New option -sigopt to dgst utility. Update dgst to use
8928 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8929 alternative signing parameters such as X9.31 or PSS in the dgst
8934 * Change ssl_cipher_apply_rule(), the internal function that does
8935 the work each time a ciphersuite string requests enabling
8936 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8937 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8938 the order of disabled ciphersuites such that those ciphersuites
8939 that most recently went from enabled to disabled not only stay
8940 in order with respect to each other, but also have higher priority
8941 than other disabled ciphersuites the next time ciphersuites are
8944 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8945 the same ciphersuites as with "HIGH" alone, but in a specific
8946 order where the PSK ciphersuites come first (since they are the
8947 most recently disabled ciphersuites when "HIGH" is parsed).
8949 Also, change ssl_create_cipher_list() (using this new
8950 functionality) such that between otherwise identical
8951 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8956 * Change ssl_create_cipher_list() so that it automatically
8957 arranges the ciphersuites in reasonable order before starting
8958 to process the rule string. Thus, the definition for "DEFAULT"
8959 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8960 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8961 This makes it much easier to arrive at a reasonable default order
8962 in applications for which anonymous ciphers are OK (meaning
8963 that you can't actually use DEFAULT).
8965 *Bodo Moeller; suggested by Victor Duchovni*
8967 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8968 processing) into multiple integers instead of setting
8969 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8970 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8971 (These masks as well as the individual bit definitions are hidden
8972 away into the non-exported interface ssl/ssl_locl.h, so this
8973 change to the definition of the SSL_CIPHER structure shouldn't
8974 affect applications.) This give us more bits for each of these
8975 categories, so there is no longer a need to coagulate AES128 and
8976 AES256 into a single algorithm bit, and to coagulate Camellia128
8977 and Camellia256 into a single algorithm bit, which has led to all
8980 Thus, among other things, the kludge introduced in 0.9.7m and
8981 0.9.8e for masking out AES256 independently of AES128 or masking
8982 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8984 With the change, we also introduce new ciphersuite aliases that
8985 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8990 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8991 Use the leftmost N bytes of the signature input if the input is
8992 larger than the prime q (with N being the size in bytes of q).
8996 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8997 it yet and it is largely untested.
9001 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9005 * Initial incomplete changes to avoid need for function casts in OpenSSL
9006 some compilers (gcc 4.2 and later) reject their use. Safestack is
9007 reimplemented. Update ASN1 to avoid use of legacy functions.
9011 * Win32/64 targets are linked with Winsock2.
9015 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9016 to external functions. This can be used to increase CRL handling
9017 efficiency especially when CRLs are very large by (for example) storing
9018 the CRL revoked certificates in a database.
9022 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9023 new CRLs added to a directory can be used. New command line option
9024 -verify_return_error to s_client and s_server. This causes real errors
9025 to be returned by the verify callback instead of carrying on no matter
9026 what. This reflects the way a "real world" verify callback would behave.
9030 * GOST engine, supporting several GOST algorithms and public key formats.
9031 Kindly donated by Cryptocom.
9035 * Partial support for Issuing Distribution Point CRL extension. CRLs
9036 partitioned by DP are handled but no indirect CRL or reason partitioning
9037 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9038 selected via a scoring technique which handles IDP and AKID in CRLs.
9042 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9043 will ultimately be used for all verify operations: this will remove the
9044 X509_STORE dependency on certificate verification and allow alternative
9045 lookup methods. X509_STORE based implementations of these two callbacks.
9049 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9050 Modify get_crl() to find a valid (unexpired) CRL if possible.
9054 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9055 this would be called X509_CRL_cmp() but that name is already used by
9056 a function that just compares CRL issuer names. Cache several CRL
9057 extensions in X509_CRL structure and cache CRLDP in X509.
9061 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9062 this maps equivalent X509_NAME structures into a consistent structure.
9063 Name comparison can then be performed rapidly using memcmp().
9067 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9072 * Allow digests to supply their own micalg string for S/MIME type using
9073 the ctrl EVP_MD_CTRL_MICALG.
9077 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9078 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9079 ctrl. It can then customise the structure before and/or after signing
9084 * New function OBJ_add_sigid() to allow application defined signature OIDs
9085 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9086 to free up any added signature OIDs.
9090 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9091 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9092 digest and cipher tables. New options added to openssl utility:
9093 list-message-digest-algorithms and list-cipher-algorithms.
9097 * Change the array representation of binary polynomials: the list
9098 of degrees of non-zero coefficients is now terminated with -1.
9099 Previously it was terminated with 0, which was also part of the
9100 value; thus, the array representation was not applicable to
9101 polynomials where t^0 has coefficient zero. This change makes
9102 the array representation useful in a more general context.
9106 * Various modifications and fixes to SSL/TLS cipher string
9107 handling. For ECC, the code now distinguishes between fixed ECDH
9108 with RSA certificates on the one hand and with ECDSA certificates
9109 on the other hand, since these are separate ciphersuites. The
9110 unused code for Fortezza ciphersuites has been removed.
9112 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9113 (not "ECDHE"). For consistency with the code for DH
9114 certificates, use of ECDH certificates is now considered ECDH
9115 authentication, not RSA or ECDSA authentication (the latter is
9116 merely the CA's signing algorithm and not actively used in the
9119 The temporary ciphersuite alias "ECCdraft" is no longer
9120 available, and ECC ciphersuites are no longer excluded from "ALL"
9121 and "DEFAULT". The following aliases now exist for RFC 4492
9122 ciphersuites, most of these by analogy with the DH case:
9124 kECDHr - ECDH cert, signed with RSA
9125 kECDHe - ECDH cert, signed with ECDSA
9126 kECDH - ECDH cert (signed with either RSA or ECDSA)
9127 kEECDH - ephemeral ECDH
9128 ECDH - ECDH cert or ephemeral ECDH
9134 AECDH - anonymous ECDH
9135 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9139 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9140 Use correct micalg parameters depending on digest(s) in signed message.
9144 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9145 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9149 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9150 an engine to register a method. Add ENGINE lookups for methods and
9151 functional reference processing.
9155 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9156 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9161 * New -resign option to smime utility. This adds one or more signers
9162 to an existing PKCS#7 signedData structure. Also -md option to use an
9163 alternative message digest algorithm for signing.
9167 * Tidy up PKCS#7 routines and add new functions to make it easier to
9168 create PKCS7 structures containing multiple signers. Update smime
9169 application to support multiple signers.
9173 * New -macalg option to pkcs12 utility to allow setting of an alternative
9178 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9179 Reorganize PBE internals to lookup from a static table using NIDs,
9180 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9181 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9182 PRF which will be automatically used with PBES2.
9186 * Replace the algorithm specific calls to generate keys in "req" with the
9191 * Update PKCS#7 enveloped data routines to use new API. This is now
9192 supported by any public key method supporting the encrypt operation. A
9193 ctrl is added to allow the public key algorithm to examine or modify
9194 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9199 * Add a ctrl to asn1 method to allow a public key algorithm to express
9200 a default digest type to use. In most cases this will be SHA1 but some
9201 algorithms (such as GOST) need to specify an alternative digest. The
9202 return value indicates how strong the preference is 1 means optional and
9203 2 is mandatory (that is it is the only supported type). Modify
9204 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9205 use the default md. Update openssl utilities to use the default digest
9206 type for signing if it is not explicitly indicated.
9210 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9211 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9212 signing method from the key type. This effectively removes the link
9213 between digests and public key types.
9217 * Add an OID cross reference table and utility functions. Its purpose is to
9218 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9219 rsaEncryption. This will allow some of the algorithm specific hackery
9220 needed to use the correct OID to be removed.
9224 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9225 structures for PKCS7_sign(). They are now set up by the relevant public
9230 * Add provisional EC pkey method with support for ECDSA and ECDH.
9234 * Add support for key derivation (agreement) in the API, DH method and
9239 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9240 public and private key formats. As a side effect these add additional
9241 command line functionality not previously available: DSA signatures can be
9242 generated and verified using pkeyutl and DH key support and generation in
9249 *Oliver Tappe <zooey@hirschkaefer.de>*
9251 * New make target "install_html_docs" installs HTML renditions of the
9254 *Oliver Tappe <zooey@hirschkaefer.de>*
9256 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9257 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9258 support key and parameter generation and add initial key generation
9259 functionality for RSA.
9263 * Add functions for main EVP_PKEY_method operations. The undocumented
9264 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9265 `EVP_PKEY_{encrypt,decrypt}_old`.
9269 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9270 key API, doesn't do much yet.
9274 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9275 public key algorithms. New option to openssl utility:
9276 "list-public-key-algorithms" to print out info.
9280 * Implement the Supported Elliptic Curves Extension for
9281 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9285 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9286 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9290 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9291 utilities such as rsa, dsa, dsaparam etc except they process any key
9296 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9297 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9298 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9303 * Initial support for pluggable public key ASN1.
9304 De-spaghettify the public key ASN1 handling. Move public and private
9305 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9306 algorithm specific handling to a single module within the relevant
9307 algorithm directory. Add functions to allow (near) opaque processing
9308 of public and private key structures.
9312 * Implement the Supported Point Formats Extension for
9313 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9317 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9318 for the psk identity [hint] and the psk callback functions to the
9319 SSL_SESSION, SSL and SSL_CTX structure.
9322 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9326 SSL_CTX_use_psk_identity_hint
9327 SSL_get_psk_identity_hint
9328 SSL_get_psk_identity
9329 SSL_use_psk_identity_hint
9331 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9333 * Add RFC 3161 compliant time stamp request creation, response generation
9334 and response verification functionality.
9336 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9338 * Add initial support for TLS extensions, specifically for the server_name
9339 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9340 have new members for a host name. The SSL data structure has an
9341 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9342 stored in that context to allow for session resumption, even after the
9343 SSL has been switched to a new SSL_CTX in reaction to a client's
9344 server_name extension.
9346 New functions (subject to change):
9348 SSL_get_servername()
9349 SSL_get_servername_type()
9352 New CTRL codes and macros (subject to change):
9354 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9355 - SSL_CTX_set_tlsext_servername_callback()
9356 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9357 - SSL_CTX_set_tlsext_servername_arg()
9358 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9360 openssl s_client has a new '-servername ...' option.
9362 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9363 '-key2 ...', '-servername_fatal' (subject to change). This allows
9364 testing the HostName extension for a specific single host name ('-cert'
9365 and '-key' remain fallbacks for handshakes without HostName
9366 negotiation). If the unrecognized_name alert has to be sent, this by
9367 default is a warning; it becomes fatal with the '-servername_fatal'
9370 *Peter Sylvester, Remy Allais, Christophe Renou*
9372 * Whirlpool hash implementation is added.
9376 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9377 bn(64,32). Because of instruction set limitations it doesn't have
9378 any negative impact on performance. This was done mostly in order
9379 to make it possible to share assembler modules, such as bn_mul_mont
9380 implementations, between 32- and 64-bit builds without hassle.
9384 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9385 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9390 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9391 dedicated Montgomery multiplication procedure, is introduced.
9392 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9393 "64-bit" performance on certain 32-bit targets.
9397 * New option SSL_OP_NO_COMP to disable use of compression selectively
9398 in SSL structures. New SSL ctrl to set maximum send fragment size.
9399 Save memory by setting the I/O buffer sizes dynamically instead of
9400 using the maximum available value.
9404 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9405 in addition to the text details.
9409 * Very, very preliminary EXPERIMENTAL support for printing of general
9410 ASN1 structures. This currently produces rather ugly output and doesn't
9411 handle several customised structures at all.
9415 * Integrated support for PVK file format and some related formats such
9416 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9417 these in the 'rsa' and 'dsa' utilities.
9421 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9425 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9426 place for the (very old) "NETSCAPE" format certificates which are now
9427 handled using new ASN1 code equivalents.
9431 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9432 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9433 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9437 * Modify CRL distribution points extension code to print out previously
9438 unsupported fields. Enhance extension setting code to allow setting of
9443 * Add print and set support for Issuing Distribution Point CRL extension.
9447 * Change 'Configure' script to enable Camellia by default.
9454 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9456 * When rejecting SSL/TLS records due to an incorrect version number, never
9457 update s->server with a new major version number. As of
9458 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9459 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9460 the previous behavior could result in a read attempt at NULL when
9461 receiving specific incorrect SSL/TLS records once record payload
9462 protection is active. ([CVE-2010-0740])
9464 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9466 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9467 could be crashed if the relevant tables were not present (e.g. chrooted).
9469 *Tomas Hoger <thoger@redhat.com>*
9471 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9473 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9475 *Martin Olsson, Neel Mehta*
9477 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9478 accommodate for stack sorting, always a write lock!).
9482 * On some versions of WIN32 Heap32Next is very slow. This can cause
9483 excessive delays in the RAND_poll(): over a minute. As a workaround
9484 include a time check in the inner Heap32Next loop too.
9488 * The code that handled flushing of data in SSL/TLS originally used the
9489 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9490 the problem outlined in PR#1949. The fix suggested there however can
9491 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9492 of Apache). So instead simplify the code to flush unconditionally.
9493 This should be fine since flushing with no data to flush is a no op.
9497 * Handle TLS versions 2.0 and later properly and correctly use the
9498 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9499 off ancient servers have a habit of sticking around for a while...
9503 * Modify compression code so it frees up structures without using the
9504 ex_data callbacks. This works around a problem where some applications
9505 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9506 restarting) then use compression (e.g. SSL with compression) later.
9507 This results in significant per-connection memory leaks and
9508 has caused some security issues including CVE-2008-1678 and
9513 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9514 change when encrypting or decrypting.
9518 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9519 connect and renegotiate with servers which do not support RI.
9520 Until RI is more widely deployed this option is enabled by default.
9524 * Add "missing" ssl ctrls to clear options and mode.
9528 * If client attempts to renegotiate and doesn't support RI respond with
9529 a no_renegotiation alert as required by RFC5746. Some renegotiating
9530 TLS clients will continue a connection gracefully when they receive
9531 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9532 waiting for a server hello which it will never receive. Now we treat a
9533 received no_renegotiation alert as a fatal error. This is because
9534 applications requesting a renegotiation might well expect it to succeed
9535 and would have no code in place to handle the server denying it so the
9536 only safe thing to do is to terminate the connection.
9540 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9541 peer supports secure renegotiation and 0 otherwise. Print out peer
9542 renegotiation support in s_client/s_server.
9546 * Replace the highly broken and deprecated SPKAC certification method with
9547 the updated NID creation version. This should correctly handle UTF8.
9551 * Implement RFC5746. Re-enable renegotiation but require the extension
9552 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9553 turns out to be a bad idea. It has been replaced by
9554 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9555 SSL_CTX_set_options(). This is really not recommended unless you
9556 know what you are doing.
9558 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9560 * Fixes to stateless session resumption handling. Use initial_ctx when
9561 issuing and attempting to decrypt tickets in case it has changed during
9562 servername handling. Use a non-zero length session ID when attempting
9563 stateless session resumption: this makes it possible to determine if
9564 a resumption has occurred immediately after receiving server hello
9565 (several places in OpenSSL subtly assume this) instead of later in
9570 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9571 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9572 fixes for a few places where the return code is not checked
9575 *Julia Lawall <julia@diku.dk>*
9577 * Add --strict-warnings option to Configure script to include devteam
9578 warnings in other configurations.
9582 * Add support for --libdir option and LIBDIR variable in makefiles. This
9583 makes it possible to install openssl libraries in locations which
9584 have names other than "lib", for example "/usr/lib64" which some
9587 *Steve Henson, based on patch from Jeremy Utley*
9589 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9590 X690 8.9.12 and can produce some misleading textual output of OIDs.
9592 *Steve Henson, reported by Dan Kaminsky*
9594 * Delete MD2 from algorithm tables. This follows the recommendation in
9595 several standards that it is not used in new applications due to
9596 several cryptographic weaknesses. For binary compatibility reasons
9597 the MD2 API is still compiled in by default.
9601 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9606 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9607 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9610 *Guenter <lists@gknw.net>*
9612 * Fix the server certificate chain building code to use X509_verify_cert(),
9613 it used to have an ad-hoc builder which was unable to cope with anything
9614 other than a simple chain.
9616 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9618 * Don't check self signed certificate signatures in X509_verify_cert()
9619 by default (a flag can override this): it just wastes time without
9620 adding any security. As a useful side effect self signed root CAs
9621 with non-FIPS digests are now usable in FIPS mode.
9625 * In dtls1_process_out_of_seq_message() the check if the current message
9626 is already buffered was missing. For every new message was memory
9627 allocated, allowing an attacker to perform an denial of service attack
9628 with sending out of seq handshake messages until there is no memory
9629 left. Additionally every future message was buffered, even if the
9630 sequence number made no sense and would be part of another handshake.
9631 So only messages with sequence numbers less than 10 in advance will be
9632 buffered. ([CVE-2009-1378])
9634 *Robin Seggelmann, discovered by Daniel Mentz*
9636 * Records are buffered if they arrive with a future epoch to be
9637 processed after finishing the corresponding handshake. There is
9638 currently no limitation to this buffer allowing an attacker to perform
9639 a DOS attack with sending records with future epochs until there is no
9640 memory left. This patch adds the pqueue_size() function to determine
9641 the size of a buffer and limits the record buffer to 100 entries.
9644 *Robin Seggelmann, discovered by Daniel Mentz*
9646 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9647 parent structure is freed. ([CVE-2009-1379])
9651 * Handle non-blocking I/O properly in SSL_shutdown() call.
9653 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9655 * Add `2.5.4.*` OIDs
9657 *Ilya O. <vrghost@gmail.com>*
9659 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9661 * Disable renegotiation completely - this fixes a severe security
9662 problem ([CVE-2009-3555]) at the cost of breaking all
9663 renegotiation. Renegotiation can be re-enabled by setting
9664 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9665 run-time. This is really not recommended unless you know what
9670 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9672 * Don't set val to NULL when freeing up structures, it is freed up by
9673 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9674 zeroing past the valid field. ([CVE-2009-0789])
9676 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9678 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9679 checked correctly. This would allow some invalid signed attributes to
9680 appear to verify correctly. ([CVE-2009-0591])
9682 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9684 * Reject UniversalString and BMPString types with invalid lengths. This
9685 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9686 a legal length. ([CVE-2009-0590])
9690 * Set S/MIME signing as the default purpose rather than setting it
9691 unconditionally. This allows applications to override it at the store
9696 * Permit restricted recursion of ASN1 strings. This is needed in practice
9697 to handle some structures.
9701 * Improve efficiency of mem_gets: don't search whole buffer each time
9704 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9706 * New -hex option for openssl rand.
9710 * Print out UTF8String and NumericString when parsing ASN1.
9714 * Support NumericString type for name components.
9718 * Allow CC in the environment to override the automatically chosen
9719 compiler. Note that nothing is done to ensure flags work with the
9724 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9726 * Properly check EVP_VerifyFinal() and similar return values
9729 *Ben Laurie, Bodo Moeller, Google Security Team*
9731 * Enable TLS extensions by default.
9735 * Allow the CHIL engine to be loaded, whether the application is
9736 multithreaded or not. (This does not release the developer from the
9737 obligation to set up the dynamic locking callbacks.)
9739 *Sander Temme <sander@temme.net>*
9741 * Use correct exit code if there is an error in dgst command.
9743 *Steve Henson; problem pointed out by Roland Dirlewanger*
9745 * Tweak Configure so that you need to say "experimental-jpake" to enable
9746 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9750 * Add experimental JPAKE support, including demo authentication in
9751 s_client and s_server.
9755 * Set the comparison function in v3_addr_canonize().
9757 *Rob Austein <sra@hactrn.net>*
9759 * Add support for XMPP STARTTLS in s_client.
9761 *Philip Paeps <philip@freebsd.org>*
9763 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9764 to ensure that even with this option, only ciphersuites in the
9765 server's preference list will be accepted. (Note that the option
9766 applies only when resuming a session, so the earlier behavior was
9767 just about the algorithm choice for symmetric cryptography.)
9771 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9773 * Fix NULL pointer dereference if a DTLS server received
9774 ChangeCipherSpec as first record ([CVE-2009-1386]).
9778 * Fix a state transition in s3_srvr.c and d1_srvr.c
9779 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9783 * The fix in 0.9.8c that supposedly got rid of unsafe
9784 double-checked locking was incomplete for RSA blinding,
9785 addressing just one layer of what turns out to have been
9786 doubly unsafe triple-checked locking.
9788 So now fix this for real by retiring the MONT_HELPER macro
9789 in crypto/rsa/rsa_eay.c.
9791 *Bodo Moeller; problem pointed out by Marius Schilder*
9793 * Various precautionary measures:
9795 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9797 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9798 (NB: This would require knowledge of the secret session ticket key
9799 to exploit, in which case you'd be SOL either way.)
9801 - Change bn_nist.c so that it will properly handle input BIGNUMs
9802 outside the expected range.
9804 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9807 *Neel Mehta, Bodo Moeller*
9809 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9810 the load fails. Useful for distros.
9812 *Ben Laurie and the FreeBSD team*
9814 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9818 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9822 * Expand ENGINE to support engine supplied SSL client certificate functions.
9824 This work was sponsored by Logica.
9828 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9829 keystores. Support for SSL/TLS client authentication too.
9830 Not compiled unless enable-capieng specified to Configure.
9832 This work was sponsored by Logica.
9836 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9837 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9838 attribute creation routines such as certificate requests and PKCS#12
9843 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9845 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9846 handshake which could lead to a client crash as found using the
9847 Codenomicon TLS test suite ([CVE-2008-1672])
9849 *Steve Henson, Mark Cox*
9851 * Fix double free in TLS server name extensions which could lead to
9852 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9856 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9858 Clear the error queue to ensure that error entries left from
9859 older function calls do not interfere with the correct operation.
9861 *Lutz Jaenicke, Erik de Castro Lopo*
9863 * Remove root CA certificates of commercial CAs:
9865 The OpenSSL project does not recommend any specific CA and does not
9866 have any policy with respect to including or excluding any CA.
9867 Therefore it does not make any sense to ship an arbitrary selection
9868 of root CA certificates with the OpenSSL software.
9872 * RSA OAEP patches to fix two separate invalid memory reads.
9873 The first one involves inputs when 'lzero' is greater than
9874 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9875 before the beginning of from). The second one involves inputs where
9876 the 'db' section contains nothing but zeroes (there is a one-byte
9877 invalid read after the end of 'db').
9879 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9881 * Partial backport from 0.9.9-dev:
9883 Introduce bn_mul_mont (dedicated Montgomery multiplication
9884 procedure) as a candidate for BIGNUM assembler implementation.
9885 While 0.9.9-dev uses assembler for various architectures, only
9886 x86_64 is available by default here in the 0.9.8 branch, and
9887 32-bit x86 is available through a compile-time setting.
9889 To try the 32-bit x86 assembler implementation, use Configure
9890 option "enable-montasm" (which exists only for this backport).
9892 As "enable-montasm" for 32-bit x86 disclaims code stability
9893 anyway, in this constellation we activate additional code
9894 backported from 0.9.9-dev for further performance improvements,
9895 namely BN_from_montgomery_word. (To enable this otherwise,
9896 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9898 *Andy Polyakov (backport partially by Bodo Moeller)*
9900 * Add TLS session ticket callback. This allows an application to set
9901 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9902 values. This is useful for key rollover for example where several key
9903 sets may exist with different names.
9907 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9908 This was broken until now in 0.9.8 releases, such that the only way
9909 a registered ENGINE could be used (assuming it initialises
9910 successfully on the host) was to explicitly set it as the default
9911 for the relevant algorithms. This is in contradiction with 0.9.7
9912 behaviour and the documentation. With this fix, when an ENGINE is
9913 registered into a given algorithm's table of implementations, the
9914 'uptodate' flag is reset so that auto-discovery will be used next
9915 time a new context for that algorithm attempts to select an
9918 *Ian Lister (tweaked by Geoff Thorpe)*
9920 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9921 implementation in the following ways:
9923 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9926 Lack of BER streaming support means one pass streaming processing is
9927 only supported if data is detached: setting the streaming flag is
9928 ignored for embedded content.
9930 CMS support is disabled by default and must be explicitly enabled
9931 with the enable-cms configuration option.
9935 * Update the GMP engine glue to do direct copies between BIGNUM and
9936 mpz_t when openssl and GMP use the same limb size. Otherwise the
9937 existing "conversion via a text string export" trick is still used.
9939 *Paul Sheer <paulsheer@gmail.com>*
9941 * Zlib compression BIO. This is a filter BIO which compressed and
9942 uncompresses any data passed through it.
9946 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9947 RFC3394 compatible AES key wrapping.
9951 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9952 sets string data without copying. X509_ALGOR_set0() and
9953 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9954 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9955 from an X509_ATTRIBUTE structure optionally checking it occurs only
9956 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9961 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9962 to get the expected BN_FLG_CONSTTIME behavior.
9964 *Bodo Moeller (Google)*
9968 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9969 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9970 - added some more tests to do_tests.pl
9971 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9972 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9973 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9974 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9975 - various changes to netware.pl to enable gcc-cross builds on Win32
9977 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9978 - various changes to fix missing prototype warnings
9979 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9980 - added AES, WHIRLPOOL and CPUID assembler code to build files
9981 - added missing AES assembler make rules to mk1mf.pl
9982 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9984 *Guenter Knauf <eflash@gmx.net>*
9986 * Implement certificate status request TLS extension defined in RFC3546.
9987 A client can set the appropriate parameters and receive the encoded
9988 OCSP response via a callback. A server can query the supplied parameters
9989 and set the encoded OCSP response in the callback. Add simplified examples
9990 to s_client and s_server.
9994 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9997 + Binary incompatibility of ssl_ctx_st structure
9998 + DTLS interoperation with non-compliant servers
9999 + Don't call get_session_cb() without proposed session
10000 + Fix ia64 assembler code
10002 *Andy Polyakov, Steve Henson*
10004 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10006 * DTLS Handshake overhaul. There were longstanding issues with
10007 OpenSSL DTLS implementation, which were making it impossible for
10008 RFC 4347 compliant client to communicate with OpenSSL server.
10009 Unfortunately just fixing these incompatibilities would "cut off"
10010 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10011 server keeps tolerating non RFC compliant syntax. The opposite is
10012 not true, 0.9.8f client can not communicate with earlier server.
10013 This update even addresses CVE-2007-4995.
10017 * Changes to avoid need for function casts in OpenSSL: some compilers
10018 (gcc 4.2 and later) reject their use.
10019 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10022 * Add RFC4507 support to OpenSSL. This includes the corrections in
10023 RFC4507bis. The encrypted ticket format is an encrypted encoded
10024 SSL_SESSION structure, that way new session features are automatically
10027 If a client application caches session in an SSL_SESSION structure
10028 support is transparent because tickets are now stored in the encoded
10031 The SSL_CTX structure automatically generates keys for ticket
10032 protection in servers so again support should be possible
10033 with no application modification.
10035 If a client or server wishes to disable RFC4507 support then the option
10036 SSL_OP_NO_TICKET can be set.
10038 Add a TLS extension debugging callback to allow the contents of any client
10039 or server extensions to be examined.
10041 This work was sponsored by Google.
10045 * Add initial support for TLS extensions, specifically for the server_name
10046 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10047 have new members for a host name. The SSL data structure has an
10048 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10049 stored in that context to allow for session resumption, even after the
10050 SSL has been switched to a new SSL_CTX in reaction to a client's
10051 server_name extension.
10053 New functions (subject to change):
10055 SSL_get_servername()
10056 SSL_get_servername_type()
10059 New CTRL codes and macros (subject to change):
10061 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10062 - SSL_CTX_set_tlsext_servername_callback()
10063 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10064 - SSL_CTX_set_tlsext_servername_arg()
10065 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10067 openssl s_client has a new '-servername ...' option.
10069 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10070 '-key2 ...', '-servername_fatal' (subject to change). This allows
10071 testing the HostName extension for a specific single host name ('-cert'
10072 and '-key' remain fallbacks for handshakes without HostName
10073 negotiation). If the unrecognized_name alert has to be sent, this by
10074 default is a warning; it becomes fatal with the '-servername_fatal'
10077 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10079 * Add AES and SSE2 assembly language support to VC++ build.
10083 * Mitigate attack on final subtraction in Montgomery reduction.
10087 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10088 (which previously caused an internal error).
10092 * Squeeze another 10% out of IGE mode when in != out.
10096 * AES IGE mode speedup.
10098 *Dean Gaudet (Google)*
10100 * Add the Korean symmetric 128-bit cipher SEED (see
10101 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10102 add SEED ciphersuites from RFC 4162:
10104 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10105 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10106 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10107 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10109 To minimize changes between patchlevels in the OpenSSL 0.9.8
10110 series, SEED remains excluded from compilation unless OpenSSL
10111 is configured with 'enable-seed'.
10113 *KISA, Bodo Moeller*
10115 * Mitigate branch prediction attacks, which can be practical if a
10116 single processor is shared, allowing a spy process to extract
10117 information. For detailed background information, see
10118 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10119 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10120 and Necessary Software Countermeasures"). The core of the change
10121 are new versions BN_div_no_branch() and
10122 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10123 respectively, which are slower, but avoid the security-relevant
10124 conditional branches. These are automatically called by BN_div()
10125 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10126 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10127 remove a conditional branch.
10129 BN_FLG_CONSTTIME is the new name for the previous
10130 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10131 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10132 in the exponent causes BN_mod_exp_mont() to use the alternative
10133 implementation in BN_mod_exp_mont_consttime().) The old name
10134 remains as a deprecated alias.
10136 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10137 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10138 constant-time implementations for more than just exponentiation.
10139 Here too the old name is kept as a deprecated alias.
10141 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10142 the BN_BLINDING structure gets an independent copy of the
10143 modulus. This means that the previous `BIGNUM *m` argument to
10144 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10145 essentially becomes `const BIGNUM *m`, although we can't actually
10146 change this in the header file before 0.9.9. It allows
10147 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10148 enable BN_FLG_CONSTTIME.
10150 *Matthew D Wood (Intel Corp)*
10152 * In the SSL/TLS server implementation, be strict about session ID
10153 context matching (which matters if an application uses a single
10154 external cache for different purposes). Previously,
10155 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10156 set. This did ensure strict client verification, but meant that,
10157 with applications using a single external cache for quite
10158 different requirements, clients could circumvent ciphersuite
10159 restrictions for a given session ID context by starting a session
10160 in a different context.
10164 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10165 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10166 authentication-only ciphersuites.
10170 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10171 not complete and could lead to a possible single byte overflow
10172 ([CVE-2007-5135]) [Ben Laurie]
10174 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10176 * Since AES128 and AES256 (and similarly Camellia128 and
10177 Camellia256) share a single mask bit in the logic of
10178 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10179 kludge to work properly if AES128 is available and AES256 isn't
10180 (or if Camellia128 is available and Camellia256 isn't).
10184 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10185 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10186 When a point or a seed is encoded in a BIT STRING, we need to
10187 prevent the removal of trailing zero bits to get the proper DER
10188 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10189 of a NamedBitList, for which trailing 0 bits need to be removed.)
10193 * Have SSL/TLS server implementation tolerate "mismatched" record
10194 protocol version while receiving ClientHello even if the
10195 ClientHello is fragmented. (The server can't insist on the
10196 particular protocol version it has chosen before the ServerHello
10197 message has informed the client about his choice.)
10201 * Add RFC 3779 support.
10203 *Rob Austein for ARIN, Ben Laurie*
10205 * Load error codes if they are not already present instead of using a
10206 static variable. This allows them to be cleanly unloaded and reloaded.
10207 Improve header file function name parsing.
10211 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10212 or CAPABILITY handshake as required by RFCs.
10214 *Goetz Babin-Ebell*
10216 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10218 * Introduce limits to prevent malicious keys being able to
10219 cause a denial of service. ([CVE-2006-2940])
10221 *Steve Henson, Bodo Moeller*
10223 * Fix ASN.1 parsing of certain invalid structures that can result
10224 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10226 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10227 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10229 * Fix SSL client code which could crash if connecting to a
10230 malicious SSLv2 server. ([CVE-2006-4343])
10232 *Tavis Ormandy and Will Drewry, Google Security Team*
10234 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10235 match only those. Before that, "AES256-SHA" would be interpreted
10236 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10237 the same strength classification in 0.9.7h) as we currently only
10238 have a single AES bit in the ciphersuite description bitmap.
10239 That change, however, also applied to ciphersuite strings such as
10240 "RC4-MD5" that intentionally matched multiple ciphersuites --
10241 namely, SSL 2.0 ciphersuites in addition to the more common ones
10242 from SSL 3.0/TLS 1.0.
10244 So we change the selection algorithm again: Naming an explicit
10245 ciphersuite selects this one ciphersuite, and any other similar
10246 ciphersuite (same bitmap) from *other* protocol versions.
10247 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10248 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10250 Since SSL 2.0 does not have any ciphersuites for which the
10251 128/256 bit distinction would be relevant, this works for now.
10252 The proper fix will be to use different bits for AES128 and
10253 AES256, which would have avoided the problems from the beginning;
10254 however, bits are scarce, so we can only do this in a new release
10255 (not just a patchlevel) when we can change the SSL_CIPHER
10256 definition to split the single 'unsigned long mask' bitmap into
10257 multiple values to extend the available space.
10261 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10263 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10264 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10266 * Add AES IGE and biIGE modes.
10270 * Change the Unix randomness entropy gathering to use poll() when
10271 possible instead of select(), since the latter has some
10272 undesirable limitations.
10274 *Darryl Miles via Richard Levitte and Bodo Moeller*
10276 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10277 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10278 cannot be implicitly activated as part of, e.g., the "AES" alias.
10279 However, please upgrade to OpenSSL 0.9.9[-dev] for
10280 non-experimental use of the ECC ciphersuites to get TLS extension
10281 support, which is required for curve and point format negotiation
10282 to avoid potential handshake problems.
10286 * Disable rogue ciphersuites:
10288 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10289 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10290 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10292 The latter two were purportedly from
10293 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10296 Also deactivate the remaining ciphersuites from
10297 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10298 unofficial, and the ID has long expired.
10302 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10303 dual-core machines) and other potential thread-safety issues.
10307 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10308 versions), which is now available for royalty-free use
10309 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10310 Also, add Camellia TLS ciphersuites from RFC 4132.
10312 To minimize changes between patchlevels in the OpenSSL 0.9.8
10313 series, Camellia remains excluded from compilation unless OpenSSL
10314 is configured with 'enable-camellia'.
10318 * Disable the padding bug check when compression is in use. The padding
10319 bug check assumes the first packet is of even length, this is not
10320 necessarily true if compression is enabled and can result in false
10321 positives causing handshake failure. The actual bug test is ancient
10322 code so it is hoped that implementations will either have fixed it by
10323 now or any which still have the bug do not support compression.
10327 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10329 * When applying a cipher rule check to see if string match is an explicit
10330 cipher suite and only match that one cipher suite if it is.
10334 * Link in manifests for VC++ if needed.
10336 *Austin Ziegler <halostatue@gmail.com>*
10338 * Update support for ECC-based TLS ciphersuites according to
10339 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10340 TLS extensions, which are supported starting with the 0.9.9
10341 branch, not in the OpenSSL 0.9.8 branch).
10345 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10346 opaque EVP_CIPHER_CTX handling.
10350 * Fixes and enhancements to zlib compression code. We now only use
10351 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10352 to conform with the standards mentioned here:
10353 <http://www.zlib.net/DLL_FAQ.txt>
10354 Static zlib linking now works on Windows and the new --with-zlib-include
10355 --with-zlib-lib options to Configure can be used to supply the location
10356 of the headers and library. Gracefully handle case where zlib library
10361 * Several fixes and enhancements to the OID generation code. The old code
10362 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10363 handle numbers larger than ULONG_MAX, truncated printing and had a
10364 non standard OBJ_obj2txt() behaviour.
10368 * Add support for building of engines under engine/ as shared libraries
10369 under VC++ build system.
10373 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10374 Hopefully, we will not see any false combination of paths any more.
10378 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10380 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10381 (part of SSL_OP_ALL). This option used to disable the
10382 countermeasure against man-in-the-middle protocol-version
10383 rollback in the SSL 2.0 server implementation, which is a bad
10384 idea. ([CVE-2005-2969])
10386 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10387 for Information Security, National Institute of Advanced Industrial
10388 Science and Technology [AIST], Japan)*
10390 * Add two function to clear and return the verify parameter flags.
10394 * Keep cipherlists sorted in the source instead of sorting them at
10395 runtime, thus removing the need for a lock.
10399 * Avoid some small subgroup attacks in Diffie-Hellman.
10401 *Nick Mathewson and Ben Laurie*
10403 * Add functions for well-known primes.
10407 * Extended Windows CE support.
10409 *Satoshi Nakamura and Andy Polyakov*
10411 * Initialize SSL_METHOD structures at compile time instead of during
10412 runtime, thus removing the need for a lock.
10416 * Make PKCS7_decrypt() work even if no certificate is supplied by
10417 attempting to decrypt each encrypted key in turn. Add support to
10422 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10424 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10427 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10431 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10432 key into the same file any more.
10436 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10440 * Add -utf8 command line and config file option to 'ca'.
10442 *Stefan <stf@udoma.org*
10444 * Removed the macro des_crypt(), as it seems to conflict with some
10445 libraries. Use DES_crypt().
10449 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10450 involves renaming the source and generated shared-libs for
10451 both. The engines will accept the corrected or legacy ids
10452 ('ncipher' and '4758_cca' respectively) when binding. NB,
10453 this only applies when building 'shared'.
10455 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10457 * Add attribute functions to EVP_PKEY structure. Modify
10458 PKCS12_create() to recognize a CSP name attribute and
10459 use it. Make -CSP option work again in pkcs12 utility.
10463 * Add new functionality to the bn blinding code:
10464 - automatic re-creation of the BN_BLINDING parameters after
10465 a fixed number of uses (currently 32)
10466 - add new function for parameter creation
10467 - introduce flags to control the update behaviour of the
10468 BN_BLINDING parameters
10469 - hide BN_BLINDING structure
10470 Add a second BN_BLINDING slot to the RSA structure to improve
10471 performance when a single RSA object is shared among several
10476 * Add support for DTLS.
10478 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10480 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10481 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10485 * Remove buggy and incomplete DH cert support from
10486 ssl/ssl_rsa.c and ssl/s3_both.c
10490 * Use SHA-1 instead of MD5 as the default digest algorithm for
10491 the `apps/openssl` commands.
10495 * Compile clean with "-Wall -Wmissing-prototypes
10496 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10497 DEBUG_SAFESTACK must also be set.
10501 * Change ./Configure so that certain algorithms can be disabled by default.
10502 The new counterpiece to "no-xxx" is "enable-xxx".
10504 The patented RC5 and MDC2 algorithms will now be disabled unless
10505 "enable-rc5" and "enable-mdc2", respectively, are specified.
10507 (IDEA remains enabled despite being patented. This is because IDEA
10508 is frequently required for interoperability, and there is no license
10509 fee for non-commercial use. As before, "no-idea" can be used to
10510 avoid this algorithm.)
10514 * Add processing of proxy certificates (see RFC 3820). This work was
10515 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10516 EGEE (Enabling Grids for E-science in Europe).
10520 * RC4 performance overhaul on modern architectures/implementations, such
10521 as Intel P4, IA-64 and AMD64.
10525 * New utility extract-section.pl. This can be used specify an alternative
10526 section number in a pod file instead of having to treat each file as
10527 a separate case in Makefile. This can be done by adding two lines to the
10530 =for comment openssl_section:XXX
10532 The blank line is mandatory.
10536 * New arguments -certform, -keyform and -pass for s_client and s_server
10537 to allow alternative format key and certificate files and passphrase
10542 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10543 update associated structures and add various utility functions.
10545 Add new policy related verify parameters, include policy checking in
10546 standard verify code. Enhance 'smime' application with extra parameters
10547 to support policy checking and print out.
10551 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10552 Nehemiah processors. These extensions support AES encryption in hardware
10553 as well as RNG (though RNG support is currently disabled).
10555 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10557 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10561 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10563 *Andy Polyakov and a number of other people*
10565 * Improved PowerPC platform support. Most notably BIGNUM assembler
10566 implementation contributed by IBM.
10568 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10570 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10571 exponent rather than 'unsigned long'. There is a corresponding change to
10572 the new 'rsa_keygen' element of the RSA_METHOD structure.
10574 *Jelte Jansen, Geoff Thorpe*
10576 * Functionality for creating the initial serial number file is now
10577 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10579 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10580 number file to 1, which is bound to cause problems. To avoid
10581 the problems while respecting compatibility between different 0.9.7
10582 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10583 CA.pl for serial number initialization. With the new release 0.9.8,
10584 we can fix the problem directly in the 'ca' utility.)
10588 * Reduced header interdependencies by declaring more opaque objects in
10589 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10590 give fewer recursive includes, which could break lazy source code - so
10591 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10592 developers should define this symbol when building and using openssl to
10593 ensure they track the recommended behaviour, interfaces, [etc], but
10594 backwards-compatible behaviour prevails when this isn't defined.
10598 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10602 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10603 This will generate a random key of the appropriate length based on the
10604 cipher context. The EVP_CIPHER can provide its own random key generation
10605 routine to support keys of a specific form. This is used in the des and
10606 3des routines to generate a key of the correct parity. Update S/MIME
10607 code to use new functions and hence generate correct parity DES keys.
10608 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10609 valid (weak or incorrect parity).
10613 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10614 as looking them up. This is useful when the verified structure may contain
10615 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10616 present unless the new PKCS7_NO_CRL flag is asserted.
10620 * Extend ASN1 oid configuration module. It now additionally accepts the
10623 shortName = some long name, 1.2.3.4
10627 * Reimplemented the BN_CTX implementation. There is now no more static
10628 limitation on the number of variables it can handle nor the depth of the
10629 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10630 information can now expand as required, and rather than having a single
10631 static array of bignums, BN_CTX now uses a linked-list of such arrays
10632 allowing it to expand on demand whilst maintaining the usefulness of
10633 BN_CTX's "bundling".
10637 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10638 to allow all RSA operations to function using a single BN_CTX.
10642 * Preliminary support for certificate policy evaluation and checking. This
10643 is initially intended to pass the tests outlined in "Conformance Testing
10644 of Relying Party Client Certificate Path Processing Logic" v1.07.
10648 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10649 remained unused and not that useful. A variety of other little bignum
10650 tweaks and fixes have also been made continuing on from the audit (see
10655 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10656 associated ASN1, EVP and SSL functions and old ASN1 macros.
10660 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10661 and this should never fail. So the return value from the use of
10662 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10663 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10667 * BN_CTX_get() should return zero-valued bignums, providing the same
10668 initialised value as BN_new().
10670 *Geoff Thorpe, suggested by Ulf Möller*
10672 * Support for inhibitAnyPolicy certificate extension.
10676 * An audit of the BIGNUM code is underway, for which debugging code is
10677 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10678 is considered valid when processing BIGNUMs, and causes execution to
10679 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10680 further steps are taken to deliberately pollute unused data in BIGNUM
10681 structures to try and expose faulty code further on. For now, openssl will
10682 (in its default mode of operation) continue to tolerate the inconsistent
10683 forms that it has tolerated in the past, but authors and packagers should
10684 consider trying openssl and their own applications when compiled with
10685 these debugging symbols defined. It will help highlight potential bugs in
10686 their own code, and will improve the test coverage for OpenSSL itself. At
10687 some point, these tighter rules will become openssl's default to improve
10688 maintainability, though the assert()s and other overheads will remain only
10689 in debugging configurations. See bn.h for more details.
10691 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10693 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10694 that can only be obtained through BN_CTX_new() (which implicitly
10695 initialises it). The presence of this function only made it possible
10696 to overwrite an existing structure (and cause memory leaks).
10700 * Because of the callback-based approach for implementing LHASH as a
10701 template type, lh_insert() adds opaque objects to hash-tables and
10702 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10703 to clean up those corresponding objects before destroying the hash table
10704 (and losing the object pointers). So some over-zealous constifications in
10705 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10706 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10707 prototyped to have "const" restrictions on the object pointers they are
10708 given (and so aren't required to cast them away any more).
10712 * The tmdiff.h API was so ugly and minimal that our own timing utility
10713 (speed) prefers to use its own implementation. The two implementations
10714 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10715 its object type properly exposed (MS_TM) instead of casting to/from
10716 `char *`. This may still change yet if someone realises MS_TM and
10718 aren't necessarily the greatest nomenclatures - but this is what was used
10719 internally to the implementation so I've used that for now.
10723 * Ensure that deprecated functions do not get compiled when
10724 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10725 the self-tests were still using deprecated key-generation functions so
10726 these have been updated also.
10730 * Reorganise PKCS#7 code to separate the digest location functionality
10731 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10732 New function PKCS7_set_digest() to set the digest type for PKCS#7
10733 digestedData type. Add additional code to correctly generate the
10734 digestedData type and add support for this type in PKCS7 initialization
10739 * New function PKCS7_set0_type_other() this initializes a PKCS7
10740 structure of type "other".
10744 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10745 sure the loop does correctly stop and breaking ("division by zero")
10746 modulus operations are not performed. The (pre-generated) prime
10747 table crypto/bn/bn_prime.h was already correct, but it could not be
10748 re-generated on some platforms because of the "division by zero"
10749 situation in the script.
10751 *Ralf S. Engelschall*
10753 * Update support for ECC-based TLS ciphersuites according to
10754 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10755 SHA-1 now is only used for "small" curves (where the
10756 representation of a field element takes up to 24 bytes); for
10757 larger curves, the field element resulting from ECDH is directly
10758 used as premaster secret.
10760 *Douglas Stebila (Sun Microsystems Laboratories)*
10762 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10763 curve secp160r1 to the tests.
10765 *Douglas Stebila (Sun Microsystems Laboratories)*
10767 * Add the possibility to load symbols globally with DSO.
10769 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10771 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10772 control of the error stack.
10776 * Add support for STORE in ENGINE.
10780 * Add the STORE type. The intention is to provide a common interface
10781 to certificate and key stores, be they simple file-based stores, or
10782 HSM-type store, or LDAP stores, or...
10783 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10787 * Add a generic structure called OPENSSL_ITEM. This can be used to
10788 pass a list of arguments to any function as well as provide a way
10789 for a function to pass data back to the caller.
10793 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10794 works like BUF_strdup() but can be used to duplicate a portion of
10795 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10800 * Add the function sk_find_ex() which works like sk_find(), but will
10801 return an index to an element even if an exact match couldn't be
10802 found. The index is guaranteed to point at the element where the
10803 searched-for key would be inserted to preserve sorting order.
10807 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10808 takes an extra flags argument for optional functionality. Currently,
10809 the following flags are defined:
10811 OBJ_BSEARCH_VALUE_ON_NOMATCH
10812 This one gets OBJ_bsearch_ex() to return a pointer to the first
10813 element where the comparing function returns a negative or zero
10816 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10817 This one gets OBJ_bsearch_ex() to return a pointer to the first
10818 element where the comparing function returns zero. This is useful
10819 if there are more than one element where the comparing function
10824 * Make it possible to create self-signed certificates with 'openssl ca'
10825 in such a way that the self-signed certificate becomes part of the
10826 CA database and uses the same mechanisms for serial number generation
10827 as all other certificate signing. The new flag '-selfsign' enables
10828 this functionality. Adapt CA.sh and CA.pl.in.
10832 * Add functionality to check the public key of a certificate request
10833 against a given private. This is useful to check that a certificate
10834 request can be signed by that key (self-signing).
10838 * Make it possible to have multiple active certificates with the same
10839 subject in the CA index file. This is done only if the keyword
10840 'unique_subject' is set to 'no' in the main CA section (default
10841 if 'CA_default') of the configuration file. The value is saved
10842 with the database itself in a separate index attribute file,
10843 named like the index file with '.attr' appended to the name.
10847 * Generate multi-valued AVAs using '+' notation in config files for
10852 * Support for nameConstraints certificate extension.
10856 * Support for policyConstraints certificate extension.
10860 * Support for policyMappings certificate extension.
10864 * Make sure the default DSA_METHOD implementation only uses its
10865 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10866 and change its own handlers to be NULL so as to remove unnecessary
10867 indirection. This lets alternative implementations fallback to the
10868 default implementation more easily.
10872 * Support for directoryName in GeneralName related extensions
10877 * Make it possible to link applications using Makefile.shared.
10878 Make that possible even when linking against static libraries!
10882 * Support for single pass processing for S/MIME signing. This now
10883 means that S/MIME signing can be done from a pipe, in addition
10884 cleartext signing (multipart/signed type) is effectively streaming
10885 and the signed data does not need to be all held in memory.
10887 This is done with a new flag PKCS7_STREAM. When this flag is set
10888 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10889 is done after the data is output (and digests calculated) in
10890 SMIME_write_PKCS7().
10894 * Add full support for -rpath/-R, both in shared libraries and
10895 applications, at least on the platforms where it's known how
10900 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10901 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10902 will now compute a table of multiples of the generator that
10903 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10904 faster (notably in the case of a single point multiplication,
10905 scalar * generator).
10907 *Nils Larsch, Bodo Moeller*
10909 * IPv6 support for certificate extensions. The various extensions
10910 which use the IP:a.b.c.d can now take IPv6 addresses using the
10911 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10916 * Added an ENGINE that implements RSA by performing private key
10917 exponentiations with the GMP library. The conversions to and from
10918 GMP's mpz_t format aren't optimised nor are any montgomery forms
10919 cached, and on x86 it appears OpenSSL's own performance has caught up.
10920 However there are likely to be other architectures where GMP could
10921 provide a boost. This ENGINE is not built in by default, but it can be
10922 specified at Configure time and should be accompanied by the necessary
10923 linker additions, eg;
10924 ./config -DOPENSSL_USE_GMP -lgmp
10928 * "openssl engine" will not display ENGINE/DSO load failure errors when
10929 testing availability of engines with "-t" - the old behaviour is
10930 produced by increasing the feature's verbosity with "-tt".
10934 * ECDSA routines: under certain error conditions uninitialized BN objects
10935 could be freed. Solution: make sure initialization is performed early
10936 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10941 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10942 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10943 software implementations. For DSA and DH, parameter generation can
10944 also be overridden by providing the appropriate method callbacks.
10948 * Change the "progress" mechanism used in key-generation and
10949 primality testing to functions that take a new BN_GENCB pointer in
10950 place of callback/argument pairs. The new API functions have `_ex`
10951 postfixes and the older functions are reimplemented as wrappers for
10952 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10953 declarations of the old functions to help (graceful) attempts to
10954 migrate to the new functions. Also, the new key-generation API
10955 functions operate on a caller-supplied key-structure and return
10956 success/failure rather than returning a key or NULL - this is to
10957 help make "keygen" another member function of RSA_METHOD etc.
10959 Example for using the new callback interface:
10961 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10962 void *my_arg = ...;
10965 BN_GENCB_set(&my_cb, my_callback, my_arg);
10967 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10968 /* For the meaning of a, b in calls to my_callback(), see the
10969 * documentation of the function that calls the callback.
10970 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10971 * my_callback should return 1 if it wants BN_is_prime_ex()
10972 * to continue, or 0 to stop.
10977 * Change the ZLIB compression method to be stateful, and make it
10978 available to TLS with the number defined in
10979 draft-ietf-tls-compression-04.txt.
10983 * Add the ASN.1 structures and functions for CertificatePair, which
10984 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10986 CertificatePair ::= SEQUENCE {
10987 forward [0] Certificate OPTIONAL,
10988 reverse [1] Certificate OPTIONAL,
10989 -- at least one of the pair shall be present -- }
10991 Also implement the PEM functions to read and write certificate
10992 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10994 This needed to be defined, mostly for the sake of the LDAP
10995 attribute crossCertificatePair, but may prove useful elsewhere as
11000 * Make it possible to inhibit symlinking of shared libraries in
11001 Makefile.shared, for Cygwin's sake.
11005 * Extend the BIGNUM API by creating a function
11006 void BN_set_negative(BIGNUM *a, int neg);
11007 and a macro that behave like
11008 int BN_is_negative(const BIGNUM *a);
11010 to avoid the need to access 'a->neg' directly in applications.
11014 * Implement fast modular reduction for pseudo-Mersenne primes
11015 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11016 EC_GROUP_new_curve_GFp() will now automatically use this
11019 *Nils Larsch <nla@trustcenter.de>*
11021 * Add new lock type (CRYPTO_LOCK_BN).
11025 * Change the ENGINE framework to automatically load engines
11026 dynamically from specific directories unless they could be
11027 found to already be built in or loaded. Move all the
11028 current engines except for the cryptodev one to a new
11029 directory engines/.
11030 The engines in engines/ are built as shared libraries if
11031 the "shared" options was given to ./Configure or ./config.
11032 Otherwise, they are inserted in libcrypto.a.
11033 /usr/local/ssl/engines is the default directory for dynamic
11034 engines, but that can be overridden at configure time through
11035 the usual use of --prefix and/or --openssldir, and at run
11036 time with the environment variable OPENSSL_ENGINES.
11038 *Geoff Thorpe and Richard Levitte*
11040 * Add Makefile.shared, a helper makefile to build shared
11041 libraries. Adapt Makefile.org.
11045 * Add version info to Win32 DLLs.
11047 *Peter 'Luna' Runestig" <peter@runestig.com>*
11049 * Add new 'medium level' PKCS#12 API. Certificates and keys
11050 can be added using this API to created arbitrary PKCS#12
11051 files while avoiding the low-level API.
11053 New options to PKCS12_create(), key or cert can be NULL and
11054 will then be omitted from the output file. The encryption
11055 algorithm NIDs can be set to -1 for no encryption, the mac
11056 iteration count can be set to 0 to omit the mac.
11058 Enhance pkcs12 utility by making the -nokeys and -nocerts
11059 options work when creating a PKCS#12 file. New option -nomac
11060 to omit the mac, NONE can be set for an encryption algorithm.
11061 New code is modified to use the enhanced PKCS12_create()
11062 instead of the low-level API.
11066 * Extend ASN1 encoder to support indefinite length constructed
11067 encoding. This can output sequences tags and octet strings in
11068 this form. Modify pk7_asn1.c to support indefinite length
11069 encoding. This is experimental and needs additional code to
11070 be useful, such as an ASN1 bio and some enhanced streaming
11073 Extend template encode functionality so that tagging is passed
11074 down to the template encoder.
11078 * Let 'openssl req' fail if an argument to '-newkey' is not
11079 recognized instead of using RSA as a default.
11083 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11084 As these are not official, they are not included in "ALL";
11085 the "ECCdraft" ciphersuite group alias can be used to select them.
11087 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11089 * Add ECDH engine support.
11091 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11093 * Add ECDH in new directory crypto/ecdh/.
11095 *Douglas Stebila (Sun Microsystems Laboratories)*
11097 * Let BN_rand_range() abort with an error after 100 iterations
11098 without success (which indicates a broken PRNG).
11102 * Change BN_mod_sqrt() so that it verifies that the input value
11103 is really the square of the return value. (Previously,
11104 BN_mod_sqrt would show GIGO behaviour.)
11108 * Add named elliptic curves over binary fields from X9.62, SECG,
11109 and WAP/WTLS; add OIDs that were still missing.
11111 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11113 * Extend the EC library for elliptic curves over binary fields
11114 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11117 EC_GF2m_simple_method
11121 EC_GROUP_new_curve_GF2m
11122 EC_GROUP_set_curve_GF2m
11123 EC_GROUP_get_curve_GF2m
11124 EC_POINT_set_affine_coordinates_GF2m
11125 EC_POINT_get_affine_coordinates_GF2m
11126 EC_POINT_set_compressed_coordinates_GF2m
11128 Point compression for binary fields is disabled by default for
11129 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11132 As binary polynomials are represented as BIGNUMs, various members
11133 of the EC_GROUP and EC_POINT data structures can be shared
11134 between the implementations for prime fields and binary fields;
11135 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11136 are essentially identical to their `..._GFp` counterparts.
11137 (For simplicity, the `..._GFp` prefix has been dropped from
11138 various internal method names.)
11140 An internal 'field_div' method (similar to 'field_mul' and
11141 'field_sqr') has been added; this is used only for binary fields.
11143 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11145 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11146 through methods ('mul', 'precompute_mult').
11148 The generic implementations (now internally called 'ec_wNAF_mul'
11149 and 'ec_wNAF_precomputed_mult') remain the default if these
11150 methods are undefined.
11152 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11154 * New function EC_GROUP_get_degree, which is defined through
11155 EC_METHOD. For curves over prime fields, this returns the bit
11156 length of the modulus.
11158 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11160 * New functions EC_GROUP_dup, EC_POINT_dup.
11161 (These simply call ..._new and ..._copy).
11163 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11165 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11166 Polynomials are represented as BIGNUMs (where the sign bit is not
11167 used) in the following functions [macros]:
11170 BN_GF2m_sub [= BN_GF2m_add]
11171 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11172 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11173 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11175 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11176 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11177 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11178 BN_GF2m_cmp [= BN_ucmp]
11180 (Note that only the 'mod' functions are actually for fields GF(2^m).
11181 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11183 For some functions, an the irreducible polynomial defining a
11184 field can be given as an 'unsigned int[]' with strictly
11185 decreasing elements giving the indices of those bits that are set;
11186 i.e., p[] represents the polynomial
11187 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11189 p[0] > p[1] > ... > p[k] = 0.
11190 This applies to the following functions:
11193 BN_GF2m_mod_mul_arr
11194 BN_GF2m_mod_sqr_arr
11195 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11196 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11197 BN_GF2m_mod_exp_arr
11198 BN_GF2m_mod_sqrt_arr
11199 BN_GF2m_mod_solve_quad_arr
11203 Conversion can be performed by the following functions:
11208 bntest.c has additional tests for binary polynomial arithmetic.
11210 Two implementations for BN_GF2m_mod_div() are available.
11211 The default algorithm simply uses BN_GF2m_mod_inv() and
11212 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11213 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11214 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11216 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11218 * Add new error code 'ERR_R_DISABLED' that can be used when some
11219 functionality is disabled at compile-time.
11221 *Douglas Stebila <douglas.stebila@sun.com>*
11223 * Change default behaviour of 'openssl asn1parse' so that more
11224 information is visible when viewing, e.g., a certificate:
11226 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11227 mode the content of non-printable OCTET STRINGs is output in a
11228 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11229 avoid the appearance of a printable string.
11231 *Nils Larsch <nla@trustcenter.de>*
11233 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11235 EC_GROUP_set_asn1_flag()
11236 EC_GROUP_get_asn1_flag()
11237 EC_GROUP_set_point_conversion_form()
11238 EC_GROUP_get_point_conversion_form()
11239 These control ASN1 encoding details:
11240 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11241 has been set to OPENSSL_EC_NAMED_CURVE.
11242 - Points are encoded in uncompressed form by default; options for
11243 asn1_for are as for point2oct, namely
11244 POINT_CONVERSION_COMPRESSED
11245 POINT_CONVERSION_UNCOMPRESSED
11246 POINT_CONVERSION_HYBRID
11248 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11250 EC_GROUP_set_seed()
11251 EC_GROUP_get0_seed()
11252 EC_GROUP_get_seed_len()
11253 This is used only for ASN1 purposes (so far).
11255 *Nils Larsch <nla@trustcenter.de>*
11257 * Add 'field_type' member to EC_METHOD, which holds the NID
11258 of the appropriate field type OID. The new function
11259 EC_METHOD_get_field_type() returns this value.
11261 *Nils Larsch <nla@trustcenter.de>*
11264 EC_POINT_point2bn()
11265 EC_POINT_bn2point()
11266 EC_POINT_point2hex()
11267 EC_POINT_hex2point()
11268 providing useful interfaces to EC_POINT_point2oct() and
11269 EC_POINT_oct2point().
11271 *Nils Larsch <nla@trustcenter.de>*
11273 * Change internals of the EC library so that the functions
11274 EC_GROUP_set_generator()
11275 EC_GROUP_get_generator()
11276 EC_GROUP_get_order()
11277 EC_GROUP_get_cofactor()
11278 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11279 to methods, which would lead to unnecessary code duplication when
11280 adding different types of curves.
11282 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11284 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11285 arithmetic, and such that modified wNAFs are generated
11286 (which avoid length expansion in many cases).
11290 * Add a function EC_GROUP_check_discriminant() (defined via
11291 EC_METHOD) that verifies that the curve discriminant is non-zero.
11293 Add a function EC_GROUP_check() that makes some sanity tests
11294 on a EC_GROUP, its generator and order. This includes
11295 EC_GROUP_check_discriminant().
11297 *Nils Larsch <nla@trustcenter.de>*
11299 * Add ECDSA in new directory crypto/ecdsa/.
11301 Add applications 'openssl ecparam' and 'openssl ecdsa'
11302 (these are based on 'openssl dsaparam' and 'openssl dsa').
11304 ECDSA support is also included in various other files across the
11305 library. Most notably,
11306 - 'openssl req' now has a '-newkey ecdsa:file' option;
11307 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11308 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11309 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11310 them suitable for ECDSA where domain parameters must be
11311 extracted before the specific public key;
11312 - ECDSA engine support has been added.
11314 *Nils Larsch <nla@trustcenter.de>*
11316 * Include some named elliptic curves, and add OIDs from X9.62,
11317 SECG, and WAP/WTLS. Each curve can be obtained from the new
11319 EC_GROUP_new_by_curve_name(),
11320 and the list of available named curves can be obtained with
11321 EC_get_builtin_curves().
11322 Also add a 'curve_name' member to EC_GROUP objects, which can be
11324 EC_GROUP_set_curve_name()
11325 EC_GROUP_get_curve_name()
11327 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11329 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11330 was actually never needed) and in BN_mul(). The removal in BN_mul()
11331 required a small change in bn_mul_part_recursive() and the addition
11332 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11333 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11334 bn_sub_words() and bn_add_words() except they take arrays with
11339 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11341 * Cleanse PEM buffers before freeing them since they may contain
11344 *Benjamin Bennett <ben@psc.edu>*
11346 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11347 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11348 authentication-only ciphersuites.
11352 * Since AES128 and AES256 share a single mask bit in the logic of
11353 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11354 kludge to work properly if AES128 is available and AES256 isn't.
11358 * Expand security boundary to match 1.1.1 module.
11362 * Remove redundant features: hash file source, editing of test vectors
11363 modify fipsld to use external fips_premain.c signature.
11367 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11368 run algorithm test programs.
11372 * Make algorithm test programs more tolerant of whitespace.
11376 * Have SSL/TLS server implementation tolerate "mismatched" record
11377 protocol version while receiving ClientHello even if the
11378 ClientHello is fragmented. (The server can't insist on the
11379 particular protocol version it has chosen before the ServerHello
11380 message has informed the client about his choice.)
11384 * Load error codes if they are not already present instead of using a
11385 static variable. This allows them to be cleanly unloaded and reloaded.
11389 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11391 * Introduce limits to prevent malicious keys being able to
11392 cause a denial of service. ([CVE-2006-2940])
11394 *Steve Henson, Bodo Moeller*
11396 * Fix ASN.1 parsing of certain invalid structures that can result
11397 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11399 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11400 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11402 * Fix SSL client code which could crash if connecting to a
11403 malicious SSLv2 server. ([CVE-2006-4343])
11405 *Tavis Ormandy and Will Drewry, Google Security Team*
11407 * Change ciphersuite string processing so that an explicit
11408 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11409 will no longer include "AES128-SHA"), and any other similar
11410 ciphersuite (same bitmap) from *other* protocol versions (so that
11411 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11412 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11413 changes from 0.9.8b and 0.9.8d.
11417 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11419 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11420 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11422 * Change the Unix randomness entropy gathering to use poll() when
11423 possible instead of select(), since the latter has some
11424 undesirable limitations.
11426 *Darryl Miles via Richard Levitte and Bodo Moeller*
11428 * Disable rogue ciphersuites:
11430 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11431 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11432 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11434 The latter two were purportedly from
11435 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11438 Also deactivate the remaining ciphersuites from
11439 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11440 unofficial, and the ID has long expired.
11444 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11445 dual-core machines) and other potential thread-safety issues.
11449 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11451 * Adapt fipsld and the build system to link against the validated FIPS
11452 module in FIPS mode.
11456 * Fixes for VC++ 2005 build under Windows.
11460 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11461 from a Windows bash shell such as MSYS. It is autodetected from the
11462 "config" script when run from a VC++ environment. Modify standard VC++
11463 build to use fipscanister.o from the GNU make build.
11467 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11469 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11470 The value now differs depending on if you build for FIPS or not.
11471 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11472 safely run with a non-FIPSed libcrypto, as it may crash because of
11473 the difference induced by this change.
11477 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11479 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11480 (part of SSL_OP_ALL). This option used to disable the
11481 countermeasure against man-in-the-middle protocol-version
11482 rollback in the SSL 2.0 server implementation, which is a bad
11483 idea. ([CVE-2005-2969])
11485 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11486 for Information Security, National Institute of Advanced Industrial
11487 Science and Technology [AIST, Japan)]*
11489 * Minimal support for X9.31 signatures and PSS padding modes. This is
11490 mainly for FIPS compliance and not fully integrated at this stage.
11494 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11495 the exponentiation using a fixed-length exponent. (Otherwise,
11496 the information leaked through timing could expose the secret key
11497 after many signatures; cf. Bleichenbacher's attack on DSA with
11502 * Make a new fixed-window mod_exp implementation the default for
11503 RSA, DSA, and DH private-key operations so that the sequence of
11504 squares and multiplies and the memory access pattern are
11505 independent of the particular secret key. This will mitigate
11506 cache-timing and potential related attacks.
11508 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11509 and this is automatically used by BN_mod_exp_mont() if the new flag
11510 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11511 will use this BN flag for private exponents unless the flag
11512 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11513 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11515 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11517 * Change the client implementation for SSLv23_method() and
11518 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11519 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11520 (Previously, the SSL 2.0 backwards compatible Client Hello
11521 message format would be used even with SSL_OP_NO_SSLv2.)
11525 * Add support for smime-type MIME parameter in S/MIME messages which some
11530 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11531 a threadsafe manner. Modify rsa code to use new function and add calls
11532 to dsa and dh code (which had race conditions before).
11536 * Include the fixed error library code in the C error file definitions
11537 instead of fixing them up at runtime. This keeps the error code
11538 structures constant.
11542 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11544 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11547 * Fixes for newer kerberos headers. NB: the casts are needed because
11548 the 'length' field is signed on one version and unsigned on another
11549 with no (?) obvious way to tell the difference, without these VC++
11550 complains. Also the "definition" of FAR (blank) is no longer included
11551 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11552 some needed definitions.
11556 * Undo Cygwin change.
11560 * Added support for proxy certificates according to RFC 3820.
11561 Because they may be a security thread to unaware applications,
11562 they must be explicitly allowed in run-time. See
11563 docs/HOWTO/proxy_certificates.txt for further information.
11567 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11569 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11570 server and client random values. Previously
11571 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11572 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11574 This change has negligible security impact because:
11576 1. Server and client random values still have 24 bytes of pseudo random
11579 2. Server and client random values are sent in the clear in the initial
11582 3. The master secret is derived using the premaster secret (48 bytes in
11583 size for static RSA ciphersuites) as well as client server and random
11586 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11589 *Stephen Henson, reported by UK NISCC*
11591 * Use Windows randomness collection on Cygwin.
11595 * Fix hang in EGD/PRNGD query when communication socket is closed
11596 prematurely by EGD/PRNGD.
11598 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11600 * Prompt for pass phrases when appropriate for PKCS12 input format.
11604 * Back-port of selected performance improvements from development
11605 branch, as well as improved support for PowerPC platforms.
11609 * Add lots of checks for memory allocation failure, error codes to indicate
11610 failure and freeing up memory if a failure occurs.
11612 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11614 * Add new -passin argument to dgst.
11618 * Perform some character comparisons of different types in X509_NAME_cmp:
11619 this is needed for some certificates that re-encode DNs into UTF8Strings
11620 (in violation of RFC3280) and can't or won't issue name rollover
11625 * Make an explicit check during certificate validation to see that
11626 the CA setting in each certificate on the chain is correct. As a
11627 side effect always do the following basic checks on extensions,
11628 not just when there's an associated purpose to the check:
11630 - if there is an unhandled critical extension (unless the user
11631 has chosen to ignore this fault)
11632 - if the path length has been exceeded (if one is set at all)
11633 - that certain extensions fit the associated purpose (if one has
11638 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11640 * Avoid a race condition when CRLs are checked in a multi threaded
11641 environment. This would happen due to the reordering of the revoked
11642 entries during signature checking and serial number lookup. Now the
11643 encoding is cached and the serial number sort performed under a lock.
11644 Add new STACK function sk_is_sorted().
11648 * Add Delta CRL to the extension code.
11652 * Various fixes to s3_pkt.c so alerts are sent properly.
11654 *David Holmes <d.holmes@f5.com>*
11656 * Reduce the chances of duplicate issuer name and serial numbers (in
11657 violation of RFC3280) using the OpenSSL certificate creation utilities.
11658 This is done by creating a random 64 bit value for the initial serial
11659 number when a serial number file is created or when a self signed
11660 certificate is created using 'openssl req -x509'. The initial serial
11661 number file is created using 'openssl x509 -next_serial' in CA.pl
11662 rather than being initialized to 1.
11666 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11668 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11669 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11671 *Joe Orton, Steve Henson*
11673 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11676 *Joe Orton, Steve Henson*
11678 * Make it possible to have multiple active certificates with the same
11679 subject in the CA index file. This is done only if the keyword
11680 'unique_subject' is set to 'no' in the main CA section (default
11681 if 'CA_default') of the configuration file. The value is saved
11682 with the database itself in a separate index attribute file,
11683 named like the index file with '.attr' appended to the name.
11687 * X509 verify fixes. Disable broken certificate workarounds when
11688 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11689 keyUsage extension present. Don't accept CRLs with unhandled critical
11690 extensions: since verify currently doesn't process CRL extensions this
11691 rejects a CRL with *any* critical extensions. Add new verify error codes
11696 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11697 A clarification of RFC2560 will require the use of OCTET STRINGs and
11698 some implementations cannot handle the current raw format. Since OpenSSL
11699 copies and compares OCSP nonces as opaque blobs without any attempt at
11700 parsing them this should not create any compatibility issues.
11704 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11705 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11706 this HMAC (and other) operations are several times slower than OpenSSL
11711 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11713 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11715 * Use the correct content when signing type "other".
11719 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11721 * Fix various bugs revealed by running the NISCC test suite:
11723 Stop out of bounds reads in the ASN1 code when presented with
11724 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11726 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11728 If verify callback ignores invalid public key errors don't try to check
11729 certificate signature with the NULL public key.
11733 * New -ignore_err option in ocsp application to stop the server
11734 exiting on the first error in a request.
11738 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11739 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11744 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11745 extra data after the compression methods not only for TLS 1.0
11746 but also for SSL 3.0 (as required by the specification).
11748 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11750 * Change X509_certificate_type() to mark the key as exported/exportable
11751 when it's 512 *bits* long, not 512 bytes.
11755 * Change AES_cbc_encrypt() so it outputs exact multiple of
11756 blocks during encryption.
11760 * Various fixes to base64 BIO and non blocking I/O. On write
11761 flushes were not handled properly if the BIO retried. On read
11762 data was not being buffered properly and had various logic bugs.
11763 This also affects blocking I/O when the data being decoded is a
11768 * Various S/MIME bugfixes and compatibility changes:
11769 output correct application/pkcs7 MIME type if
11770 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11771 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11772 of files as .eml work). Correctly handle very long lines in MIME
11777 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11779 * Countermeasure against the Klima-Pokorny-Rosa extension of
11780 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11781 a protocol version number mismatch like a decryption error
11782 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11786 * Turn on RSA blinding by default in the default implementation
11787 to avoid a timing attack. Applications that don't want it can call
11788 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11789 They would be ill-advised to do so in most cases.
11791 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11793 * Change RSA blinding code so that it works when the PRNG is not
11794 seeded (in this case, the secret RSA exponent is abused as
11795 an unpredictable seed -- if it is not unpredictable, there
11796 is no point in blinding anyway). Make RSA blinding thread-safe
11797 by remembering the creator's thread ID in rsa->blinding and
11798 having all other threads use local one-time blinding factors
11799 (this requires more computation than sharing rsa->blinding, but
11800 avoids excessive locking; and if an RSA object is not shared
11801 between threads, blinding will still be very fast).
11805 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11806 ENGINE as defaults for all supported algorithms irrespective of
11807 the 'flags' parameter. 'flags' is now honoured, so applications
11808 should make sure they are passing it correctly.
11812 * Target "mingw" now allows native Windows code to be generated in
11813 the Cygwin environment as well as with the MinGW compiler.
11817 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11819 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11820 via timing by performing a MAC computation even if incorrect
11821 block cipher padding has been found. This is a countermeasure
11822 against active attacks where the attacker has to distinguish
11823 between bad padding and a MAC verification error. ([CVE-2003-0078])
11825 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11826 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11827 Martin Vuagnoux (EPFL, Ilion)*
11829 * Make the no-err option work as intended. The intention with no-err
11830 is not to have the whole error stack handling routines removed from
11831 libcrypto, it's only intended to remove all the function name and
11832 reason texts, thereby removing some of the footprint that may not
11833 be interesting if those errors aren't displayed anyway.
11835 NOTE: it's still possible for any application or module to have its
11836 own set of error texts inserted. The routines are there, just not
11837 used by default when no-err is given.
11841 * Add support for FreeBSD on IA64.
11843 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11845 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11846 Kerberos function mit_des_cbc_cksum(). Before this change,
11847 the value returned by DES_cbc_cksum() was like the one from
11848 mit_des_cbc_cksum(), except the bytes were swapped.
11850 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11852 * Allow an application to disable the automatic SSL chain building.
11853 Before this a rather primitive chain build was always performed in
11854 ssl3_output_cert_chain(): an application had no way to send the
11855 correct chain if the automatic operation produced an incorrect result.
11857 Now the chain builder is disabled if either:
11859 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11861 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11863 The reasoning behind this is that an application would not want the
11864 auto chain building to take place if extra chain certificates are
11865 present and it might also want a means of sending no additional
11866 certificates (for example the chain has two certificates and the
11871 * Add the possibility to build without the ENGINE framework.
11873 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11875 * Under Win32 gmtime() can return NULL: check return value in
11876 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11880 * DSA routines: under certain error conditions uninitialized BN objects
11881 could be freed. Solution: make sure initialization is performed early
11882 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11883 Nils Larsch <nla@trustcenter.de> via PR#459)
11887 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11888 checked on reconnect on the client side, therefore session resumption
11889 could still fail with a "ssl session id is different" error. This
11890 behaviour is masked when SSL_OP_ALL is used due to
11891 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11892 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11893 followup to PR #377.
11897 * IA-32 assembler support enhancements: unified ELF targets, support
11898 for SCO/Caldera platforms, fix for Cygwin shared build.
11902 * Add support for FreeBSD on sparc64. As a consequence, support for
11903 FreeBSD on non-x86 processors is separate from x86 processors on
11904 the config script, much like the NetBSD support.
11906 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11908 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11910 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11913 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11914 code (06) was taken as the first octet of the session ID and the last
11915 octet was ignored consequently. As a result SSLv2 client side session
11916 caching could not have worked due to the session ID mismatch between
11918 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11923 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11924 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11929 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11930 seems that in spite of existing for more than a year, many application
11931 author have done nothing to provide the necessary callbacks, which
11932 means that this particular engine will not work properly anywhere.
11933 This is a very unfortunate situation which forces us, in the name
11934 of usability, to give the hw_ncipher.c a static lock, which is part
11936 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11937 appear in 0.9.8 or later. We EXPECT application authors to have
11938 dealt properly with this when 0.9.8 is released (unless we actually
11939 make such changes in the libcrypto locking code that changes will
11940 have to be made anyway).
11944 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11945 octets have been read, EOF or an error occurs. Without this change
11946 some truncated ASN1 structures will not produce an error.
11950 * Disable Heimdal support, since it hasn't been fully implemented.
11951 Still give the possibility to force the use of Heimdal, but with
11952 warnings and a request that patches get sent to openssl-dev.
11956 * Add the VC-CE target, introduce the WINCE sysname, and add
11957 INSTALL.WCE and appropriate conditionals to make it build.
11959 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11961 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11962 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11963 edit numbers of the version.
11965 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11967 * Introduce safe string copy and catenation functions
11968 (BUF_strlcpy() and BUF_strlcat()).
11970 *Ben Laurie (CHATS) and Richard Levitte*
11972 * Avoid using fixed-size buffers for one-line DNs.
11974 *Ben Laurie (CHATS)*
11976 * Add BUF_MEM_grow_clean() to avoid information leakage when
11977 resizing buffers containing secrets, and use where appropriate.
11979 *Ben Laurie (CHATS)*
11981 * Avoid using fixed size buffers for configuration file location.
11983 *Ben Laurie (CHATS)*
11985 * Avoid filename truncation for various CA files.
11987 *Ben Laurie (CHATS)*
11989 * Use sizeof in preference to magic numbers.
11991 *Ben Laurie (CHATS)*
11993 * Avoid filename truncation in cert requests.
11995 *Ben Laurie (CHATS)*
11997 * Add assertions to check for (supposedly impossible) buffer
12000 *Ben Laurie (CHATS)*
12002 * Don't cache truncated DNS entries in the local cache (this could
12003 potentially lead to a spoofing attack).
12005 *Ben Laurie (CHATS)*
12007 * Fix various buffers to be large enough for hex/decimal
12008 representations in a platform independent manner.
12010 *Ben Laurie (CHATS)*
12012 * Add CRYPTO_realloc_clean() to avoid information leakage when
12013 resizing buffers containing secrets, and use where appropriate.
12015 *Ben Laurie (CHATS)*
12017 * Add BIO_indent() to avoid much slightly worrying code to do
12020 *Ben Laurie (CHATS)*
12022 * Convert sprintf()/BIO_puts() to BIO_printf().
12024 *Ben Laurie (CHATS)*
12026 * buffer_gets() could terminate with the buffer only half
12029 *Ben Laurie (CHATS)*
12031 * Add assertions to prevent user-supplied crypto functions from
12032 overflowing internal buffers by having large block sizes, etc.
12034 *Ben Laurie (CHATS)*
12036 * New OPENSSL_assert() macro (similar to assert(), but enabled
12039 *Ben Laurie (CHATS)*
12041 * Eliminate unused copy of key in RC4.
12043 *Ben Laurie (CHATS)*
12045 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12047 *Ben Laurie (CHATS)*
12049 * Fix off-by-one error in EGD path.
12051 *Ben Laurie (CHATS)*
12053 * If RANDFILE path is too long, ignore instead of truncating.
12055 *Ben Laurie (CHATS)*
12057 * Eliminate unused and incorrectly sized X.509 structure
12060 *Ben Laurie (CHATS)*
12062 * Eliminate unused and dangerous function knumber().
12064 *Ben Laurie (CHATS)*
12066 * Eliminate unused and dangerous structure, KSSL_ERR.
12068 *Ben Laurie (CHATS)*
12070 * Protect against overlong session ID context length in an encoded
12071 session object. Since these are local, this does not appear to be
12074 *Ben Laurie (CHATS)*
12076 * Change from security patch (see 0.9.6e below) that did not affect
12077 the 0.9.6 release series:
12079 Remote buffer overflow in SSL3 protocol - an attacker could
12080 supply an oversized master key in Kerberos-enabled versions.
12083 *Ben Laurie (CHATS)*
12085 * Change the SSL kerb5 codes to match RFC 2712.
12089 * Make -nameopt work fully for req and add -reqopt switch.
12091 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12093 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12095 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12097 * Make sure tests can be performed even if the corresponding algorithms
12098 have been removed entirely. This was also the last step to make
12099 OpenSSL compilable with DJGPP under all reasonable conditions.
12101 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12103 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12104 to allow version independent disabling of normally unselected ciphers,
12105 which may be activated as a side-effect of selecting a single cipher.
12107 (E.g., cipher list string "RSA" enables ciphersuites that are left
12108 out of "ALL" because they do not provide symmetric encryption.
12109 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12111 *Lutz Jaenicke, Bodo Moeller*
12113 * Add appropriate support for separate platform-dependent build
12114 directories. The recommended way to make a platform-dependent
12115 build directory is the following (tested on Linux), maybe with
12118 # Place yourself outside of the OpenSSL source tree. In
12119 # this example, the environment variable OPENSSL_SOURCE
12120 # is assumed to contain the absolute OpenSSL source directory.
12121 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12122 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12123 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12124 mkdir -p `dirname $F`
12125 ln -s $OPENSSL_SOURCE/$F $F
12128 To be absolutely sure not to disturb the source tree, a "make clean"
12129 is a good thing. If it isn't successful, don't worry about it,
12130 it probably means the source directory is very clean.
12134 * Make sure any ENGINE control commands make local copies of string
12135 pointers passed to them whenever necessary. Otherwise it is possible
12136 the caller may have overwritten (or deallocated) the original string
12137 data when a later ENGINE operation tries to use the stored values.
12139 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12141 * Improve diagnostics in file reading and command-line digests.
12143 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12145 * Add AES modes CFB and OFB to the object database. Correct an
12146 error in AES-CFB decryption.
12150 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12151 allows existing EVP_CIPHER_CTX structures to be reused after
12152 calling `EVP_*Final()`. This behaviour is used by encryption
12153 BIOs and some applications. This has the side effect that
12154 applications must explicitly clean up cipher contexts with
12155 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12159 * Check the values of dna and dnb in bn_mul_recursive before calling
12160 bn_mul_comba (a non zero value means the a or b arrays do not contain
12161 n2 elements) and fallback to bn_mul_normal if either is not zero.
12165 * Fix escaping of non-ASCII characters when using the -subj option
12166 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12170 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12171 form for "surname", serialNumber has no short form.
12172 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12173 therefore remove "mail" short name for "internet 7".
12174 The OID for unique identifiers in X509 certificates is
12175 x500UniqueIdentifier, not uniqueIdentifier.
12176 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12180 * Add an "init" command to the ENGINE config module and auto initialize
12181 ENGINEs. Without any "init" command the ENGINE will be initialized
12182 after all ctrl commands have been executed on it. If init=1 the
12183 ENGINE is initialized at that point (ctrls before that point are run
12184 on the uninitialized ENGINE and after on the initialized one). If
12185 init=0 then the ENGINE will not be initialized at all.
12189 * Fix the 'app_verify_callback' interface so that the user-defined
12190 argument is actually passed to the callback: In the
12191 SSL_CTX_set_cert_verify_callback() prototype, the callback
12192 declaration has been changed from
12195 int (*cb)(X509_STORE_CTX *,void *);
12196 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12197 i=s->ctx->app_verify_callback(&ctx)
12198 has been changed into
12199 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12201 To update applications using SSL_CTX_set_cert_verify_callback(),
12202 a dummy argument can be added to their callback functions.
12204 *D. K. Smetters <smetters@parc.xerox.com>*
12206 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12208 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12210 * Add and OPENSSL_LOAD_CONF define which will cause
12211 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12212 This allows older applications to transparently support certain
12213 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12214 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12215 load the config file and OPENSSL_add_all_algorithms_conf() which will
12216 always load it have also been added.
12220 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12221 Adjust NIDs and EVP layer.
12223 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12225 * Config modules support in openssl utility.
12227 Most commands now load modules from the config file,
12228 though in a few (such as version) this isn't done
12229 because it couldn't be used for anything.
12231 In the case of ca and req the config file used is
12232 the same as the utility itself: that is the -config
12233 command line option can be used to specify an
12238 * Move default behaviour from OPENSSL_config(). If appname is NULL
12239 use "openssl_conf" if filename is NULL use default openssl config file.
12243 * Add an argument to OPENSSL_config() to allow the use of an alternative
12244 config section name. Add a new flag to tolerate a missing config file
12245 and move code to CONF_modules_load_file().
12249 * Support for crypto accelerator cards from Accelerated Encryption
12250 Processing, www.aep.ie. (Use engine 'aep')
12251 The support was copied from 0.9.6c [engine] and adapted/corrected
12252 to work with the new engine framework.
12254 *AEP Inc. and Richard Levitte*
12256 * Support for SureWare crypto accelerator cards from Baltimore
12257 Technologies. (Use engine 'sureware')
12258 The support was copied from 0.9.6c [engine] and adapted
12259 to work with the new engine framework.
12263 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12264 make the newer ENGINE framework commands for the CHIL engine work.
12266 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12268 * Make it possible to produce shared libraries on ReliantUNIX.
12270 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12272 * Add the configuration target debug-linux-ppro.
12273 Make 'openssl rsa' use the general key loading routines
12274 implemented in `apps.c`, and make those routines able to
12275 handle the key format FORMAT_NETSCAPE and the variant
12278 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12280 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12282 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12284 * Add -keyform to rsautl, and document -engine.
12286 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12288 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12289 BIO_R_NO_SUCH_FILE error code rather than the generic
12290 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12294 * Add new functions
12295 ERR_peek_last_error
12296 ERR_peek_last_error_line
12297 ERR_peek_last_error_line_data.
12298 These are similar to
12300 ERR_peek_error_line
12301 ERR_peek_error_line_data,
12302 but report on the latest error recorded rather than the first one
12303 still in the error queue.
12305 *Ben Laurie, Bodo Moeller*
12307 * default_algorithms option in ENGINE config module. This allows things
12309 default_algorithms = ALL
12310 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12314 * Preliminary ENGINE config module.
12318 * New experimental application configuration code.
12322 * Change the AES code to follow the same name structure as all other
12323 symmetric ciphers, and behave the same way. Move everything to
12324 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12326 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12328 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12330 *Ben Laurie and Theo de Raadt*
12332 * Add option to output public keys in req command.
12334 *Massimiliano Pala madwolf@openca.org*
12336 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12337 (up to about 10% better than before for P-192 and P-224).
12341 * New functions/macros
12343 SSL_CTX_set_msg_callback(ctx, cb)
12344 SSL_CTX_set_msg_callback_arg(ctx, arg)
12345 SSL_set_msg_callback(ssl, cb)
12346 SSL_set_msg_callback_arg(ssl, arg)
12348 to request calling a callback function
12350 void cb(int write_p, int version, int content_type,
12351 const void *buf, size_t len, SSL *ssl, void *arg)
12353 whenever a protocol message has been completely received
12354 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12355 protocol version according to which the SSL library interprets
12356 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12357 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12358 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12359 specification (change_cipher_spec(20), alert(21), handshake(22)).
12360 'buf' and 'len' point to the actual message, 'ssl' to the
12361 SSL object, and 'arg' is the application-defined value set by
12362 SSL[_CTX]_set_msg_callback_arg().
12364 'openssl s_client' and 'openssl s_server' have new '-msg' options
12365 to enable a callback that displays all protocol messages.
12369 * Change the shared library support so shared libraries are built as
12370 soon as the corresponding static library is finished, and thereby get
12371 openssl and the test programs linked against the shared library.
12372 This still only happens when the keyword "shard" has been given to
12373 the configuration scripts.
12375 NOTE: shared library support is still an experimental thing, and
12376 backward binary compatibility is still not guaranteed.
12378 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12380 * Add support for Subject Information Access extension.
12382 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12384 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12385 additional bytes when new memory had to be allocated, not just
12386 when reusing an existing buffer.
12390 * New command line and configuration option 'utf8' for the req command.
12391 This allows field values to be specified as UTF8 strings.
12395 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12396 runs for the former and machine-readable output for the latter.
12400 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12401 of the e-mail address in the DN (i.e., it will go into a certificate
12402 extension only). The new configuration file option 'email_in_dn = no'
12403 has the same effect.
12405 *Massimiliano Pala madwolf@openca.org*
12407 * Change all functions with names starting with `des_` to be starting
12408 with `DES_` instead. Add wrappers that are compatible with libdes,
12409 but are named `_ossl_old_des_*`. Finally, add macros that map the
12410 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12411 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12412 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12415 Since we provide two compatibility mappings, the user needs to
12416 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12417 compatibility is desired. The default (i.e., when that macro
12418 isn't defined) is OpenSSL 0.9.6c compatibility.
12420 There are also macros that enable and disable the support of old
12421 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12422 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12423 are defined, the default will apply: to support the old des routines.
12425 In either case, one must include openssl/des.h to get the correct
12426 definitions. Do not try to just include openssl/des_old.h, that
12429 NOTE: This is a major break of an old API into a new one. Software
12430 authors are encouraged to switch to the `DES_` style functions. Some
12431 time in the future, des_old.h and the libdes compatibility functions
12432 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12433 default), and then completely removed.
12437 * Test for certificates which contain unsupported critical extensions.
12438 If such a certificate is found during a verify operation it is
12439 rejected by default: this behaviour can be overridden by either
12440 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12441 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12442 X509_supported_extension() has also been added which returns 1 if a
12443 particular extension is supported.
12447 * Modify the behaviour of EVP cipher functions in similar way to digests
12448 to retain compatibility with existing code.
12452 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12453 compatibility with existing code. In particular the 'ctx' parameter does
12454 not have to be to be initialized before the call to EVP_DigestInit() and
12455 it is tidied up after a call to EVP_DigestFinal(). New function
12456 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12457 EVP_MD_CTX_copy() changed to not require the destination to be
12458 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12459 requires the destination to be valid.
12461 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12462 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12466 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12467 so that complete 'Handshake' protocol structures are kept in memory
12468 instead of overwriting 'msg_type' and 'length' with 'body' data.
12472 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12474 *Massimo Santin via Richard Levitte*
12476 * Major restructuring to the underlying ENGINE code. This includes
12477 reduction of linker bloat, separation of pure "ENGINE" manipulation
12478 (initialisation, etc) from functionality dealing with implementations
12479 of specific crypto interfaces. This change also introduces integrated
12480 support for symmetric ciphers and digest implementations - so ENGINEs
12481 can now accelerate these by providing EVP_CIPHER and EVP_MD
12482 implementations of their own. This is detailed in
12483 [crypto/engine/README.md](crypto/engine/README.md)
12484 as it couldn't be adequately described here. However, there are a few
12485 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12486 were changed in the original introduction of ENGINE code have now
12487 reverted back - the hooking from this code to ENGINE is now a good
12488 deal more passive and at run-time, operations deal directly with
12489 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12490 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12491 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12492 they were not being used by the framework as there is no concept of a
12493 BIGNUM_METHOD and they could not be generalised to the new
12494 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12495 ENGINE_cpy() has been removed as it cannot be consistently defined in
12500 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12504 * Change mkdef.pl to sort symbols that get the same entry number,
12505 and make sure the automatically generated functions `ERR_load_*`
12506 become part of libeay.num as well.
12510 * New function SSL_renegotiate_pending(). This returns true once
12511 renegotiation has been requested (either SSL_renegotiate() call
12512 or HelloRequest/ClientHello received from the peer) and becomes
12513 false once a handshake has been completed.
12514 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12515 sends a HelloRequest, but does not ensure that a handshake takes
12516 place. SSL_renegotiate_pending() is useful for checking if the
12517 client has followed the request.)
12521 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12522 By default, clients may request session resumption even during
12523 renegotiation (if session ID contexts permit); with this option,
12524 session resumption is possible only in the first handshake.
12526 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12527 more bits available for options that should not be part of
12528 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12532 * Add some demos for certificate and certificate request creation.
12536 * Make maximum certificate chain size accepted from the peer application
12537 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12538 "Douglas E. Engert" <deengert@anl.gov>.
12542 * Add support for shared libraries for Unixware-7
12543 (Boyd Lynn Gerber <gerberb@zenez.com>).
12547 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12548 be done prior to destruction. Use this to unload error strings from
12549 ENGINEs that load their own error strings. NB: This adds two new API
12550 functions to "get" and "set" this destroy handler in an ENGINE.
12554 * Alter all existing ENGINE implementations (except "openssl" and
12555 "openbsd") to dynamically instantiate their own error strings. This
12556 makes them more flexible to be built both as statically-linked ENGINEs
12557 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12558 Also, add stub code to each that makes building them as self-contained
12559 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12563 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12564 implementations into applications that are completely implemented in
12565 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12566 commands that can be used to configure what shared-library to load and
12567 to control aspects of the way it is handled. Also, made an update to
12568 the [README-Engine.md](README-Engine.md) file
12569 that brings its information up-to-date and
12570 provides some information and instructions on the "dynamic" ENGINE
12571 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12575 * Make it possible to unload ranges of ERR strings with a new
12576 "ERR_unload_strings" function.
12580 * Add a copy() function to EVP_MD.
12584 * Make EVP_MD routines take a context pointer instead of just the
12585 md_data void pointer.
12589 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12590 that the digest can only process a single chunk of data
12591 (typically because it is provided by a piece of
12592 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12593 is only going to provide a single chunk of data, and hence the
12594 framework needn't accumulate the data for oneshot drivers.
12598 * As with "ERR", make it possible to replace the underlying "ex_data"
12599 functions. This change also alters the storage and management of global
12600 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12601 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12602 index counters. The API functions that use this state have been changed
12603 to take a "class_index" rather than pointers to the class's local STACK
12604 and counter, and there is now an API function to dynamically create new
12605 classes. This centralisation allows us to (a) plug a lot of the
12606 thread-safety problems that existed, and (b) makes it possible to clean
12607 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12608 such data would previously have always leaked in application code and
12609 workarounds were in place to make the memory debugging turn a blind eye
12610 to it. Application code that doesn't use this new function will still
12611 leak as before, but their memory debugging output will announce it now
12612 rather than letting it slide.
12614 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12615 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12616 has a return value to indicate success or failure.
12620 * Make it possible to replace the underlying "ERR" functions such that the
12621 global state (2 LHASH tables and 2 locks) is only used by the "default"
12622 implementation. This change also adds two functions to "get" and "set"
12623 the implementation prior to it being automatically set the first time
12624 any other ERR function takes place. Ie. an application can call "get",
12625 pass the return value to a module it has just loaded, and that module
12626 can call its own "set" function using that value. This means the
12627 module's "ERR" operations will use (and modify) the error state in the
12628 application and not in its own statically linked copy of OpenSSL code.
12632 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12633 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12634 the operation, and provides a more encapsulated way for external code
12635 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12636 to use these functions rather than manually incrementing the counts.
12638 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12642 * Add EVP test program.
12646 * Add symmetric cipher support to ENGINE. Expect the API to change!
12650 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12651 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12652 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12653 These allow a CRL to be built without having to access X509_CRL fields
12654 directly. Modify 'ca' application to use new functions.
12658 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12659 bug workarounds. Rollback attack detection is a security feature.
12660 The problem will only arise on OpenSSL servers when TLSv1 is not
12661 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12662 Software authors not wanting to support TLSv1 will have special reasons
12663 for their choice and can explicitly enable this option.
12665 *Bodo Moeller, Lutz Jaenicke*
12667 * Rationalise EVP so it can be extended: don't include a union of
12668 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12669 (similar to those existing for EVP_CIPHER_CTX).
12674 EVP_MD_CTX_init(&md); /* new function call */
12675 EVP_DigestInit(&md, EVP_sha1());
12676 EVP_DigestUpdate(&md, in, len);
12677 EVP_DigestFinal(&md, out, NULL);
12678 EVP_MD_CTX_cleanup(&md); /* new function call */
12682 * Make DES key schedule conform to the usual scheme, as well as
12683 correcting its structure. This means that calls to DES functions
12684 now have to pass a pointer to a des_key_schedule instead of a
12685 plain des_key_schedule (which was actually always a pointer
12688 des_key_schedule ks;
12690 des_set_key_checked(..., &ks);
12691 des_ncbc_encrypt(..., &ks, ...);
12693 (Note that a later change renames 'des_...' into 'DES_...'.)
12697 * Initial reduction of linker bloat: the use of some functions, such as
12698 PEM causes large amounts of unused functions to be linked in due to
12699 poor organisation. For example pem_all.c contains every PEM function
12700 which has a knock on effect of linking in large amounts of (unused)
12701 ASN1 code. Grouping together similar functions and splitting unrelated
12702 functions prevents this.
12706 * Cleanup of EVP macros.
12710 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12711 correct `_ecb suffix`.
12715 * Add initial OCSP responder support to ocsp application. The
12716 revocation information is handled using the text based index
12717 use by the ca application. The responder can either handle
12718 requests generated internally, supplied in files (for example
12719 via a CGI script) or using an internal minimal server.
12723 * Add configuration choices to get zlib compression for TLS.
12727 * Changes to Kerberos SSL for RFC 2712 compliance:
12728 1. Implemented real KerberosWrapper, instead of just using
12729 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12730 2. Implemented optional authenticator field of KerberosWrapper.
12732 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12733 and authenticator structs; see crypto/krb5/.
12735 Generalized Kerberos calls to support multiple Kerberos libraries.
12736 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12737 via Richard Levitte*
12739 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12740 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12741 values for each of the key sizes rather than having just
12742 parameters (and 'speed' generating keys each time).
12746 * Speed up EVP routines.
12749 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12750 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12751 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12752 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12754 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12755 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12756 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12759 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12761 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12765 * Added the OS2-EMX target.
12767 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12769 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12770 New functions to support `NCONF `routines in extension code.
12771 New function `CONF_set_nconf()`
12772 to allow functions which take an `NCONF` to also handle the old `LHASH`
12773 structure: this means that the old `CONF` compatible routines can be
12774 retained (in particular w.rt. extensions) without having to duplicate the
12775 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12779 * Enhance the general user interface with mechanisms for inner control
12780 and with possibilities to have yes/no kind of prompts.
12784 * Change all calls to low-level digest routines in the library and
12785 applications to use EVP. Add missing calls to HMAC_cleanup() and
12786 don't assume HMAC_CTX can be copied using memcpy().
12788 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12790 * Add the possibility to control engines through control names but with
12791 arbitrary arguments instead of just a string.
12792 Change the key loaders to take a UI_METHOD instead of a callback
12793 function pointer. NOTE: this breaks binary compatibility with earlier
12794 versions of OpenSSL [engine].
12795 Adapt the nCipher code for these new conditions and add a card insertion
12800 * Enhance the general user interface with mechanisms to better support
12801 dialog box interfaces, application-defined prompts, the possibility
12802 to use defaults (for example default passwords from somewhere else)
12803 and interrupts/cancellations.
12807 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12808 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12812 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12813 tidy up some unnecessarily weird code in 'sk_new()').
12815 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12817 * Change the key loading routines for ENGINEs to use the same kind
12818 callback (pem_password_cb) as all other routines that need this
12823 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12824 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12825 than this minimum value is recommended.
12829 * New random seeder for OpenVMS, using the system process statistics
12830 that are easily reachable.
12834 * Windows apparently can't transparently handle global
12835 variables defined in DLLs. Initialisations such as:
12837 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12839 won't compile. This is used by the any applications that need to
12840 declare their own ASN1 modules. This was fixed by adding the option
12841 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12842 needed for static libraries under Win32.
12846 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12847 setting of purpose and trust fields. New X509_STORE trust and
12848 purpose functions and tidy up setting in other SSL functions.
12852 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12853 structure. These are inherited by X509_STORE_CTX when it is
12854 initialised. This allows various defaults to be set in the
12855 X509_STORE structure (such as flags for CRL checking and custom
12856 purpose or trust settings) for functions which only use X509_STORE_CTX
12857 internally such as S/MIME.
12859 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12860 trust settings if they are not set in X509_STORE. This allows X509_STORE
12861 purposes and trust (in S/MIME for example) to override any set by default.
12863 Add command line options for CRL checking to smime, s_client and s_server
12868 * Initial CRL based revocation checking. If the CRL checking flag(s)
12869 are set then the CRL is looked up in the X509_STORE structure and
12870 its validity and signature checked, then if the certificate is found
12871 in the CRL the verify fails with a revoked error.
12873 Various new CRL related callbacks added to X509_STORE_CTX structure.
12875 Command line options added to 'verify' application to support this.
12877 This needs some additional work, such as being able to handle multiple
12878 CRLs with different times, extension based lookup (rather than just
12879 by subject name) and ultimately more complete V2 CRL extension
12884 * Add a general user interface API (crypto/ui/). This is designed
12885 to replace things like des_read_password and friends (backward
12886 compatibility functions using this new API are provided).
12887 The purpose is to remove prompting functions from the DES code
12888 section as well as provide for prompting through dialog boxes in
12889 a window system and the like.
12893 * Add "ex_data" support to ENGINE so implementations can add state at a
12894 per-structure level rather than having to store it globally.
12898 * Make it possible for ENGINE structures to be copied when retrieved by
12899 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12900 This causes the "original" ENGINE structure to act like a template,
12901 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12902 operational state can be localised to each ENGINE structure, despite the
12903 fact they all share the same "methods". New ENGINE structures returned in
12904 this case have no functional references and the return value is the single
12905 structural reference. This matches the single structural reference returned
12906 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12911 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12912 needs to match any other type at all we need to manually clear the
12917 * Changes to the "openssl engine" utility to include;
12918 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12919 about an ENGINE's available control commands.
12920 - executing control commands from command line arguments using the
12921 '-pre' and '-post' switches. '-post' is only used if '-t' is
12922 specified and the ENGINE is successfully initialised. The syntax for
12923 the individual commands are colon-separated, for example;
12924 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12928 * New dynamic control command support for ENGINEs. ENGINEs can now
12929 declare their own commands (numbers), names (strings), descriptions,
12930 and input types for run-time discovery by calling applications. A
12931 subset of these commands are implicitly classed as "executable"
12932 depending on their input type, and only these can be invoked through
12933 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12934 can be based on user input, config files, etc). The distinction is
12935 that "executable" commands cannot return anything other than a boolean
12936 result and can only support numeric or string input, whereas some
12937 discoverable commands may only be for direct use through
12938 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12939 pointers, or other custom uses. The "executable" commands are to
12940 support parameterisations of ENGINE behaviour that can be
12941 unambiguously defined by ENGINEs and used consistently across any
12942 OpenSSL-based application. Commands have been added to all the
12943 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12944 control over shared-library paths without source code alterations.
12948 * Changed all ENGINE implementations to dynamically allocate their
12949 ENGINEs rather than declaring them statically. Apart from this being
12950 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12951 this also allows the implementations to compile without using the
12952 internal engine_int.h header.
12956 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12957 'const' value. Any code that should be able to modify a RAND_METHOD
12958 should already have non-const pointers to it (ie. they should only
12959 modify their own ones).
12963 * Made a variety of little tweaks to the ENGINE code.
12964 - "atalla" and "ubsec" string definitions were moved from header files
12965 to C code. "nuron" string definitions were placed in variables
12966 rather than hard-coded - allowing parameterisation of these values
12967 later on via ctrl() commands.
12968 - Removed unused "#if 0"'d code.
12969 - Fixed engine list iteration code so it uses ENGINE_free() to release
12970 structural references.
12971 - Constified the RAND_METHOD element of ENGINE structures.
12972 - Constified various get/set functions as appropriate and added
12973 missing functions (including a catch-all ENGINE_cpy that duplicates
12974 all ENGINE values onto a new ENGINE except reference counts/state).
12975 - Removed NULL parameter checks in get/set functions. Setting a method
12976 or function to NULL is a way of cancelling out a previously set
12977 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12978 and doesn't justify the extra error symbols and code.
12979 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12980 flags from engine_int.h to engine.h.
12981 - Changed prototypes for ENGINE handler functions (init(), finish(),
12982 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12986 * Implement binary inversion algorithm for BN_mod_inverse in addition
12987 to the algorithm using long division. The binary algorithm can be
12988 used only if the modulus is odd. On 32-bit systems, it is faster
12989 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12990 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12991 up to 450 bits. In 64-bit environments, the binary algorithm
12992 appears to be advantageous for much longer moduli; here we use it
12993 for moduli up to 2048 bits.
12997 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12998 could not support the combine flag in choice fields.
13002 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13003 extensions from a certificate request to the certificate.
13007 * Allow multiple 'certopt' and 'nameopt' options to be separated
13008 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13009 file: this allows the display of the certificate about to be
13010 signed to be customised, to allow certain fields to be included
13011 or excluded and extension details. The old system didn't display
13012 multicharacter strings properly, omitted fields not in the policy
13013 and couldn't display additional details such as extensions.
13017 * Function EC_POINTs_mul for multiple scalar multiplication
13018 of an arbitrary number of elliptic curve points
13019 \sum scalars[i]*points[i],
13020 optionally including the generator defined for the EC_GROUP:
13021 scalar*generator + \sum scalars[i]*points[i].
13023 EC_POINT_mul is a simple wrapper function for the typical case
13024 that the point list has just one item (besides the optional
13029 * First EC_METHODs for curves over GF(p):
13031 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13032 operations and provides various method functions that can also
13033 operate with faster implementations of modular arithmetic.
13035 EC_GFp_mont_method() reuses most functions that are part of
13036 EC_GFp_simple_method, but uses Montgomery arithmetic.
13038 *Bodo Moeller; point addition and point doubling
13039 implementation directly derived from source code provided by
13040 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13042 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13043 crypto/ec/ec_lib.c):
13045 Curves are EC_GROUP objects (with an optional group generator)
13046 based on EC_METHODs that are built into the library.
13048 Points are EC_POINT objects based on EC_GROUP objects.
13050 Most of the framework would be able to handle curves over arbitrary
13051 finite fields, but as there are no obvious types for fields other
13052 than GF(p), some functions are limited to that for now.
13056 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13057 that the file contains a complete HTTP response.
13061 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13062 change the def and num file printf format specifier from "%-40sXXX"
13063 to "%-39s XXX". The latter will always guarantee a space after the
13064 field while the former will cause them to run together if the field
13065 is 40 of more characters long.
13069 * Constify the cipher and digest 'method' functions and structures
13070 and modify related functions to take constant EVP_MD and EVP_CIPHER
13075 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13076 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13080 * Modify `EVP_Digest*()` routines so they now return values. Although the
13081 internal software routines can never fail additional hardware versions
13086 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13088 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13089 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13092 ERR_R_NESTED_ASN1_ERROR
13094 ERR_R_MISSING_ASN1_EOS
13095 were 4 .. 9, conflicting with
13096 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13098 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13099 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13101 Add new error code 'ERR_R_INTERNAL_ERROR'.
13105 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13110 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13111 sets the subject name for a new request or supersedes the
13112 subject name in a given request. Formats that can be parsed are
13113 'CN=Some Name, OU=myOU, C=IT'
13115 'CN=Some Name/OU=myOU/C=IT'.
13117 Add options '-batch' and '-verbose' to 'openssl req'.
13119 *Massimiliano Pala <madwolf@hackmasters.net>*
13121 * Introduce the possibility to access global variables through
13122 functions on platform were that's the best way to handle exporting
13123 global variables in shared libraries. To enable this functionality,
13124 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13125 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13126 is normally done by Configure or something similar).
13128 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13129 in the source file (foo.c) like this:
13131 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13132 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13134 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13135 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13137 OPENSSL_DECLARE_GLOBAL(int,foo);
13138 #define foo OPENSSL_GLOBAL_REF(foo)
13139 OPENSSL_DECLARE_GLOBAL(double,bar);
13140 #define bar OPENSSL_GLOBAL_REF(bar)
13142 The #defines are very important, and therefore so is including the
13143 header file everywhere where the defined globals are used.
13145 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13146 of ASN.1 items, but that structure is a bit different.
13148 The largest change is in util/mkdef.pl which has been enhanced with
13149 better and easier to understand logic to choose which symbols should
13150 go into the Windows .def files as well as a number of fixes and code
13151 cleanup (among others, algorithm keywords are now sorted
13152 lexicographically to avoid constant rewrites).
13156 * In BN_div() keep a copy of the sign of 'num' before writing the
13157 result to 'rm' because if rm==num the value will be overwritten
13158 and produce the wrong result if 'num' is negative: this caused
13159 problems with BN_mod() and BN_nnmod().
13163 * Function OCSP_request_verify(). This checks the signature on an
13164 OCSP request and verifies the signer certificate. The signer
13165 certificate is just checked for a generic purpose and OCSP request
13170 * Add OCSP_check_validity() function to check the validity of OCSP
13171 responses. OCSP responses are prepared in real time and may only
13172 be a few seconds old. Simply checking that the current time lies
13173 between thisUpdate and nextUpdate max reject otherwise valid responses
13174 caused by either OCSP responder or client clock inaccuracy. Instead
13175 we allow thisUpdate and nextUpdate to fall within a certain period of
13176 the current time. The age of the response can also optionally be
13177 checked. Two new options -validity_period and -status_age added to
13182 * If signature or public key algorithm is unrecognized print out its
13183 OID rather that just UNKNOWN.
13187 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13188 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13189 ID to be generated from the issuer certificate alone which can then be
13190 passed to OCSP_id_issuer_cmp().
13194 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13195 ASN1 modules to export functions returning ASN1_ITEM pointers
13196 instead of the ASN1_ITEM structures themselves. This adds several
13197 new macros which allow the underlying ASN1 function/structure to
13198 be accessed transparently. As a result code should not use ASN1_ITEM
13199 references directly (such as &X509_it) but instead use the relevant
13200 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13201 use of the new ASN1 code on platforms where exporting structures
13202 is problematical (for example in shared libraries) but exporting
13203 functions returning pointers to structures is not.
13207 * Add support for overriding the generation of SSL/TLS session IDs.
13208 These callbacks can be registered either in an SSL_CTX or per SSL.
13209 The purpose of this is to allow applications to control, if they wish,
13210 the arbitrary values chosen for use as session IDs, particularly as it
13211 can be useful for session caching in multiple-server environments. A
13212 command-line switch for testing this (and any client code that wishes
13213 to use such a feature) has been added to "s_server".
13215 *Geoff Thorpe, Lutz Jaenicke*
13217 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13218 of the form `#if defined(...) || defined(...) || ...` and
13219 `#if !defined(...) && !defined(...) && ...`. This also avoids
13220 the growing number of special cases it was previously handling.
13224 * Make all configuration macros available for application by making
13225 sure they are available in opensslconf.h, by giving them names starting
13226 with `OPENSSL_` to avoid conflicts with other packages and by making
13227 sure e_os2.h will cover all platform-specific cases together with
13229 Additionally, it is now possible to define configuration/platform-
13230 specific names (called "system identities"). In the C code, these
13231 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13232 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13233 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13238 * New option -set_serial to 'req' and 'x509' this allows the serial
13239 number to use to be specified on the command line. Previously self
13240 signed certificates were hard coded with serial number 0 and the
13241 CA options of 'x509' had to use a serial number in a file which was
13246 * New options to 'ca' utility to support V2 CRL entry extensions.
13247 Currently CRL reason, invalidity date and hold instruction are
13248 supported. Add new CRL extensions to V3 code and some new objects.
13252 * New function EVP_CIPHER_CTX_set_padding() this is used to
13253 disable standard block padding (aka PKCS#5 padding) in the EVP
13254 API, which was previously mandatory. This means that the data is
13255 not padded in any way and so the total length much be a multiple
13256 of the block size, otherwise an error occurs.
13260 * Initial (incomplete) OCSP SSL support.
13264 * New function OCSP_parse_url(). This splits up a URL into its host,
13265 port and path components: primarily to parse OCSP URLs. New -url
13266 option to ocsp utility.
13270 * New nonce behavior. The return value of OCSP_check_nonce() now
13271 reflects the various checks performed. Applications can decide
13272 whether to tolerate certain situations such as an absent nonce
13273 in a response when one was present in a request: the ocsp application
13274 just prints out a warning. New function OCSP_add1_basic_nonce()
13275 this is to allow responders to include a nonce in a response even if
13276 the request is nonce-less.
13280 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13281 skipped when using openssl x509 multiple times on a single input file,
13282 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13286 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13287 set string type: to handle setting ASN1_TIME structures. Fix ca
13288 utility to correctly initialize revocation date of CRLs.
13292 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13293 the clients preferred ciphersuites and rather use its own preferences.
13294 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13295 Internet Explorer by ensuring unchanged hash method during stepup.
13296 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13300 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13301 to aes and add a new 'exist' option to print out symbols that don't
13306 * Additional options to ocsp utility to allow flags to be set and
13307 additional certificates supplied.
13311 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13312 OCSP client a number of certificate to only verify the response
13317 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13318 handle the new API. Currently only ECB, CBC modes supported. Add new
13321 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13322 Encryption Standard (AES) Ciphersuites for Transport Layer
13323 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13324 not enabled by default and were not part of the "ALL" ciphersuite
13325 alias because they were not yet official; they could be
13326 explicitly requested by specifying the "AESdraft" ciphersuite
13327 group alias. In the final release of OpenSSL 0.9.7, the group
13328 alias is called "AES" and is part of "ALL".)
13330 *Ben Laurie, Steve Henson, Bodo Moeller*
13332 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13333 request to response.
13337 * Functions for OCSP responders. OCSP_request_onereq_count(),
13338 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13339 extract information from a certificate request. OCSP_response_create()
13340 creates a response and optionally adds a basic response structure.
13341 OCSP_basic_add1_status() adds a complete single response to a basic
13342 response and returns the OCSP_SINGLERESP structure just added (to allow
13343 extensions to be included for example). OCSP_basic_add1_cert() adds a
13344 certificate to a basic response and OCSP_basic_sign() signs a basic
13345 response with various flags. New helper functions ASN1_TIME_check()
13346 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13347 (converts ASN1_TIME to GeneralizedTime).
13351 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13352 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13353 structure from a certificate. X509_pubkey_digest() digests the public_key
13354 contents: this is used in various key identifiers.
13358 * Make sk_sort() tolerate a NULL argument.
13360 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13362 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13363 passed by the function are trusted implicitly. If any of them signed the
13364 response then it is assumed to be valid and is not verified.
13368 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13369 to data. This was previously part of the PKCS7 ASN1 code. This
13370 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13371 *Steve Henson, reported by Kenneth R. Robinette
13372 <support@securenetterm.com>*
13374 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13375 routines: without these tracing memory leaks is very painful.
13376 Fix leaks in PKCS12 and PKCS7 routines.
13380 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13381 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13382 effectively meant GeneralizedTime would never be used. Now it
13383 is initialised to -1 but X509_time_adj() now has to check the value
13384 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13385 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13386 *Steve Henson, reported by Kenneth R. Robinette
13387 <support@securenetterm.com>*
13389 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13390 result in a zero length in the ASN1_INTEGER structure which was
13391 not consistent with the structure when d2i_ASN1_INTEGER() was used
13392 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13393 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13394 where it did not print out a minus for negative ASN1_INTEGER.
13398 * Add summary printout to ocsp utility. The various functions which
13399 convert status values to strings have been renamed to:
13400 OCSP_response_status_str(), OCSP_cert_status_str() and
13401 OCSP_crl_reason_str() and are no longer static. New options
13402 to verify nonce values and to disable verification. OCSP response
13403 printout format cleaned up.
13407 * Add additional OCSP certificate checks. These are those specified
13408 in RFC2560. This consists of two separate checks: the CA of the
13409 certificate being checked must either be the OCSP signer certificate
13410 or the issuer of the OCSP signer certificate. In the latter case the
13411 OCSP signer certificate must contain the OCSP signing extended key
13412 usage. This check is performed by attempting to match the OCSP
13413 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13414 in the OCSP_CERTID structures of the response.
13418 * Initial OCSP certificate verification added to OCSP_basic_verify()
13419 and related routines. This uses the standard OpenSSL certificate
13420 verify routines to perform initial checks (just CA validity) and
13421 to obtain the certificate chain. Then additional checks will be
13422 performed on the chain. Currently the root CA is checked to see
13423 if it is explicitly trusted for OCSP signing. This is used to set
13424 a root CA as a global signing root: that is any certificate that
13425 chains to that CA is an acceptable OCSP signing certificate.
13429 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13430 extensions from a separate configuration file.
13431 As when reading extensions from the main configuration file,
13432 the '-extensions ...' option may be used for specifying the
13435 *Massimiliano Pala <madwolf@comune.modena.it>*
13437 * New OCSP utility. Allows OCSP requests to be generated or
13438 read. The request can be sent to a responder and the output
13439 parsed, outputted or printed in text form. Not complete yet:
13440 still needs to check the OCSP response validity.
13444 * New subcommands for 'openssl ca':
13445 `openssl ca -status <serial>` prints the status of the cert with
13446 the given serial number (according to the index file).
13447 `openssl ca -updatedb` updates the expiry status of certificates
13450 *Massimiliano Pala <madwolf@comune.modena.it>*
13452 * New '-newreq-nodes' command option to CA.pl. This is like
13453 '-newreq', but calls 'openssl req' with the '-nodes' option
13454 so that the resulting key is not encrypted.
13456 *Damien Miller <djm@mindrot.org>*
13458 * New configuration for the GNU Hurd.
13460 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13462 * Initial code to implement OCSP basic response verify. This
13463 is currently incomplete. Currently just finds the signer's
13464 certificate and verifies the signature on the response.
13468 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13469 value of OPENSSLDIR. This is available via the new '-d' option
13470 to 'openssl version', and is also included in 'openssl version -a'.
13474 * Allowing defining memory allocation callbacks that will be given
13475 file name and line number information in additional arguments
13476 (a `const char*` and an int). The basic functionality remains, as
13477 well as the original possibility to just replace malloc(),
13478 realloc() and free() by functions that do not know about these
13479 additional arguments. To register and find out the current
13480 settings for extended allocation functions, the following
13481 functions are provided:
13483 CRYPTO_set_mem_ex_functions
13484 CRYPTO_set_locked_mem_ex_functions
13485 CRYPTO_get_mem_ex_functions
13486 CRYPTO_get_locked_mem_ex_functions
13488 These work the same way as CRYPTO_set_mem_functions and friends.
13489 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13490 extended allocation function is enabled.
13491 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13492 a conventional allocation function is enabled.
13494 *Richard Levitte, Bodo Moeller*
13496 * Finish off removing the remaining LHASH function pointer casts.
13497 There should no longer be any prototype-casting required when using
13498 the LHASH abstraction, and any casts that remain are "bugs". See
13499 the callback types and macros at the head of lhash.h for details
13500 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13504 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13505 If /dev/[u]random devices are not available or do not return enough
13506 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13508 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13509 /etc/entropy will be queried once each in this sequence, querying stops
13510 when enough entropy was collected without querying more sockets.
13514 * Change the Unix RAND_poll() variant to be able to poll several
13515 random devices, as specified by DEVRANDOM, until a sufficient amount
13516 of data has been collected. We spend at most 10 ms on each file
13517 (select timeout) and read in non-blocking mode. DEVRANDOM now
13518 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13519 (previously it was just the string "/dev/urandom"), so on typical
13520 platforms the 10 ms delay will never occur.
13521 Also separate out the Unix variant to its own file, rand_unix.c.
13522 For VMS, there's a currently-empty rand_vms.c.
13526 * Move OCSP client related routines to ocsp_cl.c. These
13527 provide utility functions which an application needing
13528 to issue a request to an OCSP responder and analyse the
13529 response will typically need: as opposed to those which an
13530 OCSP responder itself would need which will be added later.
13532 OCSP_request_sign() signs an OCSP request with an API similar
13533 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13534 response. OCSP_response_get1_basic() extracts basic response
13535 from response. OCSP_resp_find_status(): finds and extracts status
13536 information from an OCSP_CERTID structure (which will be created
13537 when the request structure is built). These are built from lower
13538 level functions which work on OCSP_SINGLERESP structures but
13539 won't normally be used unless the application wishes to examine
13540 extensions in the OCSP response for example.
13542 Replace nonce routines with a pair of functions.
13543 OCSP_request_add1_nonce() adds a nonce value and optionally
13544 generates a random value. OCSP_check_nonce() checks the
13545 validity of the nonce in an OCSP response.
13549 * Change function OCSP_request_add() to OCSP_request_add0_id().
13550 This doesn't copy the supplied OCSP_CERTID and avoids the
13551 need to free up the newly created id. Change return type
13552 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13553 This can then be used to add extensions to the request.
13554 Deleted OCSP_request_new(), since most of its functionality
13555 is now in OCSP_REQUEST_new() (and the case insensitive name
13556 clash) apart from the ability to set the request name which
13557 will be added elsewhere.
13561 * Update OCSP API. Remove obsolete extensions argument from
13562 various functions. Extensions are now handled using the new
13563 OCSP extension code. New simple OCSP HTTP function which
13564 can be used to send requests and parse the response.
13568 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13569 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13570 uses the special reorder version of SET OF to sort the attributes
13571 and reorder them to match the encoded order. This resolves a long
13572 standing problem: a verify on a PKCS7 structure just after signing
13573 it used to fail because the attribute order did not match the
13574 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13575 it uses the received order. This is necessary to tolerate some broken
13576 software that does not order SET OF. This is handled by encoding
13577 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13578 to produce the required SET OF.
13582 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13583 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13584 files to get correct declarations of the ASN.1 item variables.
13588 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13589 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13590 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13591 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13592 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13593 ASN1_ITEM and no wrapper functions.
13597 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13598 replace the old function pointer based I/O routines. Change most of
13599 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13603 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13604 lines, recognize more "algorithms" that can be deselected, and make
13605 it complain about algorithm deselection that isn't recognised.
13609 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13610 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13611 to use new functions. Add NO_ASN1_OLD which can be set to remove
13612 some old style ASN1 functions: this can be used to determine if old
13613 code will still work when these eventually go away.
13617 * New extension functions for OCSP structures, these follow the
13618 same conventions as certificates and CRLs.
13622 * New function X509V3_add1_i2d(). This automatically encodes and
13623 adds an extension. Its behaviour can be customised with various
13624 flags to append, replace or delete. Various wrappers added for
13625 certificates and CRLs.
13629 * Fix to avoid calling the underlying ASN1 print routine when
13630 an extension cannot be parsed. Correct a typo in the
13631 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13635 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13636 entries for variables.
13640 * Add functionality to `apps/openssl.c` for detecting locking
13641 problems: As the program is single-threaded, all we have
13642 to do is register a locking callback using an array for
13643 storing which locks are currently held by the program.
13647 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13648 SSL_get_ex_data_X509_STORE_idx(), which is used in
13649 ssl_verify_cert_chain() and thus can be called at any time
13650 during TLS/SSL handshakes so that thread-safety is essential.
13651 Unfortunately, the ex_data design is not at all suited
13652 for multi-threaded use, so it probably should be abolished.
13656 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13658 *Broadcom, tweaked and integrated by Geoff Thorpe*
13660 * Move common extension printing code to new function
13661 X509V3_print_extensions(). Reorganise OCSP print routines and
13662 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13666 * New function X509_signature_print() to remove duplication in some
13671 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13672 set (this was treated exactly the same as SET OF previously). This
13673 is used to reorder the STACK representing the structure to match the
13674 encoding. This will be used to get round a problem where a PKCS7
13675 structure which was signed could not be verified because the STACK
13676 order did not reflect the encoded order.
13680 * Reimplement the OCSP ASN1 module using the new code.
13684 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13685 for its ASN1 operations. The old style function pointers still exist
13686 for now but they will eventually go away.
13690 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13691 completely replaces the old ASN1 functionality with a table driven
13692 encoder and decoder which interprets an ASN1_ITEM structure describing
13693 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13694 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13695 has also been converted to the new form.
13699 * Change BN_mod_exp_recp so that negative moduli are tolerated
13700 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13701 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13702 for negative moduli.
13706 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13707 of not touching the result's sign bit.
13711 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13716 * Changed the LHASH code to use prototypes for callbacks, and created
13717 macros to declare and implement thin (optionally static) functions
13718 that provide type-safety and avoid function pointer casting for the
13719 type-specific callbacks.
13723 * Added Kerberos Cipher Suites to be used with TLS, as written in
13725 *Veers Staats <staatsvr@asc.hpc.mil>,
13726 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13728 * Reformat the FAQ so the different questions and answers can be divided
13729 in sections depending on the subject.
13733 * Have the zlib compression code load ZLIB.DLL dynamically under
13738 * New function BN_mod_sqrt for computing square roots modulo a prime
13739 (using the probabilistic Tonelli-Shanks algorithm unless
13740 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13741 be handled deterministically).
13743 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13745 * Make BN_mod_inverse faster by explicitly handling small quotients
13746 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13747 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13751 * New function BN_kronecker.
13755 * Fix BN_gcd so that it works on negative inputs; the result is
13756 positive unless both parameters are zero.
13757 Previously something reasonably close to an infinite loop was
13758 possible because numbers could be growing instead of shrinking
13759 in the implementation of Euclid's algorithm.
13763 * Fix BN_is_word() and BN_is_one() macros to take into account the
13764 sign of the number in question.
13766 Fix BN_is_word(a,w) to work correctly for w == 0.
13768 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13769 because its test if the absolute value of 'a' equals 'w'.
13770 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13771 it exists mostly for use in the implementations of BN_is_zero(),
13772 BN_is_one(), and BN_is_word().
13776 * New function BN_swap.
13780 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13781 the exponentiation functions are more likely to produce reasonable
13782 results on negative inputs.
13786 * Change BN_mod_mul so that the result is always non-negative.
13787 Previously, it could be negative if one of the factors was negative;
13788 I don't think anyone really wanted that behaviour.
13792 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13793 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13794 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13795 and add new functions:
13804 BN_mod_lshift1_quick
13806 BN_mod_lshift_quick
13808 These functions always generate non-negative results.
13810 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13811 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13813 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13814 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13815 be reduced modulo `m`.
13817 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13820 The following entry accidentally appeared in the CHANGES file
13821 distributed with OpenSSL 0.9.7. The modifications described in
13822 it do *not* apply to OpenSSL 0.9.7.
13824 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13825 was actually never needed) and in BN_mul(). The removal in BN_mul()
13826 required a small change in bn_mul_part_recursive() and the addition
13827 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13828 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13829 bn_sub_words() and bn_add_words() except they take arrays with
13835 * In 'openssl passwd', verify passwords read from the terminal
13836 unless the '-salt' option is used (which usually means that
13837 verification would just waste user's time since the resulting
13838 hash is going to be compared with some given password hash)
13839 or the new '-noverify' option is used.
13841 This is an incompatible change, but it does not affect
13842 non-interactive use of 'openssl passwd' (passwords on the command
13843 line, '-stdin' option, '-in ...' option) and thus should not
13844 cause any problems.
13848 * Remove all references to RSAref, since there's no more need for it.
13852 * Make DSO load along a path given through an environment variable
13853 (SHLIB_PATH) with shl_load().
13857 * Constify the ENGINE code as a result of BIGNUM constification.
13858 Also constify the RSA code and most things related to it. In a
13859 few places, most notable in the depth of the ASN.1 code, ugly
13860 casts back to non-const were required (to be solved at a later
13865 * Make it so the openssl application has all engines loaded by default.
13869 * Constify the BIGNUM routines a little more.
13873 * Add the following functions:
13875 ENGINE_load_cswift()
13877 ENGINE_load_atalla()
13878 ENGINE_load_nuron()
13879 ENGINE_load_builtin_engines()
13881 That way, an application can itself choose if external engines that
13882 are built-in in OpenSSL shall ever be used or not. The benefit is
13883 that applications won't have to be linked with libdl or other dso
13884 libraries unless it's really needed.
13886 Changed 'openssl engine' to load all engines on demand.
13887 Changed the engine header files to avoid the duplication of some
13888 declarations (they differed!).
13892 * 'openssl engine' can now list capabilities.
13896 * Better error reporting in 'openssl engine'.
13900 * Never call load_dh_param(NULL) in s_server.
13904 * Add engine application. It can currently list engines by name and
13905 identity, and test if they are actually available.
13909 * Improve RPM specification file by forcing symbolic linking and making
13910 sure the installed documentation is also owned by root.root.
13912 *Damien Miller <djm@mindrot.org>*
13914 * Give the OpenSSL applications more possibilities to make use of
13915 keys (public as well as private) handled by engines.
13919 * Add OCSP code that comes from CertCo.
13923 * Add VMS support for the Rijndael code.
13927 * Added untested support for Nuron crypto accelerator.
13931 * Add support for external cryptographic devices. This code was
13932 previously distributed separately as the "engine" branch.
13934 *Geoff Thorpe, Richard Levitte*
13936 * Rework the filename-translation in the DSO code. It is now possible to
13937 have far greater control over how a "name" is turned into a filename
13938 depending on the operating environment and any oddities about the
13939 different shared library filenames on each system.
13943 * Support threads on FreeBSD-elf in Configure.
13947 * Fix for SHA1 assembly problem with MASM: it produces
13948 warnings about corrupt line number information when assembling
13949 with debugging information. This is caused by the overlapping
13952 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13955 NCONF_get_number() has no error checking at all. As a replacement,
13956 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13957 promoted strongly. The old NCONF_get_number is kept around for
13958 binary backward compatibility.
13959 Make it possible for methods to load from something other than a BIO,
13960 by providing a function pointer that is given a name instead of a BIO.
13961 For example, this could be used to load configuration data from an
13966 * Fix for non blocking accept BIOs. Added new I/O special reason
13967 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13968 with non blocking I/O was not possible because no retry code was
13969 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13974 * Added the beginnings of Rijndael support.
13978 * Fix for bug in DirectoryString mask setting. Add support for
13979 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13980 to allow certificate printing to more controllable, additional
13981 'certopt' option to 'x509' to allow new printing options to be
13986 * Clean old EAY MD5 hack from e_os.h.
13990 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13992 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13993 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13995 *Joe Orton, Steve Henson*
13997 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13999 * Fix additional bug revealed by the NISCC test suite:
14001 Stop bug triggering large recursion when presented with
14002 certain ASN.1 tags ([CVE-2003-0851])
14006 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14008 * Fix various bugs revealed by running the NISCC test suite:
14010 Stop out of bounds reads in the ASN1 code when presented with
14011 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14013 If verify callback ignores invalid public key errors don't try to check
14014 certificate signature with the NULL public key.
14018 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14019 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14024 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14025 extra data after the compression methods not only for TLS 1.0
14026 but also for SSL 3.0 (as required by the specification).
14028 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14030 * Change X509_certificate_type() to mark the key as exported/exportable
14031 when it's 512 *bits* long, not 512 bytes.
14035 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14037 * Countermeasure against the Klima-Pokorny-Rosa extension of
14038 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14039 a protocol version number mismatch like a decryption error
14040 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14044 * Turn on RSA blinding by default in the default implementation
14045 to avoid a timing attack. Applications that don't want it can call
14046 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14047 They would be ill-advised to do so in most cases.
14049 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14051 * Change RSA blinding code so that it works when the PRNG is not
14052 seeded (in this case, the secret RSA exponent is abused as
14053 an unpredictable seed -- if it is not unpredictable, there
14054 is no point in blinding anyway). Make RSA blinding thread-safe
14055 by remembering the creator's thread ID in rsa->blinding and
14056 having all other threads use local one-time blinding factors
14057 (this requires more computation than sharing rsa->blinding, but
14058 avoids excessive locking; and if an RSA object is not shared
14059 between threads, blinding will still be very fast).
14063 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14065 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14066 via timing by performing a MAC computation even if incorrect
14067 block cipher padding has been found. This is a countermeasure
14068 against active attacks where the attacker has to distinguish
14069 between bad padding and a MAC verification error. ([CVE-2003-0078])
14071 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14072 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14073 Martin Vuagnoux (EPFL, Ilion)*
14075 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14077 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14078 memory from its contents. This is done with a counter that will
14079 place alternating values in each byte. This can be used to solve
14080 two issues: 1) the removal of calls to memset() by highly optimizing
14081 compilers, and 2) cleansing with other values than 0, since those can
14082 be read through on certain media, for example a swap space on disk.
14086 * Bugfix: client side session caching did not work with external caching,
14087 because the session->cipher setting was not restored when reloading
14088 from the external cache. This problem was masked, when
14089 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14090 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14094 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14095 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14097 *Zeev Lieber <zeev-l@yahoo.com>*
14099 * Undo an undocumented change introduced in 0.9.6e which caused
14100 repeated calls to OpenSSL_add_all_ciphers() and
14101 OpenSSL_add_all_digests() to be ignored, even after calling
14106 * Change the default configuration reader to deal with last line not
14107 being properly terminated.
14111 * Change X509_NAME_cmp() so it applies the special rules on handling
14112 DN values that are of type PrintableString, as well as RDNs of type
14113 emailAddress where the value has the type ia5String.
14115 *stefank@valicert.com via Richard Levitte*
14117 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14118 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14119 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14120 the bitwise-OR of the two for use by the majority of applications
14121 wanting this behaviour, and update the docs. The documented
14122 behaviour and actual behaviour were inconsistent and had been
14123 changing anyway, so this is more a bug-fix than a behavioural
14126 *Geoff Thorpe, diagnosed by Nadav Har'El*
14128 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14129 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14133 * Fix initialization code race conditions in
14134 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14135 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14136 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14137 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14138 ssl2_get_cipher_by_char(),
14139 ssl3_get_cipher_by_char().
14141 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14143 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14144 the cached sessions are flushed, as the remove_cb() might use ex_data
14145 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14146 (see [openssl.org #212]).
14148 *Geoff Thorpe, Lutz Jaenicke*
14150 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14151 length, instead of the encoding length to d2i_ASN1_OBJECT.
14155 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14157 * [In 0.9.6g-engine release:]
14158 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14160 *Lynn Gazis <lgazis@rainbow.com>*
14162 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14164 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14165 and get fix the header length calculation.
14166 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14167 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14169 * Use proper error handling instead of 'assertions' in buffer
14170 overflow checks added in 0.9.6e. This prevents DoS (the
14171 assertions could call abort()).
14173 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14175 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14177 * Add various sanity checks to asn1_get_length() to reject
14178 the ASN1 length bytes if they exceed sizeof(long), will appear
14179 negative or the content length exceeds the length of the
14182 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14184 * Fix cipher selection routines: ciphers without encryption had no flags
14185 for the cipher strength set and where therefore not handled correctly
14186 by the selection routines (PR #130).
14190 * Fix EVP_dsa_sha macro.
14195 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14196 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14197 that was added in OpenSSL 0.9.6d.
14199 As the countermeasure turned out to be incompatible with some
14200 broken SSL implementations, the new option is part of SSL_OP_ALL.
14201 SSL_OP_ALL is usually employed when compatibility with weird SSL
14202 implementations is desired (e.g. '-bugs' option to 's_client' and
14203 's_server'), so the new option is automatically set in many
14208 * Changes in security patch:
14210 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14211 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14212 Air Force Materiel Command, USAF, under agreement number
14215 * Add various sanity checks to asn1_get_length() to reject
14216 the ASN1 length bytes if they exceed sizeof(long), will appear
14217 negative or the content length exceeds the length of the
14218 supplied buffer. ([CVE-2002-0659])
14220 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14222 * Assertions for various potential buffer overflows, not known to
14223 happen in practice.
14225 *Ben Laurie (CHATS)*
14227 * Various temporary buffers to hold ASCII versions of integers were
14228 too small for 64 bit platforms. ([CVE-2002-0655])
14229 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14231 * Remote buffer overflow in SSL3 protocol - an attacker could
14232 supply an oversized session ID to a client. ([CVE-2002-0656])
14234 *Ben Laurie (CHATS)*
14236 * Remote buffer overflow in SSL2 protocol - an attacker could
14237 supply an oversized client master key. ([CVE-2002-0656])
14239 *Ben Laurie (CHATS)*
14241 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14243 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14244 encoded as NULL) with id-dsa-with-sha1.
14246 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14248 * Check various `X509_...()` return values in `apps/req.c`.
14250 *Nils Larsch <nla@trustcenter.de>*
14252 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14253 an end-of-file condition would erroneously be flagged, when the CRLF
14254 was just at the end of a processed block. The bug was discovered when
14255 processing data through a buffering memory BIO handing the data to a
14256 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14257 <ptsekov@syntrex.com> and Nedelcho Stanev.
14261 * Implement a countermeasure against a vulnerability recently found
14262 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14263 before application data chunks to avoid the use of known IVs
14264 with data potentially chosen by the attacker.
14268 * Fix length checks in ssl3_get_client_hello().
14272 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14273 to prevent ssl3_read_internal() from incorrectly assuming that
14274 ssl3_read_bytes() found application data while handshake
14275 processing was enabled when in fact s->s3->in_read_app_data was
14276 merely automatically cleared during the initial handshake.
14278 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14280 * Fix object definitions for Private and Enterprise: they were not
14281 recognized in their shortname (=lowercase) representation. Extend
14282 obj_dat.pl to issue an error when using undefined keywords instead
14283 of silently ignoring the problem (Svenning Sorensen
14284 <sss@sss.dnsalias.net>).
14288 * Fix DH_generate_parameters() so that it works for 'non-standard'
14289 generators, i.e. generators other than 2 and 5. (Previously, the
14290 code did not properly initialise the 'add' and 'rem' values to
14291 BN_generate_prime().)
14293 In the new general case, we do not insist that 'generator' is
14294 actually a primitive root: This requirement is rather pointless;
14295 a generator of the order-q subgroup is just as good, if not
14300 * Map new X509 verification errors to alerts. Discovered and submitted by
14301 Tom Wu <tom@arcot.com>.
14305 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14306 returning non-zero before the data has been completely received
14307 when using non-blocking I/O.
14309 *Bodo Moeller; problem pointed out by John Hughes*
14311 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14313 *Ben Laurie, Lutz Jaenicke*
14315 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14316 Yoram Zahavi <YoramZ@gilian.com>).
14320 * Add information about CygWin 1.3 and on, and preserve proper
14321 configuration for the versions before that.
14323 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14325 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14326 check whether we deal with a copy of a session and do not delete from
14327 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14328 <izhar@checkpoint.com>.
14332 * Do not store session data into the internal session cache, if it
14333 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14334 flag is set). Proposed by Aslam <aslam@funk.com>.
14338 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14343 * [In 0.9.6d-engine release:]
14344 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14346 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14348 * Add the configuration target linux-s390x.
14350 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14352 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14353 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14354 variable as an indication that a ClientHello message has been
14355 received. As the flag value will be lost between multiple
14356 invocations of ssl3_accept when using non-blocking I/O, the
14357 function may not be aware that a handshake has actually taken
14358 place, thus preventing a new session from being added to the
14361 To avoid this problem, we now set s->new_session to 2 instead of
14362 using a local variable.
14364 *Lutz Jaenicke, Bodo Moeller*
14366 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14367 if the SSL_R_LENGTH_MISMATCH error is detected.
14369 *Geoff Thorpe, Bodo Moeller*
14371 * New 'shared_ldflag' column in Configure platform table.
14375 * Fix EVP_CIPHER_mode macro.
14377 *"Dan S. Camper" <dan@bti.net>*
14379 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14380 type, we must throw them away by setting rr->length to 0.
14382 *D P Chang <dpc@qualys.com>*
14384 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14386 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14387 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14388 worked incorrectly for those cases where range = `10..._2` and
14389 `3*range` is two bits longer than range.)
14393 * Only add signing time to PKCS7 structures if it is not already
14398 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14399 OBJ_ld_ce should be OBJ_id_ce.
14400 Also some ip-pda OIDs in crypto/objects/objects.txt were
14401 incorrect (cf. RFC 3039).
14403 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14405 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14406 returns early because it has nothing to do.
14408 *Andy Schneider <andy.schneider@bjss.co.uk>*
14410 * [In 0.9.6c-engine release:]
14411 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14413 *Andy Schneider <andy.schneider@bjss.co.uk>*
14415 * [In 0.9.6c-engine release:]
14416 Add support for Cryptographic Appliance's keyserver technology.
14417 (Use engine 'keyclient')
14419 *Cryptographic Appliances and Geoff Thorpe*
14421 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14422 is called via tools/c89.sh because arguments have to be
14423 rearranged (all '-L' options must appear before the first object
14426 *Richard Shapiro <rshapiro@abinitio.com>*
14428 * [In 0.9.6c-engine release:]
14429 Add support for Broadcom crypto accelerator cards, backported
14432 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14434 * [In 0.9.6c-engine release:]
14435 Add support for SureWare crypto accelerator cards from
14436 Baltimore Technologies. (Use engine 'sureware')
14438 *Baltimore Technologies and Mark Cox*
14440 * [In 0.9.6c-engine release:]
14441 Add support for crypto accelerator cards from Accelerated
14442 Encryption Processing, www.aep.ie. (Use engine 'aep')
14444 *AEP Inc. and Mark Cox*
14446 * Add a configuration entry for gcc on UnixWare.
14448 *Gary Benson <gbenson@redhat.com>*
14450 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14451 messages are stored in a single piece (fixed-length part and
14452 variable-length part combined) and fix various bugs found on the way.
14456 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14457 instead. BIO_gethostbyname() does not know what timeouts are
14458 appropriate, so entries would stay in cache even when they have
14460 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14462 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14463 faced with a pathologically small ClientHello fragment that does
14464 not contain client_version: Instead of aborting with an error,
14465 simply choose the highest available protocol version (i.e.,
14466 TLS 1.0 unless it is disabled). In practice, ClientHello
14467 messages are never sent like this, but this change gives us
14468 strictly correct behaviour at least for TLS.
14472 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14473 never resets s->method to s->ctx->method when called from within
14474 one of the SSL handshake functions.
14476 *Bodo Moeller; problem pointed out by Niko Baric*
14478 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14479 (sent using the client's version number) if client_version is
14480 smaller than the protocol version in use. Also change
14481 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14482 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14483 the client will at least see that alert.
14487 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14492 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14493 client receives HelloRequest while in a handshake.
14495 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14497 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14498 should end in 'break', not 'goto end' which circumvents various
14499 cleanups done in state SSL_ST_OK. But session related stuff
14500 must be disabled for SSL_ST_OK in the case that we just sent a
14503 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14504 before just sending a HelloRequest.
14506 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14508 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14509 reveal whether illegal block cipher padding was found or a MAC
14510 verification error occurred. (Neither SSLerr() codes nor alerts
14511 are directly visible to potential attackers, but the information
14512 may leak via logfiles.)
14514 Similar changes are not required for the SSL 2.0 implementation
14515 because the number of padding bytes is sent in clear for SSL 2.0,
14516 and the extra bytes are just ignored. However ssl/s2_pkt.c
14517 failed to verify that the purported number of padding bytes is in
14522 * Add OpenUNIX-8 support including shared libraries
14523 (Boyd Lynn Gerber <gerberb@zenez.com>).
14527 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14528 'wristwatch attack' using huge encoding parameters (cf.
14529 James H. Manger's CRYPTO 2001 paper). Note that the
14530 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14531 encoding parameters and hence was not vulnerable.
14535 * BN_sqr() bug fix.
14537 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14539 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14540 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14541 followed by modular reduction.
14543 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14545 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14546 equivalent based on BN_pseudo_rand() instead of BN_rand().
14550 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14551 This function was broken, as the check for a new client hello message
14552 to handle SGC did not allow these large messages.
14553 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14557 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14561 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14562 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14566 * Rework the configuration and shared library support for Tru64 Unix.
14567 The configuration part makes use of modern compiler features and
14568 still retains old compiler behavior for those that run older versions
14569 of the OS. The shared library support part includes a variant that
14570 uses the RPATH feature, and is available through the special
14571 configuration target "alpha-cc-rpath", which will never be selected
14574 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14576 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14577 with the same message size as in ssl3_get_certificate_request().
14578 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14579 messages might inadvertently be reject as too long.
14581 *Petr Lampa <lampa@fee.vutbr.cz>*
14583 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14587 * Modified SSL library such that the verify_callback that has been set
14588 specifically for an SSL object with SSL_set_verify() is actually being
14589 used. Before the change, a verify_callback set with this function was
14590 ignored and the verify_callback() set in the SSL_CTX at the time of
14591 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14592 to allow the necessary settings.
14596 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14597 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14598 done automatically (in contradiction to the requirements of the C
14599 standard). This made problems when used from OpenSSH.
14603 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14604 dh->length and always used
14606 BN_rand_range(priv_key, dh->p).
14608 BN_rand_range() is not necessary for Diffie-Hellman, and this
14609 specific range makes Diffie-Hellman unnecessarily inefficient if
14610 dh->length (recommended exponent length) is much smaller than the
14611 length of dh->p. We could use BN_rand_range() if the order of
14612 the subgroup was stored in the DH structure, but we only have
14617 BN_rand(priv_key, l, ...)
14619 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14626 RSA_eay_public_encrypt
14627 RSA_eay_private_decrypt
14628 RSA_eay_private_encrypt (signing)
14629 RSA_eay_public_decrypt (signature verification)
14631 (default implementations for RSA_public_encrypt,
14632 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14633 always reject numbers >= n.
14637 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14638 to synchronize access to 'locking_thread'. This is necessary on
14639 systems where access to 'locking_thread' (an 'unsigned long'
14640 variable) is not atomic.
14644 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14645 *before* setting the 'crypto_lock_rand' flag. The previous code had
14646 a race condition if 0 is a valid thread ID.
14648 *Travis Vitek <vitek@roguewave.com>*
14650 * Add support for shared libraries under Irix.
14652 *Albert Chin-A-Young <china@thewrittenword.com>*
14654 * Add configuration option to build on Linux on both big-endian and
14655 little-endian MIPS.
14657 *Ralf Baechle <ralf@uni-koblenz.de>*
14659 * Add the possibility to create shared libraries on HP-UX.
14663 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14665 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14666 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14667 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14668 PRNG state recovery was possible based on the output of
14669 one PRNG request appropriately sized to gain knowledge on
14670 'md' followed by enough consecutive 1-byte PRNG requests
14671 to traverse all of 'state'.
14673 1. When updating 'md_local' (the current thread's copy of 'md')
14674 during PRNG output generation, hash all of the previous
14675 'md_local' value, not just the half used for PRNG output.
14677 2. Make the number of bytes from 'state' included into the hash
14678 independent from the number of PRNG bytes requested.
14680 The first measure alone would be sufficient to avoid
14681 Markku-Juhani's attack. (Actually it had never occurred
14682 to me that the half of 'md_local' used for chaining was the
14683 half from which PRNG output bytes were taken -- I had always
14684 assumed that the secret half would be used.) The second
14685 measure makes sure that additional data from 'state' is never
14686 mixed into 'md_local' in small portions; this heuristically
14687 further strengthens the PRNG.
14691 * Fix crypto/bn/asm/mips3.s.
14695 * When only the key is given to "enc", the IV is undefined. Print out
14696 an error message in this case.
14700 * Handle special case when X509_NAME is empty in X509 printing routines.
14704 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14705 positive and less than q.
14709 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14710 used: it isn't thread safe and the add_lock_callback should handle
14713 *Paul Rose <Paul.Rose@bridge.com>*
14715 * Verify that incoming data obeys the block size in
14716 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14722 *Ulf Möller, Bodo Möller*
14724 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14725 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14726 when fixing the server behaviour for backwards-compatible 'client
14727 hello' messages. (Note that the attack is impractical against
14728 SSL 3.0 and TLS 1.0 anyway because length and version checking
14729 means that the probability of guessing a valid ciphertext is
14730 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14733 Before 0.9.5, the countermeasure (hide the error by generating a
14734 random 'decryption result') did not work properly because
14735 ERR_clear_error() was missing, meaning that SSL_get_error() would
14736 detect the supposedly ignored error.
14738 Both problems are now fixed.
14742 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14743 (previously it was 1024).
14747 * Fix for compatibility mode trust settings: ignore trust settings
14748 unless some valid trust or reject settings are present.
14752 * Fix for blowfish EVP: its a variable length cipher.
14756 * Fix various bugs related to DSA S/MIME verification. Handle missing
14757 parameters in DSA public key structures and return an error in the
14758 DSA routines if parameters are absent.
14762 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14763 in the current directory if neither $RANDFILE nor $HOME was set.
14764 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14765 caused some confusion to Windows users who haven't defined $HOME.
14766 Thus RAND_file_name() is changed again: e_os.h can define a
14767 DEFAULT_HOME, which will be used if $HOME is not set.
14768 For Windows, we use "C:"; on other platforms, we still require
14769 environment variables.
14771 * Move 'if (!initialized) RAND_poll()' into regions protected by
14772 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14773 having multiple threads call RAND_poll() concurrently.
14777 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14778 combination of a flag and a thread ID variable.
14779 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14780 flag), *other* threads can enter ssleay_add_bytes without obeying
14781 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14782 that they do not hold after the first thread unsets add_do_not_lock).
14786 * Change bctest again: '-x' expressions are not available in all
14787 versions of 'test'.
14791 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14793 * Fix a couple of memory leaks in PKCS7_dataDecode()
14795 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14797 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14798 the default extension for executables, if any. Also, make the perl
14799 scripts that use symlink() to test if it really exists and use "cp"
14800 if it doesn't. All this made OpenSSL compilable and installable in
14805 * Fix for asn1_GetSequence() for indefinite length constructed data.
14806 If SEQUENCE is length is indefinite just set c->slen to the total
14807 amount of data available.
14809 *Steve Henson, reported by shige@FreeBSD.org*
14811 *This change does not apply to 0.9.7.*
14813 * Change bctest to avoid here-documents inside command substitution
14814 (workaround for FreeBSD /bin/sh bug).
14815 For compatibility with Ultrix, avoid shell functions (introduced
14816 in the bctest version that searches along $PATH).
14820 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14821 with des_encrypt() defined on some operating systems, like Solaris
14826 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14827 On the Importance of Eliminating Errors in Cryptographic
14828 Computations, J. Cryptology 14 (2001) 2, 101-119,
14829 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14833 * MIPS assembler BIGNUM division bug fix.
14837 * Disabled incorrect Alpha assembler code.
14841 * Fix PKCS#7 decode routines so they correctly update the length
14842 after reading an EOC for the EXPLICIT tag.
14846 *This change does not apply to 0.9.7.*
14848 * Fix bug in PKCS#12 key generation routines. This was triggered
14849 if a 3DES key was generated with a 0 initial byte. Include
14850 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14851 (but broken) behaviour.
14855 * Enhance bctest to search for a working bc along $PATH and print
14858 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14860 * Fix memory leaks in err.c: free err_data string if necessary;
14861 don't write to the wrong index in ERR_set_error_data.
14865 * Implement ssl23_peek (analogous to ssl23_read), which previously
14870 * Replace rdtsc with `_emit` statements for VC++ version 5.
14872 *Jeremy Cooper <jeremy@baymoo.org>*
14874 * Make it possible to reuse SSLv2 sessions.
14878 * In copy_email() check for >= 0 as a return value for
14879 X509_NAME_get_index_by_NID() since 0 is a valid index.
14881 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14883 * Avoid coredump with unsupported or invalid public keys by checking if
14884 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14885 PKCS7_verify() fails with non detached data.
14889 * Don't use getenv in library functions when run as setuid/setgid.
14890 New function OPENSSL_issetugid().
14894 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14895 due to incorrect handling of multi-threading:
14897 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14899 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14901 3. Count how many times MemCheck_off() has been called so that
14902 nested use can be treated correctly. This also avoids
14903 inband-signalling in the previous code (which relied on the
14904 assumption that thread ID 0 is impossible).
14908 * Add "-rand" option also to s_client and s_server.
14912 * Fix CPU detection on Irix 6.x.
14913 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14914 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14916 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14921 *This change does not apply to 0.9.7.*
14923 * Use the cached encoding of an X509_NAME structure rather than
14924 copying it. This is apparently the reason for the libsafe "errors"
14925 but the code is actually correct.
14929 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14930 Bleichenbacher's DSA attack.
14931 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14932 to be set and top=0 forces the highest bit to be set; top=-1 is new
14933 and leaves the highest bit random.
14935 *Ulf Moeller, Bodo Moeller*
14937 * In the `NCONF_...`-based implementations for `CONF_...` queries
14938 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14939 a temporary CONF structure with the data component set to NULL
14940 (which gives segmentation faults in lh_retrieve).
14941 Instead, use NULL for the CONF pointer in CONF_get_string and
14942 CONF_get_number (which may use environment variables) and directly
14943 return NULL from CONF_get_section.
14947 * Fix potential buffer overrun for EBCDIC.
14951 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14952 keyUsage if basicConstraints absent for a CA.
14956 * Make SMIME_write_PKCS7() write mail header values with a format that
14957 is more generally accepted (no spaces before the semicolon), since
14958 some programs can't parse those values properly otherwise. Also make
14959 sure BIO's that break lines after each write do not create invalid
14964 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14965 macros previously used would not encode an empty SEQUENCE OF
14966 and break the signature.
14970 *This change does not apply to 0.9.7.*
14972 * Zero the premaster secret after deriving the master secret in
14977 * Add some EVP_add_digest_alias registrations (as found in
14978 OpenSSL_add_all_digests()) to SSL_library_init()
14979 aka OpenSSL_add_ssl_algorithms(). This provides improved
14980 compatibility with peers using X.509 certificates
14981 with unconventional AlgorithmIdentifier OIDs.
14985 * Fix for Irix with NO_ASM.
14987 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14989 * ./config script fixes.
14991 *Ulf Moeller, Richard Levitte*
14993 * Fix 'openssl passwd -1'.
14997 * Change PKCS12_key_gen_asc() so it can cope with non null
14998 terminated strings whose length is passed in the passlen
14999 parameter, for example from PEM callbacks. This was done
15000 by adding an extra length parameter to asc2uni().
15002 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15004 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15005 call failed, free the DSA structure.
15009 * Fix to uni2asc() to cope with zero length Unicode strings.
15010 These are present in some PKCS#12 files.
15014 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15015 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15016 when writing a 32767 byte record.
15018 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15020 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15021 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15023 (RSA objects have a reference count access to which is protected
15024 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15025 so they are meant to be shared between threads.)
15026 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15027 "Reddie, Steven" <Steven.Reddie@ca.com>*
15029 * Fix a deadlock in CRYPTO_mem_leaks().
15033 * Use better test patterns in bntest.
15037 * rand_win.c fix for Borland C.
15041 * BN_rshift bugfix for n == 0.
15045 * Add a 'bctest' script that checks for some known 'bc' bugs
15046 so that 'make test' does not abort just because 'bc' is broken.
15050 * Store verify_result within SSL_SESSION also for client side to
15051 avoid potential security hole. (Re-used sessions on the client side
15052 always resulted in verify_result==X509_V_OK, not using the original
15053 result of the server certificate verification.)
15057 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15058 SSL3_RT_APPLICATION_DATA, return 0.
15059 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15064 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15065 releases, have been re-implemented by renaming the previous
15066 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15067 and ssl3_read_internal, respectively, and adding 'peek' parameters
15068 to them. The new ssl[23]_{read,peek} functions are calls to
15069 ssl[23]_read_internal with the 'peek' flag set appropriately.
15070 A 'peek' parameter has also been added to ssl3_read_bytes, which
15071 does the actual work for ssl3_read_internal.
15075 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15076 the method-specific "init()" handler. Also clean up ex_data after
15077 calling the method-specific "finish()" handler. Previously, this was
15078 happening the other way round.
15082 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15083 The previous value, 12, was not always sufficient for BN_mod_exp().
15087 * Make sure that shared libraries get the internal name engine with
15088 the full version number and not just 0. This should mark the
15089 shared libraries as not backward compatible. Of course, this should
15090 be changed again when we can guarantee backward binary compatibility.
15094 * Fix typo in get_cert_by_subject() in by_dir.c
15096 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15098 * Rework the system to generate shared libraries:
15100 - Make note of the expected extension for the shared libraries and
15101 if there is a need for symbolic links from for example libcrypto.so.0
15102 to libcrypto.so.0.9.7. There is extended info in Configure for
15105 - Make as few rebuilds of the shared libraries as possible.
15107 - Still avoid linking the OpenSSL programs with the shared libraries.
15109 - When installing, install the shared libraries separately from the
15114 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15116 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15117 and not in SSL_clear because the latter is also used by the
15118 accept/connect functions; previously, the settings made by
15119 SSL_set_read_ahead would be lost during the handshake.
15121 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15123 * Correct util/mkdef.pl to be selective about disabled algorithms.
15124 Previously, it would create entries for disabled algorithms no
15129 * Added several new manual pages for SSL_* function.
15133 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15135 * In ssl23_get_client_hello, generate an error message when faced
15136 with an initial SSL 3.0/TLS record that is too small to contain the
15137 first two bytes of the ClientHello message, i.e. client_version.
15138 (Note that this is a pathologic case that probably has never happened
15139 in real life.) The previous approach was to use the version number
15140 from the record header as a substitute; but our protocol choice
15141 should not depend on that one because it is not authenticated
15142 by the Finished messages.
15146 * More robust randomness gathering functions for Windows.
15148 *Jeffrey Altman <jaltman@columbia.edu>*
15150 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15151 not set then we don't setup the error code for issuer check errors
15152 to avoid possibly overwriting other errors which the callback does
15153 handle. If an application does set the flag then we assume it knows
15154 what it is doing and can handle the new informational codes
15159 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15160 a general "ANY" type, as such it should be able to decode anything
15161 including tagged types. However it didn't check the class so it would
15162 wrongly interpret tagged types in the same way as their universal
15163 counterpart and unknown types were just rejected. Changed so that the
15164 tagged and unknown types are handled in the same way as a SEQUENCE:
15165 that is the encoding is stored intact. There is also a new type
15166 "V_ASN1_OTHER" which is used when the class is not universal, in this
15167 case we have no idea what the actual type is so we just lump them all
15172 * On VMS, stdout may very well lead to a file that is written to
15173 in a record-oriented fashion. That means that every write() will
15174 write a separate record, which will be read separately by the
15175 programs trying to read from it. This can be very confusing.
15177 The solution is to put a BIO filter in the way that will buffer
15178 text until a linefeed is reached, and then write everything a
15179 line at a time, so every record written will be an actual line,
15180 not chunks of lines and not (usually doesn't happen, but I've
15181 seen it once) several lines in one record. BIO_f_linebuffer() is
15184 Currently, it's a VMS-only method, because that's where it has
15185 been tested well enough.
15189 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15190 it can return incorrect results.
15191 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15192 but it was in 0.9.6-beta[12].)
15196 * Disable the check for content being present when verifying detached
15197 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15198 include zero length content when signing messages.
15202 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15203 BIO_ctrl (for BIO pairs).
15207 * Add DSO method for VMS.
15211 * Bug fix: Montgomery multiplication could produce results with the
15216 * Add RPM specification openssl.spec and modify it to build three
15217 packages. The default package contains applications, application
15218 documentation and run-time libraries. The devel package contains
15219 include files, static libraries and function documentation. The
15220 doc package contains the contents of the doc directory. The original
15221 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15225 * Add a large number of documentation files for many SSL routines.
15227 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15229 * Add a configuration entry for Sony News 4.
15231 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15233 * Don't set the two most significant bits to one when generating a
15234 random number < q in the DSA library.
15238 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15239 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15240 the underlying transport is blocking) if a handshake took place.
15241 (The default behaviour is needed by applications such as s_client
15242 and s_server that use select() to determine when to use SSL_read;
15243 but for applications that know in advance when to expect data, it
15244 just makes things more complicated.)
15248 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15253 * Add a few more EBCDIC conditionals that make `req` and `x509`
15254 work better on such systems.
15256 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15258 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15259 Update PKCS12_parse() so it copies the friendlyName and the
15260 keyid to the certificates aux info.
15264 * Fix bug in PKCS7_verify() which caused an infinite loop
15265 if there was more than one signature.
15267 *Sven Uszpelkat <su@celocom.de>*
15269 * Major change in util/mkdef.pl to include extra information
15270 about each symbol, as well as presenting variables as well
15271 as functions. This change means that there's n more need
15272 to rebuild the .num files when some algorithms are excluded.
15276 * Allow the verify time to be set by an application,
15277 rather than always using the current time.
15281 * Phase 2 verify code reorganisation. The certificate
15282 verify code now looks up an issuer certificate by a
15283 number of criteria: subject name, authority key id
15284 and key usage. It also verifies self signed certificates
15285 by the same criteria. The main comparison function is
15286 X509_check_issued() which performs these checks.
15288 Lot of changes were necessary in order to support this
15289 without completely rewriting the lookup code.
15291 Authority and subject key identifier are now cached.
15293 The LHASH 'certs' is X509_STORE has now been replaced
15294 by a STACK_OF(X509_OBJECT). This is mainly because an
15295 LHASH can't store or retrieve multiple objects with
15296 the same hash value.
15298 As a result various functions (which were all internal
15299 use only) have changed to handle the new X509_STORE
15300 structure. This will break anything that messed round
15301 with X509_STORE internally.
15303 The functions X509_STORE_add_cert() now checks for an
15304 exact match, rather than just subject name.
15306 The X509_STORE API doesn't directly support the retrieval
15307 of multiple certificates matching a given criteria, however
15308 this can be worked round by performing a lookup first
15309 (which will fill the cache with candidate certificates)
15310 and then examining the cache for matches. This is probably
15311 the best we can do without throwing out X509_LOOKUP
15312 entirely (maybe later...).
15314 The X509_VERIFY_CTX structure has been enhanced considerably.
15316 All certificate lookup operations now go via a get_issuer()
15317 callback. Although this currently uses an X509_STORE it
15318 can be replaced by custom lookups. This is a simple way
15319 to bypass the X509_STORE hackery necessary to make this
15320 work and makes it possible to use more efficient techniques
15321 in future. A very simple version which uses a simple
15322 STACK for its trusted certificate store is also provided
15323 using X509_STORE_CTX_trusted_stack().
15325 The verify_cb() and verify() callbacks now have equivalents
15326 in the X509_STORE_CTX structure.
15328 X509_STORE_CTX also has a 'flags' field which can be used
15329 to customise the verify behaviour.
15333 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15334 excludes S/MIME capabilities.
15338 * When a certificate request is read in keep a copy of the
15339 original encoding of the signed data and use it when outputting
15340 again. Signatures then use the original encoding rather than
15341 a decoded, encoded version which may cause problems if the
15342 request is improperly encoded.
15346 * For consistency with other BIO_puts implementations, call
15347 buffer_write(b, ...) directly in buffer_puts instead of calling
15350 In BIO_puts, increment b->num_write as in BIO_write.
15352 *Peter.Sylvester@EdelWeb.fr*
15354 * Fix BN_mul_word for the case where the word is 0. (We have to use
15355 BN_zero, we may not return a BIGNUM with an array consisting of
15356 words set to zero.)
15360 * Avoid calling abort() from within the library when problems are
15361 detected, except if preprocessor symbols have been defined
15362 (such as REF_CHECK, BN_DEBUG etc.).
15366 * New openssl application 'rsautl'. This utility can be
15367 used for low-level RSA operations. DER public key
15368 BIO/fp routines also added.
15372 * New Configure entry and patches for compiling on QNX 4.
15374 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15376 * A demo state-machine implementation was sponsored by
15377 Nuron (<http://www.nuron.com/>) and is now available in
15378 demos/state_machine.
15382 * New options added to the 'dgst' utility for signature
15383 generation and verification.
15387 * Unrecognized PKCS#7 content types are now handled via a
15388 catch all ASN1_TYPE structure. This allows unsupported
15389 types to be stored as a "blob" and an application can
15390 encode and decode it manually.
15394 * Fix various signed/unsigned issues to make a_strex.c
15395 compile under VC++.
15397 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15399 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15400 length if passed a buffer. ASN1_INTEGER_to_BN failed
15401 if passed a NULL BN and its argument was negative.
15403 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15405 * Modification to PKCS#7 encoding routines to output definite
15406 length encoding. Since currently the whole structures are in
15407 memory there's not real point in using indefinite length
15408 constructed encoding. However if OpenSSL is compiled with
15409 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15413 * Added BIO_vprintf() and BIO_vsnprintf().
15417 * Added more prefixes to parse for in the strings written
15418 through a logging bio, to cover all the levels that are available
15419 through syslog. The prefixes are now:
15421 PANIC, EMERG, EMR => LOG_EMERG
15422 ALERT, ALR => LOG_ALERT
15423 CRIT, CRI => LOG_CRIT
15424 ERROR, ERR => LOG_ERR
15425 WARNING, WARN, WAR => LOG_WARNING
15426 NOTICE, NOTE, NOT => LOG_NOTICE
15427 INFO, INF => LOG_INFO
15428 DEBUG, DBG => LOG_DEBUG
15430 and as before, if none of those prefixes are present at the
15431 beginning of the string, LOG_ERR is chosen.
15433 On Win32, the `LOG_*` levels are mapped according to this:
15435 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15436 LOG_WARNING => EVENTLOG_WARNING_TYPE
15437 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15441 * Made it possible to reconfigure with just the configuration
15442 argument "reconf" or "reconfigure". The command line arguments
15443 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15444 and are retrieved from there when reconfiguring.
15450 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15452 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15456 * The obj_dat.pl script was messing up the sorting of object
15457 names. The reason was that it compared the quoted version
15458 of strings as a result "OCSP" > "OCSP Signing" because
15459 " > SPACE. Changed script to store unquoted versions of
15460 names and add quotes on output. It was also omitting some
15461 names from the lookup table if they were given a default
15462 value (that is if SN is missing it is given the same
15463 value as LN and vice versa), these are now added on the
15464 grounds that if an object has a name we should be able to
15465 look it up. Finally added warning output when duplicate
15466 short or long names are found.
15470 * Changes needed for Tandem NSK.
15472 *Scott Uroff <scott@xypro.com>*
15474 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15475 RSA_padding_check_SSLv23(), special padding was never detected
15476 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15477 version rollback attacks was not effective.
15479 In s23_clnt.c, don't use special rollback-attack detection padding
15480 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15481 client; similarly, in s23_srvr.c, don't do the rollback check if
15482 SSL 2.0 is the only protocol enabled in the server.
15486 * Make it possible to get hexdumps of unprintable data with 'openssl
15487 asn1parse'. By implication, the functions ASN1_parse_dump() and
15488 BIO_dump_indent() are added.
15492 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15493 these print out strings and name structures based on various
15494 flags including RFC2253 support and proper handling of
15495 multibyte characters. Added options to the 'x509' utility
15496 to allow the various flags to be set.
15500 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15501 Also change the functions X509_cmp_current_time() and
15502 X509_gmtime_adj() work with an ASN1_TIME structure,
15503 this will enable certificates using GeneralizedTime in validity
15504 dates to be checked.
15508 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15509 negative public key encodings) on by default,
15510 NO_NEG_PUBKEY_BUG can be set to disable it.
15514 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15515 content octets. An i2c_ASN1_OBJECT is unnecessary because
15516 the encoding can be trivially obtained from the structure.
15520 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15521 not read locks (`CRYPTO_r_[un]lock`).
15525 * A first attempt at creating official support for shared
15526 libraries through configuration. I've kept it so the
15527 default is static libraries only, and the OpenSSL programs
15528 are always statically linked for now, but there are
15529 preparations for dynamic linking in place.
15530 This has been tested on Linux and Tru64.
15534 * Randomness polling function for Win9x, as described in:
15535 Peter Gutmann, Software Generation of Practically Strong
15540 * Fix so PRNG is seeded in req if using an already existing
15545 * New options to smime application. -inform and -outform
15546 allow alternative formats for the S/MIME message including
15547 PEM and DER. The -content option allows the content to be
15548 specified separately. This should allow things like Netscape
15549 form signing output easier to verify.
15553 * Fix the ASN1 encoding of tags using the 'long form'.
15557 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15558 STRING types. These convert content octets to and from the
15559 underlying type. The actual tag and length octets are
15560 already assumed to have been read in and checked. These
15561 are needed because all other string types have virtually
15562 identical handling apart from the tag. By having versions
15563 of the ASN1 functions that just operate on content octets
15564 IMPLICIT tagging can be handled properly. It also allows
15565 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15566 and ASN1_INTEGER are identical apart from the tag.
15570 * Change the handling of OID objects as follows:
15572 - New object identifiers are inserted in objects.txt, following
15573 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15574 - objects.pl is used to process obj_mac.num and create a new
15576 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15579 This is currently kind of a hack, and the perl code in objects.pl
15580 isn't very elegant, but it works as I intended. The simplest way
15581 to check that it worked correctly is to look in obj_dat.h and
15582 check the array nid_objs and make sure the objects haven't moved
15583 around (this is important!). Additions are OK, as well as
15584 consistent name changes.
15588 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15592 * Addition of the command line parameter '-rand file' to 'openssl req'.
15593 The given file adds to whatever has already been seeded into the
15594 random pool through the RANDFILE configuration file option or
15595 environment variable, or the default random state file.
15599 * mkstack.pl now sorts each macro group into lexical order.
15600 Previously the output order depended on the order the files
15601 appeared in the directory, resulting in needless rewriting
15606 * Patches to make OpenSSL compile under Win32 again. Mostly
15607 work arounds for the VC++ problem that it treats func() as
15608 func(void). Also stripped out the parts of mkdef.pl that
15609 added extra typesafe functions: these no longer exist.
15613 * Reorganisation of the stack code. The macros are now all
15614 collected in safestack.h . Each macro is defined in terms of
15615 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15616 DEBUG_SAFESTACK is now handled in terms of function casts,
15617 this has the advantage of retaining type safety without the
15618 use of additional functions. If DEBUG_SAFESTACK is not defined
15619 then the non typesafe macros are used instead. Also modified the
15620 mkstack.pl script to handle the new form. Needs testing to see
15621 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15622 the default if no major problems. Similar behaviour for ASN1_SET_OF
15623 and PKCS12_STACK_OF.
15627 * When some versions of IIS use the 'NET' form of private key the
15628 key derivation algorithm is different. Normally MD5(password) is
15629 used as a 128 bit RC4 key. In the modified case
15630 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15631 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15632 as the old Netscape_RSA functions except they have an additional
15633 'sgckey' parameter which uses the modified algorithm. Also added
15634 an -sgckey command line option to the rsa utility. Thanks to
15635 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15636 algorithm to openssl-dev.
15640 * The evp_local.h macros were using 'c.##kname' which resulted in
15641 invalid expansion on some systems (SCO 5.0.5 for example).
15642 Corrected to 'c.kname'.
15644 *Phillip Porch <root@theporch.com>*
15646 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15647 a STACK of email addresses from a certificate or request, these look
15648 in the subject name and the subject alternative name extensions and
15649 omit any duplicate addresses.
15653 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15654 This makes DSA verification about 2 % faster.
15658 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15659 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15660 plus overhead for 1024 bit moduli).
15661 This makes exponentiations about 0.5 % faster for 1024 bit
15662 exponents (as measured by "openssl speed rsa2048").
15666 * Rename memory handling macros to avoid conflicts with other
15668 Malloc => OPENSSL_malloc
15669 Malloc_locked => OPENSSL_malloc_locked
15670 Realloc => OPENSSL_realloc
15671 Free => OPENSSL_free
15675 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15676 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15680 * CygWin32 support.
15682 *John Jarvie <jjarvie@newsguy.com>*
15684 * The type-safe stack code has been rejigged. It is now only compiled
15685 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15686 by default all type-specific stack functions are "#define"d back to
15687 standard stack functions. This results in more streamlined output
15688 but retains the type-safety checking possibilities of the original
15693 * The STACK code has been cleaned up, and certain type declarations
15694 that didn't make a lot of sense have been brought in line. This has
15695 also involved a cleanup of sorts in safestack.h to more correctly
15696 map type-safe stack functions onto their plain stack counterparts.
15697 This work has also resulted in a variety of "const"ifications of
15698 lots of the code, especially `_cmp` operations which should normally
15699 be prototyped with "const" parameters anyway.
15703 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15704 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15705 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15706 where all of 'md' is used each time the PRNG is used, but 'state'
15707 is used only indexed by a cyclic counter. As entropy may not be
15708 well distributed from the beginning, 'md' is important as a
15709 chaining variable. However, the output function chains only half
15710 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15711 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15712 in all of 'state' being rewritten, with the new values depending
15713 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15717 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15718 the handshake is continued after ssl_verify_cert_chain();
15719 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15720 can lead to 'unexplainable' connection aborts later.
15722 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15724 * Major EVP API cipher revision.
15725 Add hooks for extra EVP features. This allows various cipher
15726 parameters to be set in the EVP interface. Support added for variable
15727 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15728 setting of RC2 and RC5 parameters.
15730 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15733 Remove lots of duplicated code from the EVP library. For example *every*
15734 cipher init() function handles the 'iv' in the same way according to the
15735 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15736 for CFB and OFB modes they zero ctx->num.
15738 New functionality allows removal of S/MIME code RC2 hack.
15740 Most of the routines have the same form and so can be declared in terms
15743 By shifting this to the top level EVP_CipherInit() it can be removed from
15744 all individual ciphers. If the cipher wants to handle IVs or keys
15745 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15748 Change lots of functions like EVP_EncryptUpdate() to now return a
15749 value: although software versions of the algorithms cannot fail
15750 any installed hardware versions can.
15754 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15755 this option is set, tolerate broken clients that send the negotiated
15756 protocol version number instead of the requested protocol version
15761 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15762 i.e. non-zero for export ciphersuites, zero otherwise.
15763 Previous versions had this flag inverted, inconsistent with
15764 rsa_tmp_cb (..._TMP_RSA_CB).
15766 *Bodo Moeller; problem reported by Amit Chopra*
15768 * Add missing DSA library text string. Work around for some IIS
15769 key files with invalid SEQUENCE encoding.
15773 * Add a document (doc/standards.txt) that list all kinds of standards
15774 and so on that are implemented in OpenSSL.
15778 * Enhance c_rehash script. Old version would mishandle certificates
15779 with the same subject name hash and wouldn't handle CRLs at all.
15780 Added -fingerprint option to crl utility, to support new c_rehash
15785 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15789 * Fix for SSL server purpose checking. Server checking was
15790 rejecting certificates which had extended key usage present
15791 but no ssl client purpose.
15793 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15795 * Make PKCS#12 code work with no password. The PKCS#12 spec
15796 is a little unclear about how a blank password is handled.
15797 Since the password in encoded as a BMPString with terminating
15798 double NULL a zero length password would end up as just the
15799 double NULL. However no password at all is different and is
15800 handled differently in the PKCS#12 key generation code. NS
15801 treats a blank password as zero length. MSIE treats it as no
15802 password on export: but it will try both on import. We now do
15803 the same: PKCS12_parse() tries zero length and no password if
15804 the password is set to "" or NULL (NULL is now a valid password:
15805 it wasn't before) as does the pkcs12 application.
15809 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15810 perror when PEM_read_bio_X509_REQ fails, the error message must
15811 be obtained from the error queue.
15815 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15816 it in ERR_remove_state if appropriate, and change ERR_get_state
15817 accordingly to avoid race conditions (this is necessary because
15818 thread_hash is no longer constant once set).
15822 * Bugfix for linux-elf makefile.one.
15826 * RSA_get_default_method() will now cause a default
15827 RSA_METHOD to be chosen if one doesn't exist already.
15828 Previously this was only set during a call to RSA_new()
15829 or RSA_new_method(NULL) meaning it was possible for
15830 RSA_get_default_method() to return NULL.
15834 * Added native name translation to the existing DSO code
15835 that will convert (if the flag to do so is set) filenames
15836 that are sufficiently small and have no path information
15837 into a canonical native form. Eg. "blah" converted to
15838 "libblah.so" or "blah.dll" etc.
15842 * New function ERR_error_string_n(e, buf, len) which is like
15843 ERR_error_string(e, buf), but writes at most 'len' bytes
15844 including the 0 terminator. For ERR_error_string_n, 'buf'
15847 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15849 * CONF library reworked to become more general. A new CONF
15850 configuration file reader "class" is implemented as well as a
15851 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15852 old `CONF_*` functions are still there, but are reimplemented to
15853 work in terms of the new functions. Also, a set of functions
15854 to handle the internal storage of the configuration data is
15855 provided to make it easier to write new configuration file
15856 reader "classes" (I can definitely see something reading a
15857 configuration file in XML format, for example), called `_CONF_*`,
15858 or "the configuration storage API"...
15860 The new configuration file reading functions are:
15862 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15863 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15865 NCONF_default, NCONF_WIN32
15867 NCONF_dump_fp, NCONF_dump_bio
15869 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15870 NCONF_new creates a new CONF object. This works in the same way
15871 as other interfaces in OpenSSL, like the BIO interface.
15872 `NCONF_dump_*` dump the internal storage of the configuration file,
15873 which is useful for debugging. All other functions take the same
15874 arguments as the old `CONF_*` functions with the exception of the
15875 first that must be a `CONF *` instead of a `LHASH *`.
15877 To make it easier to use the new classes with the old `CONF_*` functions,
15878 the function CONF_set_default_method is provided.
15882 * Add '-tls1' option to 'openssl ciphers', which was already
15883 mentioned in the documentation but had not been implemented.
15884 (This option is not yet really useful because even the additional
15885 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15889 * Initial DSO code added into libcrypto for letting OpenSSL (and
15890 OpenSSL-based applications) load shared libraries and bind to
15891 them in a portable way.
15893 *Geoff Thorpe, with contributions from Richard Levitte*
15895 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15897 * Make sure _lrotl and _lrotr are only used with MSVC.
15899 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15900 (the default implementation of RAND_status).
15902 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15903 to '-clrext' (= clear extensions), as intended and documented.
15904 *Bodo Moeller; inconsistency pointed out by Michael Attili
15905 <attili@amaxo.com>*
15907 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15908 was larger than the MD block size.
15910 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15912 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15913 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15914 using the passed key: if the passed key was a private key the result
15915 of X509_print(), for example, would be to print out all the private key
15920 * des_quad_cksum() byte order bug fix.
15921 *Ulf Möller, using the problem description in krb4-0.9.7, where
15922 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15924 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15927 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15929 * For easily testing in shell scripts whether some command
15930 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15931 returns with exit code 0 iff no command of the given name is available.
15932 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15933 the output goes to stdout and nothing is printed to stderr.
15934 Additional arguments are always ignored.
15936 Since for each cipher there is a command of the same name,
15937 the 'no-cipher' compilation switches can be tested this way.
15939 ('openssl no-XXX' is not able to detect pseudo-commands such
15940 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15944 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15948 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15949 is set; it will be thrown away anyway because each handshake creates
15951 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15952 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15953 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15954 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15958 * New s_client option -ign_eof: EOF at stdin is ignored, and
15959 'Q' and 'R' lose their special meanings (quit/renegotiate).
15960 This is part of what -quiet does; unlike -quiet, -ign_eof
15961 does not suppress any output.
15965 * Add compatibility options to the purpose and trust code. The
15966 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15967 accepts a certificate or CA, this was the previous behaviour,
15968 with all the associated security issues.
15970 X509_TRUST_COMPAT is the old trust behaviour: only and
15971 automatically trust self signed roots in certificate store. A
15972 new trust setting X509_TRUST_DEFAULT is used to specify that
15973 a purpose has no associated trust setting and it should instead
15974 use the value in the default purpose.
15978 * Fix the PKCS#8 DSA private key code so it decodes keys again
15979 and fix a memory leak.
15983 * In util/mkerr.pl (which implements 'make errors'), preserve
15984 reason strings from the previous version of the .c file, as
15985 the default to have only downcase letters (and digits) in
15986 automatically generated reasons codes is not always appropriate.
15990 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15991 using strerror. Previously, ERR_reason_error_string() returned
15992 library names as reason strings for SYSerr; but SYSerr is a special
15993 case where small numbers are errno values, not library numbers.
15997 * Add '-dsaparam' option to 'openssl dhparam' application. This
15998 converts DSA parameters into DH parameters. (When creating parameters,
15999 DSA_generate_parameters is used.)
16003 * Include 'length' (recommended exponent length) in C code generated
16004 by 'openssl dhparam -C'.
16008 * The second argument to set_label in perlasm was already being used
16009 so couldn't be used as a "file scope" flag. Moved to third argument
16014 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16015 instead of RAND_bytes for encryption IVs and salts.
16019 * Include RAND_status() into RAND_METHOD instead of implementing
16020 it only for md_rand.c Otherwise replacing the PRNG by calling
16021 RAND_set_rand_method would be impossible.
16025 * Don't let DSA_generate_key() enter an infinite loop if the random
16026 number generation fails.
16030 * New 'rand' application for creating pseudo-random output.
16034 * Added configuration support for Linux/IA64
16036 *Rolf Haberrecker <rolf@suse.de>*
16038 * Assembler module support for Mingw32.
16042 * Shared library support for HPUX (in shlib/).
16044 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16046 * Shared library support for Solaris gcc.
16048 *Lutz Behnke <behnke@trustcenter.de>*
16050 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16052 * PKCS7_encrypt() was adding text MIME headers twice because they
16053 were added manually and by SMIME_crlf_copy().
16057 * In bntest.c don't call BN_rand with zero bits argument.
16059 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16061 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16062 case was implemented. This caused BN_div_recp() to fail occasionally.
16066 * Add an optional second argument to the set_label() in the perl
16067 assembly language builder. If this argument exists and is set
16068 to 1 it signals that the assembler should use a symbol whose
16069 scope is the entire file, not just the current function. This
16070 is needed with MASM which uses the format label:: for this scope.
16072 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16074 * Change the ASN1 types so they are typedefs by default. Before
16075 almost all types were #define'd to ASN1_STRING which was causing
16076 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16081 * Change names of new functions to the new get1/get0 naming
16082 convention: After 'get1', the caller owns a reference count
16083 and has to call `..._free`; 'get0' returns a pointer to some
16084 data structure without incrementing reference counters.
16085 (Some of the existing 'get' functions increment a reference
16086 counter, some don't.)
16087 Similarly, 'set1' and 'add1' functions increase reference
16088 counters or duplicate objects.
16092 * Allow for the possibility of temp RSA key generation failure:
16093 the code used to assume it always worked and crashed on failure.
16097 * Fix potential buffer overrun problem in BIO_printf().
16098 *Ulf Möller, using public domain code by Patrick Powell; problem
16099 pointed out by David Sacerdote <das33@cornell.edu>*
16101 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16102 RAND_egd() and RAND_status(). In the command line application,
16103 the EGD socket can be specified like a seed file using RANDFILE
16108 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16109 Some CAs (e.g. Verisign) distribute certificates in this form.
16113 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16114 list to exclude them. This means that no special compilation option
16115 is needed to use anonymous DH: it just needs to be included in the
16120 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16121 EVP_MD_type. The old functionality is available in a new macro called
16122 EVP_MD_md(). Change code that uses it and update docs.
16126 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16127 where the `void *` argument is replaced by a function pointer argument.
16128 Previously `void *` was abused to point to functions, which works on
16129 many platforms, but is not correct. As these functions are usually
16130 called by macros defined in OpenSSL header files, most source code
16131 should work without changes.
16135 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16136 sections with information on -D... compiler switches used for
16137 compiling the library so that applications can see them. To enable
16138 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16139 must be defined. E.g.,
16140 #define OPENSSL_ALGORITHM_DEFINES
16141 #include <openssl/opensslconf.h>
16142 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16144 *Richard Levitte, Ulf and Bodo Möller*
16146 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16151 * Change the 'other' type in certificate aux info to a STACK_OF
16152 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16153 the required ASN1 format: arbitrary types determined by an OID.
16157 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16158 argument to 'req'. This is not because the function is newer or
16159 better than others it just uses the work 'NEW' in the certificate
16160 request header lines. Some software needs this.
16164 * Reorganise password command line arguments: now passwords can be
16165 obtained from various sources. Delete the PEM_cb function and make
16166 it the default behaviour: i.e. if the callback is NULL and the
16167 usrdata argument is not NULL interpret it as a null terminated pass
16168 phrase. If usrdata and the callback are NULL then the pass phrase
16169 is prompted for as usual.
16173 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16174 the support is automatically enabled. The resulting binaries will
16175 autodetect the card and use it if present.
16177 *Ben Laurie and Compaq Inc.*
16179 * Work around for Netscape hang bug. This sends certificate request
16180 and server done in one record. Since this is perfectly legal in the
16181 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16182 the bugs/SSLv3 entry for more info.
16186 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16190 * Add -rand argument to smime and pkcs12 applications and read/write
16195 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16199 * Add command line password options to the remaining applications.
16203 * Bug fix for BN_div_recp() for numerators with an even number of
16208 * More tests in bntest.c, and changed test_bn output.
16212 * ./config recognizes MacOS X now.
16216 * Bug fix for BN_div() when the first words of num and divisor are
16217 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16221 * Add support for various broken PKCS#8 formats, and command line
16222 options to produce them.
16226 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16227 get temporary BIGNUMs from a BN_CTX.
16231 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16236 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16237 include a #define from the old name to the new. The original intent
16238 was that statically linked binaries could for example just call
16239 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16240 link with digests. This never worked because SSLeay_add_all_digests()
16241 and SSLeay_add_all_ciphers() were in the same source file so calling
16242 one would link with the other. They are now in separate source files.
16246 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16250 * Use a less unusual form of the Miller-Rabin primality test (it used
16251 a binary algorithm for exponentiation integrated into the Miller-Rabin
16252 loop, our standard modexp algorithms are faster).
16256 * Support for the EBCDIC character set completed.
16258 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16260 * Source code cleanups: use const where appropriate, eliminate casts,
16261 use `void *` instead of `char *` in lhash.
16265 * Bugfix: ssl3_send_server_key_exchange was not restartable
16266 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16267 this the server could overwrite ephemeral keys that the client
16272 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16273 using 50 iterations of the Rabin-Miller test.
16275 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16276 iterations of the Rabin-Miller test as required by the appendix
16277 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16278 As BN_is_prime_fasttest includes trial division, DSA parameter
16279 generation becomes much faster.
16281 This implies a change for the callback functions in DSA_is_prime
16282 and DSA_generate_parameters: The callback function is called once
16283 for each positive witness in the Rabin-Miller test, not just
16284 occasionally in the inner loop; and the parameters to the
16285 callback function now provide an iteration count for the outer
16286 loop rather than for the current invocation of the inner loop.
16287 DSA_generate_parameters additionally can call the callback
16288 function with an 'iteration count' of -1, meaning that a
16289 candidate has passed the trial division test (when q is generated
16290 from an application-provided seed, trial division is skipped).
16294 * New function BN_is_prime_fasttest that optionally does trial
16295 division before starting the Rabin-Miller test and has
16296 an additional BN_CTX * argument (whereas BN_is_prime always
16297 has to allocate at least one BN_CTX).
16298 'callback(1, -1, cb_arg)' is called when a number has passed the
16299 trial division stage.
16303 * Fix for bug in CRL encoding. The validity dates weren't being handled
16308 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16312 * New function BN_pseudo_rand().
16316 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16317 bignum version of BN_from_montgomery() with the working code from
16318 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16323 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16324 made it impossible to use the same SSL_SESSION data structure in
16325 SSL2 clients in multiple threads.
16329 * The return value of RAND_load_file() no longer counts bytes obtained
16330 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16331 to seed the PRNG (previously an explicit byte count was required).
16333 *Ulf Möller, Bodo Möller*
16335 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16336 used `char *` instead of `void *` and had casts all over the place.
16340 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16344 * Retain source code compatibility for BN_prime_checks macro:
16345 BN_is_prime(..., BN_prime_checks, ...) now uses
16346 BN_prime_checks_for_size to determine the appropriate number of
16347 Rabin-Miller iterations.
16351 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16352 DH_CHECK_P_NOT_SAFE_PRIME.
16353 (Check if this is true? OpenPGP calls them "strong".)
16357 * Merge the functionality of "dh" and "gendh" programs into a new program
16358 "dhparam". The old programs are retained for now but will handle DH keys
16359 (instead of parameters) in future.
16363 * Make the ciphers, s_server and s_client programs check the return values
16364 when a new cipher list is set.
16368 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16369 ciphers. Before when the 56bit ciphers were enabled the sorting was
16372 The syntax for the cipher sorting has been extended to support sorting by
16373 cipher-strength (using the strength_bits hard coded in the tables).
16374 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16376 Fix a bug in the cipher-command parser: when supplying a cipher command
16377 string with an "undefined" symbol (neither command nor alphanumeric
16378 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16379 an error is flagged.
16381 Due to the strength-sorting extension, the code of the
16382 ssl_create_cipher_list() function was completely rearranged. I hope that
16383 the readability was also increased :-)
16385 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16387 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16388 for the first serial number and places 2 in the serial number file. This
16389 avoids problems when the root CA is created with serial number zero and
16390 the first user certificate has the same issuer name and serial number
16395 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16396 the new code. Add documentation for this stuff.
16400 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16401 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16402 structures and behave in an analogous way to the X509v3 functions:
16403 they shouldn't be called directly but wrapper functions should be used
16406 So we also now have some wrapper functions that call the X509at functions
16407 when passed certificate requests. (TO DO: similar things can be done with
16408 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16409 things. Some of these need some d2i or i2d and print functionality
16410 because they handle more complex structures.)
16414 * Add missing #ifndefs that caused missing symbols when building libssl
16415 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16416 NO_RSA in `ssl/s2*.c`.
16418 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16420 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16421 has a return value which indicates the quality of the random data
16422 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16423 error queue. New function RAND_pseudo_bytes() generates output that is
16424 guaranteed to be unique but not unpredictable. RAND_add is like
16425 RAND_seed, but takes an extra argument for an entropy estimate
16426 (RAND_seed always assumes full entropy).
16430 * Do more iterations of Rabin-Miller probable prime test (specifically,
16431 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16432 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16433 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16434 false-positive rate of at most 2^-80 for random input.
16438 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16442 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16443 in the 0.9.5 release), this returns the chain
16444 from an X509_CTX structure with a dup of the stack and all
16445 the X509 reference counts upped: so the stack will exist
16446 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16449 Also make SSL_SESSION_print() print out the verify return
16454 * Add manpage for the pkcs12 command. Also change the default
16455 behaviour so MAC iteration counts are used unless the new
16456 -nomaciter option is used. This improves file security and
16457 only older versions of MSIE (4.0 for example) need it.
16461 * Honor the no-xxx Configure options when creating .DEF files.
16465 * Add PKCS#10 attributes to field table: challengePassword,
16466 unstructuredName and unstructuredAddress. These are taken from
16467 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16468 international characters are used.
16470 More changes to X509_ATTRIBUTE code: allow the setting of types
16471 based on strings. Remove the 'loc' parameter when adding
16472 attributes because these will be a SET OF encoding which is sorted
16477 * Initial changes to the 'req' utility to allow request generation
16478 automation. This will allow an application to just generate a template
16479 file containing all the field values and have req construct the
16482 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16483 used all over the place including certificate requests and PKCS#7
16484 structures. They are currently handled manually where necessary with
16485 some primitive wrappers for PKCS#7. The new functions behave in a
16486 manner analogous to the X509 extension functions: they allow
16487 attributes to be looked up by NID and added.
16489 Later something similar to the X509V3 code would be desirable to
16490 automatically handle the encoding, decoding and printing of the
16491 more complex types. The string types like challengePassword can
16492 be handled by the string table functions.
16494 Also modified the multi byte string table handling. Now there is
16495 a 'global mask' which masks out certain types. The table itself
16496 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16497 is useful when for example there is only one permissible type
16498 (as in countryName) and using the mask might result in no valid
16503 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16504 SSL_get_peer_finished to allow applications to obtain the latest
16505 Finished messages sent to the peer or expected from the peer,
16506 respectively. (SSL_get_peer_finished is usually the Finished message
16507 actually received from the peer, otherwise the protocol will be aborted.)
16509 As the Finished message are message digests of the complete handshake
16510 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16511 be used for external authentication procedures when the authentication
16512 provided by SSL/TLS is not desired or is not enough.
16516 * Enhanced support for Alpha Linux is added. Now ./config checks if
16517 the host supports BWX extension and if Compaq C is present on the
16518 $PATH. Just exploiting of the BWX extension results in 20-30%
16519 performance kick for some algorithms, e.g. DES and RC4 to mention
16520 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16525 * Add support for MS "fast SGC". This is arguably a violation of the
16526 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16527 weak crypto and after checking the certificate is SGC a second one
16528 with strong crypto. MS SGC stops the first handshake after receiving
16529 the server certificate message and sends a second client hello. Since
16530 a server will typically do all the time consuming operations before
16531 expecting any further messages from the client (server key exchange
16532 is the most expensive) there is little difference between the two.
16534 To get OpenSSL to support MS SGC we have to permit a second client
16535 hello message after we have sent server done. In addition we have to
16536 reset the MAC if we do get this second client hello.
16540 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16541 if a DER encoded private key is RSA or DSA traditional format. Changed
16542 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16543 format DER encoded private key. Newer code should use PKCS#8 format which
16544 has the key type encoded in the ASN1 structure. Added DER private key
16545 support to pkcs8 application.
16549 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16550 ciphersuites has been selected (as required by the SSL 3/TLS 1
16551 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16552 is set, we interpret this as a request to violate the specification
16553 (the worst that can happen is a handshake failure, and 'correct'
16554 behaviour would result in a handshake failure anyway).
16558 * In SSL_CTX_add_session, take into account that there might be multiple
16559 SSL_SESSION structures with the same session ID (e.g. when two threads
16560 concurrently obtain them from an external cache).
16561 The internal cache can handle only one SSL_SESSION with a given ID,
16562 so if there's a conflict, we now throw out the old one to achieve
16567 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16568 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16569 some routines that use cipher OIDs: some ciphers do not have OIDs
16570 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16575 * Simplify the trust setting structure and code. Now we just have
16576 two sequences of OIDs for trusted and rejected settings. These will
16577 typically have values the same as the extended key usage extension
16578 and any application specific purposes.
16580 The trust checking code now has a default behaviour: it will just
16581 check for an object with the same NID as the passed id. Functions can
16582 be provided to override either the default behaviour or the behaviour
16583 for a given id. SSL client, server and email already have functions
16584 in place for compatibility: they check the NID and also return "trusted"
16585 if the certificate is self signed.
16589 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16590 traditional format into an EVP_PKEY structure.
16594 * Add a password callback function PEM_cb() which either prompts for
16595 a password if usr_data is NULL or otherwise assumes it is a null
16596 terminated password. Allow passwords to be passed on command line
16597 environment or config files in a few more utilities.
16601 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16602 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16603 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16604 Update documentation.
16608 * Support for ASN1 "NULL" type. This could be handled before by using
16609 ASN1_TYPE but there wasn't any function that would try to read a NULL
16610 and produce an error if it couldn't. For compatibility we also have
16611 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16612 don't allocate anything because they don't need to.
16616 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16619 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16621 * Rebuild of the memory allocation routines used by OpenSSL code and
16622 possibly others as well. The purpose is to make an interface that
16623 provide hooks so anyone can build a separate set of allocation and
16624 deallocation routines to be used by OpenSSL, for example memory
16625 pool implementations, or something else, which was previously hard
16626 since Malloc(), Realloc() and Free() were defined as macros having
16627 the values malloc, realloc and free, respectively (except for Win32
16628 compilations). The same is provided for memory debugging code.
16629 OpenSSL already comes with functionality to find memory leaks, but
16630 this gives people a chance to debug other memory problems.
16632 With these changes, a new set of functions and macros have appeared:
16634 CRYPTO_set_mem_debug_functions() [F]
16635 CRYPTO_get_mem_debug_functions() [F]
16636 CRYPTO_dbg_set_options() [F]
16637 CRYPTO_dbg_get_options() [F]
16638 CRYPTO_malloc_debug_init() [M]
16640 The memory debug functions are NULL by default, unless the library
16641 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16642 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16643 gives the standard debugging functions that come with OpenSSL) or
16644 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16645 provided by the library user) must be used. When the standard
16646 debugging functions are used, CRYPTO_dbg_set_options can be used to
16647 request additional information:
16648 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16649 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16651 Also, things like CRYPTO_set_mem_functions will always give the
16652 expected result (the new set of functions is used for allocation
16653 and deallocation) at all times, regardless of platform and compiler
16656 To finish it up, some functions that were never use in any other
16657 way than through macros have a new API and new semantic:
16659 CRYPTO_dbg_malloc()
16660 CRYPTO_dbg_realloc()
16663 All macros of value have retained their old syntax.
16665 *Richard Levitte and Bodo Moeller*
16667 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16668 ordering of SMIMECapabilities wasn't in "strength order" and there
16669 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16674 * Some ASN1 types with illegal zero length encoding (INTEGER,
16675 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16677 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16679 * Merge in my S/MIME library for OpenSSL. This provides a simple
16680 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16681 functionality to handle multipart/signed properly) and a utility
16682 called 'smime' to call all this stuff. This is based on code I
16683 originally wrote for Celo who have kindly allowed it to be
16684 included in OpenSSL.
16688 * Add variants des_set_key_checked and des_set_key_unchecked of
16689 des_set_key (aka des_key_sched). Global variable des_check_key
16690 decides which of these is called by des_set_key; this way
16691 des_check_key behaves as it always did, but applications and
16692 the library itself, which was buggy for des_check_key == 1,
16693 have a cleaner way to pick the version they need.
16697 * New function PKCS12_newpass() which changes the password of a
16702 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16703 dynamic mix. In both cases the ids can be used as an index into the
16704 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16705 functions so they accept a list of the field values and the
16706 application doesn't need to directly manipulate the X509_TRUST
16711 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16716 * Modify the way the V3 extension code looks up extensions. This now
16717 works in a similar way to the object code: we have some "standard"
16718 extensions in a static table which is searched with OBJ_bsearch()
16719 and the application can add dynamic ones if needed. The file
16720 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16721 updated whenever a new extension is added to the core code and kept
16722 in ext_nid order. There is a simple program 'tabtest.c' which checks
16723 this. New extensions are not added too often so this file can readily
16724 be maintained manually.
16726 There are two big advantages in doing things this way. The extensions
16727 can be looked up immediately and no longer need to be "added" using
16728 X509V3_add_standard_extensions(): this function now does nothing.
16729 Side note: I get *lots* of email saying the extension code doesn't
16730 work because people forget to call this function.
16731 Also no dynamic allocation is done unless new extensions are added:
16732 so if we don't add custom extensions there is no need to call
16733 X509V3_EXT_cleanup().
16737 * Modify enc utility's salting as follows: make salting the default. Add a
16738 magic header, so unsalted files fail gracefully instead of just decrypting
16739 to garbage. This is because not salting is a big security hole, so people
16740 should be discouraged from doing it.
16744 * Fixes and enhancements to the 'x509' utility. It allowed a message
16745 digest to be passed on the command line but it only used this
16746 parameter when signing a certificate. Modified so all relevant
16747 operations are affected by the digest parameter including the
16748 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16749 DSA key was used because it didn't fix the digest.
16753 * Initial certificate chain verify code. Currently tests the untrusted
16754 certificates for consistency with the verify purpose (which is set
16755 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16757 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16758 this is because it will reject chains with invalid extensions whereas
16759 every previous version of OpenSSL and SSLeay made no checks at all.
16761 Trust code: checks the root CA for the relevant trust settings. Trust
16762 settings have an initial value consistent with the verify purpose: e.g.
16763 if the verify purpose is for SSL client use it expects the CA to be
16764 trusted for SSL client use. However the default value can be changed to
16765 permit custom trust settings: one example of this would be to only trust
16766 certificates from a specific "secure" set of CAs.
16768 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16769 which should be used for version portability: especially since the
16770 verify structure is likely to change more often now.
16772 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16773 to set them. If not set then assume SSL clients will verify SSL servers
16776 Two new options to the verify program: -untrusted allows a set of
16777 untrusted certificates to be passed in and -purpose which sets the
16778 intended purpose of the certificate. If a purpose is set then the
16779 new chain verify code is used to check extension consistency.
16783 * Support for the authority information access extension.
16787 * Modify RSA and DSA PEM read routines to transparently handle
16788 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16789 public keys in a format compatible with certificate
16790 SubjectPublicKeyInfo structures. Unfortunately there were already
16791 functions called *_PublicKey_* which used various odd formats so
16792 these are retained for compatibility: however the DSA variants were
16793 never in a public release so they have been deleted. Changed dsa/rsa
16794 utilities to handle the new format: note no releases ever handled public
16795 keys so we should be OK.
16797 The primary motivation for this change is to avoid the same fiasco
16798 that dogs private keys: there are several incompatible private key
16799 formats some of which are standard and some OpenSSL specific and
16800 require various evil hacks to allow partial transparent handling and
16801 even then it doesn't work with DER formats. Given the option anything
16802 other than PKCS#8 should be dumped: but the other formats have to
16803 stay in the name of compatibility.
16805 With public keys and the benefit of hindsight one standard format
16806 is used which works with EVP_PKEY, RSA or DSA structures: though
16807 it clearly returns an error if you try to read the wrong kind of key.
16809 Added a -pubkey option to the 'x509' utility to output the public key.
16810 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16811 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16812 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16813 that do the same as the `EVP_PKEY_assign_*()` except they up the
16814 reference count of the added key (they don't "swallow" the
16819 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16820 CRLs would fail if the file contained no certificates or no CRLs:
16821 added a new function to read in both types and return the number
16822 read: this means that if none are read it will be an error. The
16823 DER versions of the certificate and CRL reader would always fail
16824 because it isn't possible to mix certificates and CRLs in DER format
16825 without choking one or the other routine. Changed this to just read
16826 a certificate: this is the best we can do. Also modified the code
16827 in `apps/verify.c` to take notice of return codes: it was previously
16828 attempting to read in certificates from NULL pointers and ignoring
16829 any errors: this is one reason why the cert and CRL reader seemed
16830 to work. It doesn't check return codes from the default certificate
16831 routines: these may well fail if the certificates aren't installed.
16835 * Code to support otherName option in GeneralName.
16839 * First update to verify code. Change the verify utility
16840 so it warns if it is passed a self signed certificate:
16841 for consistency with the normal behaviour. X509_verify
16842 has been modified to it will now verify a self signed
16843 certificate if *exactly* the same certificate appears
16844 in the store: it was previously impossible to trust a
16845 single self signed certificate. This means that:
16846 openssl verify ss.pem
16847 now gives a warning about a self signed certificate but
16848 openssl verify -CAfile ss.pem ss.pem
16853 * For servers, store verify_result in SSL_SESSION data structure
16854 (and add it to external session representation).
16855 This is needed when client certificate verifications fails,
16856 but an application-provided verification callback (set by
16857 SSL_CTX_set_cert_verify_callback) allows accepting the session
16858 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16859 but returns 1): When the session is reused, we have to set
16860 ssl->verify_result to the appropriate error code to avoid
16863 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16865 * Fix a bug in the new PKCS#7 code: it didn't consider the
16866 case in PKCS7_dataInit() where the signed PKCS7 structure
16867 didn't contain any existing data because it was being created.
16869 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16871 * Add a salt to the key derivation routines in enc.c. This
16872 forms the first 8 bytes of the encrypted file. Also add a
16873 -S option to allow a salt to be input on the command line.
16877 * New function X509_cmp(). Oddly enough there wasn't a function
16878 to compare two certificates. We do this by working out the SHA1
16879 hash and comparing that. X509_cmp() will be needed by the trust
16884 * SSL_get1_session() is like SSL_get_session(), but increments
16885 the reference count in the SSL_SESSION returned.
16887 *Geoff Thorpe <geoff@eu.c2.net>*
16889 * Fix for 'req': it was adding a null to request attributes.
16890 Also change the X509_LOOKUP and X509_INFO code to handle
16891 certificate auxiliary information.
16895 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16900 * Add the possibility to add extra information to the memory leak
16901 detecting output, to form tracebacks, showing from where each
16902 allocation was originated: CRYPTO_push_info("constant string") adds
16903 the string plus current file name and line number to a per-thread
16904 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16905 is like calling CYRPTO_pop_info() until the stack is empty.
16906 Also updated memory leak detection code to be multi-thread-safe.
16910 * Add options -text and -noout to pkcs7 utility and delete the
16911 encryption options which never did anything. Update docs.
16915 * Add options to some of the utilities to allow the pass phrase
16916 to be included on either the command line (not recommended on
16917 OSes like Unix) or read from the environment. Update the
16918 manpages and fix a few bugs.
16922 * Add a few manpages for some of the openssl commands.
16926 * Fix the -revoke option in ca. It was freeing up memory twice,
16927 leaking and not finding already revoked certificates.
16931 * Extensive changes to support certificate auxiliary information.
16932 This involves the use of X509_CERT_AUX structure and X509_AUX
16933 functions. An X509_AUX function such as PEM_read_X509_AUX()
16934 can still read in a certificate file in the usual way but it
16935 will also read in any additional "auxiliary information". By
16936 doing things this way a fair degree of compatibility can be
16937 retained: existing certificates can have this information added
16938 using the new 'x509' options.
16940 Current auxiliary information includes an "alias" and some trust
16941 settings. The trust settings will ultimately be used in enhanced
16942 certificate chain verification routines: currently a certificate
16943 can only be trusted if it is self signed and then it is trusted
16948 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16949 The problem was that one of the replacement routines had not been working
16950 since SSLeay releases. For now the offending routine has been replaced
16951 with non-optimised assembler. Even so, this now gives around 95%
16952 performance improvement for 1024 bit RSA signs.
16956 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16957 handling. Most clients have the effective key size in bits equal to
16958 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16959 A few however don't do this and instead use the size of the decrypted key
16960 to determine the RC2 key length and the AlgorithmIdentifier to determine
16961 the effective key length. In this case the effective key length can still
16962 be 40 bits but the key length can be 168 bits for example. This is fixed
16963 by manually forcing an RC2 key into the EVP_PKEY structure because the
16964 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16965 the key length and effective key length are equal.
16969 * Add a bunch of functions that should simplify the creation of
16970 X509_NAME structures. Now you should be able to do:
16971 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16972 and have it automatically work out the correct field type and fill in
16973 the structures. The more adventurous can try:
16974 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16975 and it will (hopefully) work out the correct multibyte encoding.
16979 * Change the 'req' utility to use the new field handling and multibyte
16980 copy routines. Before the DN field creation was handled in an ad hoc
16981 way in req, ca, and x509 which was rather broken and didn't support
16982 BMPStrings or UTF8Strings. Since some software doesn't implement
16983 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16984 using the dirstring_type option. See the new comment in the default
16985 openssl.cnf for more info.
16989 * Make crypto/rand/md_rand.c more robust:
16990 - Assure unique random numbers after fork().
16991 - Make sure that concurrent threads access the global counter and
16992 md serializably so that we never lose entropy in them
16993 or use exactly the same state in multiple threads.
16994 Access to the large state is not always serializable because
16995 the additional locking could be a performance killer, and
16996 md should be large enough anyway.
17000 * New file `apps/app_rand.c` with commonly needed functionality
17001 for handling the random seed file.
17003 Use the random seed file in some applications that previously did not:
17005 dsaparam -genkey (which also ignored its '-rand' option),
17008 x509 (when signing).
17009 Except on systems with /dev/urandom, it is crucial to have a random
17010 seed file at least for key creation, DSA signing, and for DH exchanges;
17011 for RSA signatures we could do without one.
17013 gendh and gendsa (unlike genrsa) used to read only the first byte
17014 of each file listed in the '-rand' option. The function as previously
17015 found in genrsa is now in app_rand.c and is used by all programs
17016 that support '-rand'.
17020 * In RAND_write_file, use mode 0600 for creating files;
17021 don't just chmod when it may be too late.
17025 * Report an error from X509_STORE_load_locations
17026 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17030 * New function ASN1_mbstring_copy() this copies a string in either
17031 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17032 into an ASN1_STRING type. A mask of permissible types is passed
17033 and it chooses the "minimal" type to use or an error if not type
17038 * Add function equivalents to the various macros in asn1.h. The old
17039 macros are retained with an `M_` prefix. Code inside the library can
17040 use the `M_` macros. External code (including the openssl utility)
17041 should *NOT* in order to be "shared library friendly".
17045 * Add various functions that can check a certificate's extensions
17046 to see if it usable for various purposes such as SSL client,
17047 server or S/MIME and CAs of these types. This is currently
17048 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17049 verification. Also added a -purpose flag to x509 utility to
17050 print out all the purposes.
17054 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17059 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17060 for, obtain and decode and extension and obtain its critical flag.
17061 This allows all the necessary extension code to be handled in a
17062 single function call.
17066 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17067 platforms. See crypto/rc4/rc4_enc.c for further details.
17071 * New -noout option to asn1parse. This causes no output to be produced
17072 its main use is when combined with -strparse and -out to extract data
17073 from a file (which may not be in ASN.1 format).
17077 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17078 when producing the local key id.
17080 *Richard Levitte <levitte@stacken.kth.se>*
17082 * New option -dhparam in s_server. This allows a DH parameter file to be
17083 stated explicitly. If it is not stated then it tries the first server
17084 certificate file. The previous behaviour hard coded the filename
17089 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17090 a public key to be input or output. For example:
17091 openssl rsa -in key.pem -pubout -out pubkey.pem
17092 Also added necessary DSA public key functions to handle this.
17096 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17097 in the message. This was handled by allowing
17098 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17100 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17102 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17103 to the end of the strings whereas this didn't. This would cause problems
17104 if strings read with d2i_ASN1_bytes() were later modified.
17106 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17108 * Fix for base64 decode bug. When a base64 bio reads only one line of
17109 data and it contains EOF it will end up returning an error. This is
17110 caused by input 46 bytes long. The cause is due to the way base64
17111 BIOs find the start of base64 encoded data. They do this by trying a
17112 trial decode on each line until they find one that works. When they
17113 do a flag is set and it starts again knowing it can pass all the
17114 data directly through the decoder. Unfortunately it doesn't reset
17115 the context it uses. This means that if EOF is reached an attempt
17116 is made to pass two EOFs through the context and this causes the
17117 resulting error. This can also cause other problems as well. As is
17118 usual with these problems it takes *ages* to find and the fix is
17119 trivial: move one line.
17121 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17123 * Ugly workaround to get s_client and s_server working under Windows. The
17124 old code wouldn't work because it needed to select() on sockets and the
17125 tty (for keypresses and to see if data could be written). Win32 only
17126 supports select() on sockets so we select() with a 1s timeout on the
17127 sockets and then see if any characters are waiting to be read, if none
17128 are present then we retry, we also assume we can always write data to
17129 the tty. This isn't nice because the code then blocks until we've
17130 received a complete line of data and it is effectively polling the
17131 keyboard at 1s intervals: however it's quite a bit better than not
17132 working at all :-) A dedicated Windows application might handle this
17133 with an event loop for example.
17137 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17138 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17139 will be called when RSA_sign() and RSA_verify() are used. This is useful
17140 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17141 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17142 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17143 This necessitated the support of an extra signature type NID_md5_sha1
17144 for SSL signatures and modifications to the SSL library to use it instead
17145 of calling RSA_public_decrypt() and RSA_private_encrypt().
17149 * Add new -verify -CAfile and -CApath options to the crl program, these
17150 will lookup a CRL issuers certificate and verify the signature in a
17151 similar way to the verify program. Tidy up the crl program so it
17152 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17153 less strict. It will now permit CRL extensions even if it is not
17154 a V2 CRL: this will allow it to tolerate some broken CRLs.
17158 * Initialize all non-automatic variables each time one of the openssl
17159 sub-programs is started (this is necessary as they may be started
17160 multiple times from the "OpenSSL>" prompt).
17162 *Lennart Bang, Bodo Moeller*
17164 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17165 removing all other RSA functionality (this is what NO_RSA does). This
17166 is so (for example) those in the US can disable those operations covered
17167 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17172 * Non-copying interface to BIO pairs.
17173 (still largely untested)
17177 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17178 ASCII string. This was handled independently in various places before.
17182 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17183 UTF8 strings a character at a time.
17187 * Use client_version from client hello to select the protocol
17188 (s23_srvr.c) and for RSA client key exchange verification
17189 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17193 * Add various utility functions to handle SPKACs, these were previously
17194 handled by poking round in the structure internals. Added new function
17195 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17196 print, verify and generate SPKACs. Based on an original idea from
17197 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17201 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17205 * Allow the config file extension section to be overwritten on the
17206 command line. Based on an original idea from Massimiliano Pala
17207 <madwolf@comune.modena.it>. The new option is called -extensions
17208 and can be applied to ca, req and x509. Also -reqexts to override
17209 the request extensions in req and -crlexts to override the crl extensions
17214 * Add new feature to the SPKAC handling in ca. Now you can include
17215 the same field multiple times by preceding it by "XXXX." for example:
17218 this is the same syntax as used in the req config file.
17222 * Allow certificate extensions to be added to certificate requests. These
17223 are specified in a 'req_extensions' option of the req section of the
17224 config file. They can be printed out with the -text option to req but
17225 are otherwise ignored at present.
17229 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17230 data read consists of only the final block it would not decrypted because
17231 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17232 A misplaced 'break' also meant the decrypted final block might not be
17233 copied until the next read.
17237 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17238 a few extra parameters to the DH structure: these will be useful if
17239 for example we want the value of 'q' or implement X9.42 DH.
17243 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17244 provides hooks that allow the default DSA functions or functions on a
17245 "per key" basis to be replaced. This allows hardware acceleration and
17246 hardware key storage to be handled without major modification to the
17247 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17248 associated functions.
17252 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17253 as "read only": it can't be written to and the buffer it points to will
17254 not be freed. Reading from a read only BIO is much more efficient than
17255 a normal memory BIO. This was added because there are several times when
17256 an area of memory needs to be read from a BIO. The previous method was
17257 to create a memory BIO and write the data to it, this results in two
17258 copies of the data and an O(n^2) reading algorithm. There is a new
17259 function BIO_new_mem_buf() which creates a read only memory BIO from
17260 an area of memory. Also modified the PKCS#7 routines to use read only
17265 * Bugfix: ssl23_get_client_hello did not work properly when called in
17266 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17267 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17268 but a retry condition occurred while trying to read the rest.
17272 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17273 NID_pkcs7_encrypted by default: this was wrong since this should almost
17274 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17275 the encrypted data type: this is a more sensible place to put it and it
17276 allows the PKCS#12 code to be tidied up that duplicated this
17281 * Changed obj_dat.pl script so it takes its input and output files on
17282 the command line. This should avoid shell escape redirection problems
17287 * Initial support for certificate extension requests, these are included
17288 in things like Xenroll certificate requests. Included functions to allow
17289 extensions to be obtained and added.
17293 * -crlf option to s_client and s_server for sending newlines as
17294 CRLF (as required by many protocols).
17298 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17300 * Install libRSAglue.a when OpenSSL is built with RSAref.
17302 *Ralf S. Engelschall*
17304 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17306 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17308 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17313 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17314 DH parameters/keys (q is lost during that conversion, but the resulting
17315 DH parameters contain its length).
17317 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17318 much faster than DH_generate_parameters (which creates parameters
17319 where `p = 2*q + 1`), and also the smaller q makes DH computations
17320 much more efficient (160-bit exponentiation instead of 1024-bit
17321 exponentiation); so this provides a convenient way to support DHE
17322 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17323 utter importance to use
17324 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17326 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17327 when such DH parameters are used, because otherwise small subgroup
17328 attacks may become possible!
17332 * Avoid memory leak in i2d_DHparams.
17336 * Allow the -k option to be used more than once in the enc program:
17337 this allows the same encrypted message to be read by multiple recipients.
17341 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17342 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17343 it will always use the numerical form of the OID, even if it has a short
17348 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17349 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17350 otherwise bn_mod_exp was called. In the case of hardware keys for example
17351 no private key components need be present and it might store extra data
17352 in the RSA structure, which cannot be accessed from bn_mod_exp.
17353 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17354 private key operations.
17358 * Added support for SPARC Linux.
17362 * pem_password_cb function type incompatibly changed from
17363 typedef int pem_password_cb(char *buf, int size, int rwflag);
17365 ....(char *buf, int size, int rwflag, void *userdata);
17366 so that applications can pass data to their callbacks:
17367 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17368 additional void * argument, which is just handed through whenever
17369 the password callback is called.
17371 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17373 New function SSL_CTX_set_default_passwd_cb_userdata.
17375 Compatibility note: As many C implementations push function arguments
17376 onto the stack in reverse order, the new library version is likely to
17377 interoperate with programs that have been compiled with the old
17378 pem_password_cb definition (PEM_whatever takes some data that
17379 happens to be on the stack as its last argument, and the callback
17380 just ignores this garbage); but there is no guarantee whatsoever that
17383 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17384 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17385 problems not only on Windows, but also on some Unix platforms.
17386 To avoid problematic command lines, these definitions are now in an
17387 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17388 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17392 * MIPS III/IV assembler module is reimplemented.
17396 * More DES library cleanups: remove references to srand/rand and
17397 delete an unused file.
17401 * Add support for the free Netwide assembler (NASM) under Win32,
17402 since not many people have MASM (ml) and it can be hard to obtain.
17403 This is currently experimental but it seems to work OK and pass all
17404 the tests. Check out INSTALL.W32 for info.
17408 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17409 without temporary keys kept an extra copy of the server key,
17410 and connections with temporary keys did not free everything in case
17415 * New function RSA_check_key and new openssl rsa option -check
17416 for verifying the consistency of RSA keys.
17418 *Ulf Moeller, Bodo Moeller*
17420 * Various changes to make Win32 compile work:
17421 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17422 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17423 comparison" warnings.
17424 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17428 * Add a debugging option to PKCS#5 v2 key generation function: when
17429 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17430 derived keys are printed to stderr.
17434 * Copy the flags in ASN1_STRING_dup().
17436 *Roman E. Pavlov <pre@mo.msk.ru>*
17438 * The x509 application mishandled signing requests containing DSA
17439 keys when the signing key was also DSA and the parameters didn't match.
17441 It was supposed to omit the parameters when they matched the signing key:
17442 the verifying software was then supposed to automatically use the CA's
17443 parameters if they were absent from the end user certificate.
17445 Omitting parameters is no longer recommended. The test was also
17446 the wrong way round! This was probably due to unusual behaviour in
17447 EVP_cmp_parameters() which returns 1 if the parameters match.
17448 This meant that parameters were omitted when they *didn't* match and
17449 the certificate was useless. Certificates signed with 'ca' didn't have
17452 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17454 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17455 The interface is as follows:
17456 Applications can use
17457 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17458 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17459 "off" is now the default.
17460 The library internally uses
17461 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17462 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17463 to disable memory-checking temporarily.
17465 Some inconsistent states that previously were possible (and were
17466 even the default) are now avoided.
17468 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17469 with each memory chunk allocated; this is occasionally more helpful
17470 than just having a counter.
17472 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17474 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17479 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17480 which largely parallels "options", but is for changing API behaviour,
17481 whereas "options" are about protocol behaviour.
17482 Initial "mode" flags are:
17484 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17485 a single record has been written.
17486 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17487 retries use the same buffer location.
17488 (But all of the contents must be
17493 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17496 * Fix problems with no-hmac etc.
17498 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17500 * New functions RSA_get_default_method(), RSA_set_method() and
17501 RSA_get_method(). These allows replacement of RSA_METHODs without having
17502 to mess around with the internals of an RSA structure.
17506 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17507 Also really enable memory leak checks in openssl.c and in some
17510 *Chad C. Mulligan, Bodo Moeller*
17512 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17513 up the length of negative integers. This has now been simplified to just
17514 store the length when it is first determined and use it later, rather
17515 than trying to keep track of where data is copied and updating it to
17517 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17519 * Add a new function PKCS7_signatureVerify. This allows the verification
17520 of a PKCS#7 signature but with the signing certificate passed to the
17521 function itself. This contrasts with PKCS7_dataVerify which assumes the
17522 certificate is present in the PKCS#7 structure. This isn't always the
17523 case: certificates can be omitted from a PKCS#7 structure and be
17524 distributed by "out of band" means (such as a certificate database).
17528 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17529 function prototypes in pem.h, also change util/mkdef.pl to add the
17530 necessary function names.
17534 * mk1mf.pl (used by Windows builds) did not properly read the
17535 options set by Configure in the top level Makefile, and Configure
17536 was not even able to write more than one option correctly.
17537 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17541 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17542 file to be loaded from a BIO or FILE pointer. The BIO version will
17543 for example allow memory BIOs to contain config info.
17547 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17548 Whoever hopes to achieve shared-library compatibility across versions
17549 must use this, not the compile-time macro.
17550 (Exercise 0.9.4: Which is the minimum library version required by
17552 Note: All this applies only to multi-threaded programs, others don't
17557 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17558 through a BIO pair triggered the default case, i.e.
17559 SSLerr(...,SSL_R_UNKNOWN_STATE).
17563 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17564 can use the SSL library even if none of the specific BIOs is
17569 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17570 for the encoded length.
17572 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17574 * Add initial documentation of the X509V3 functions.
17578 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17579 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17580 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17581 secure PKCS#8 private key format with a high iteration count.
17585 * Fix determination of Perl interpreter: A perl or perl5
17586 *directory* in $PATH was also accepted as the interpreter.
17588 *Ralf S. Engelschall*
17590 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17591 wrong with it but it was very old and did things like calling
17592 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17593 unusual formatting.
17597 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17598 to use the new extension code.
17602 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17603 with macros. This should make it easier to change their form, add extra
17604 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17609 * Add to configuration table a new entry that can specify an alternative
17610 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17611 according to Mark Crispin <MRC@Panda.COM>.
17615 * DES CBC did not update the IV. Weird.
17619 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17620 Changing the behaviour of the former might break existing programs --
17621 where IV updating is needed, des_ncbc_encrypt can be used.
17624 * When bntest is run from "make test" it drives bc to check its
17625 calculations, as well as internally checking them. If an internal check
17626 fails, it needs to cause bc to give a non-zero result or make test carries
17627 on without noticing the failure. Fixed.
17631 * DES library cleanups.
17635 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17636 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17637 ciphers. NOTE: although the key derivation function has been verified
17638 against some published test vectors it has not been extensively tested
17639 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17644 * Instead of "mkdir -p", which is not fully portable, use new
17645 Perl script "util/mkdir-p.pl".
17649 * Rewrite the way password based encryption (PBE) is handled. It used to
17650 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17651 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17652 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17653 the 'parameter' field of the AlgorithmIdentifier is passed to the
17654 underlying key generation function so it must do its own ASN1 parsing.
17655 This has also changed the EVP_PBE_CipherInit() function which now has a
17656 'parameter' argument instead of literal salt and iteration count values
17657 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17661 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17662 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17663 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17664 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17665 value was just used as a "magic string" and not used directly its
17666 value doesn't matter.
17670 * Introduce some semblance of const correctness to BN. Shame C doesn't
17675 * "linux-sparc64" configuration (ultrapenguin).
17677 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17678 "linux-sparc" configuration.
17680 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17682 * config now generates no-xxx options for missing ciphers.
17686 * Support the EBCDIC character set (work in progress).
17687 File ebcdic.c not yet included because it has a different license.
17689 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17691 * Support BS2000/OSD-POSIX.
17693 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17695 * Make callbacks for key generation use `void *` instead of `char *`.
17699 * Make S/MIME samples compile (not yet tested).
17703 * Additional typesafe stacks.
17707 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17711 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17713 * New configuration variant "sco5-gcc".
17715 * Updated some demos.
17717 *Sean O Riordain, Wade Scholine*
17719 * Add missing BIO_free at exit of pkcs12 application.
17723 * Fix memory leak in conf.c.
17727 * Updates for Win32 to assembler version of MD5.
17731 * Set #! path to perl in `apps/der_chop` to where we found it
17732 instead of using a fixed path.
17736 * SHA library changes for irix64-mips4-cc.
17740 * Improvements for VMS support.
17744 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17746 * Bignum library bug fix. IRIX 6 passes "make test" now!
17747 This also avoids the problems with SC4.2 and unpatched SC5.
17749 *Andy Polyakov <appro@fy.chalmers.se>*
17751 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17752 These are required because of the typesafe stack would otherwise break
17753 existing code. If old code used a structure member which used to be STACK
17754 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17755 sk_num or sk_value it would produce an error because the num, data members
17756 are not present in STACK_OF. Now it just produces a warning. sk_set
17757 replaces the old method of assigning a value to sk_value
17758 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17759 that does this will no longer work (and should use sk_set instead) but
17760 this could be regarded as a "questionable" behaviour anyway.
17764 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17765 correctly handle encrypted S/MIME data.
17769 * Change type of various DES function arguments from des_cblock
17770 (which means, in function argument declarations, pointer to char)
17771 to des_cblock * (meaning pointer to array with 8 char elements),
17772 which allows the compiler to do more typechecking; it was like
17773 that back in SSLeay, but with lots of ugly casts.
17775 Introduce new type const_des_cblock.
17779 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17780 problems: find RecipientInfo structure that matches recipient certificate
17781 and initialise the ASN1 structures properly based on passed cipher.
17785 * Belatedly make the BN tests actually check the results.
17789 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17790 to and from BNs: it was completely broken. New compilation option
17791 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17792 key elements as negative integers.
17796 * Reorganize and speed up MD5.
17798 *Andy Polyakov <appro@fy.chalmers.se>*
17802 *Richard Levitte <richard@levitte.org>*
17804 * New option -out to asn1parse to allow the parsed structure to be
17805 output to a file. This is most useful when combined with the -strparse
17806 option to examine the output of things like OCTET STRINGS.
17810 * Make SSL library a little more fool-proof by not requiring any longer
17811 that `SSL_set_{accept,connect}_state` be called before
17812 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17813 in many applications because usually everything *appeared* to work as
17814 intended anyway -- now it really works as intended).
17818 * Move openssl.cnf out of lib/.
17822 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17823 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17824 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17826 *Ralf S. Engelschall*
17828 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17829 handle PKCS#7 enveloped data properly.
17831 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17833 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17834 copying pointers. The cert_st handling is changed by this in
17835 various ways (and thus what used to be known as ctx->default_cert
17836 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17837 any longer when s->cert does not give us what we need).
17838 ssl_cert_instantiate becomes obsolete by this change.
17839 As soon as we've got the new code right (possibly it already is?),
17840 we have solved a couple of bugs of the earlier code where s->cert
17841 was used as if it could not have been shared with other SSL structures.
17843 Note that using the SSL API in certain dirty ways now will result
17844 in different behaviour than observed with earlier library versions:
17845 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17846 does not influence s as it used to.
17848 In order to clean up things more thoroughly, inside SSL_SESSION
17849 we don't use CERT any longer, but a new structure SESS_CERT
17850 that holds per-session data (if available); currently, this is
17851 the peer's certificate chain and, for clients, the server's certificate
17852 and temporary key. CERT holds only those values that can have
17853 meaningful defaults in an SSL_CTX.
17857 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17858 from the internal representation. Various PKCS#7 fixes: remove some
17859 evil casts and set the enc_dig_alg field properly based on the signing
17864 * Allow PKCS#12 password to be set from the command line or the
17865 environment. Let 'ca' get its config file name from the environment
17866 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17871 * Allow certificate policies extension to use an IA5STRING for the
17872 organization field. This is contrary to the PKIX definition but
17873 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17878 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17879 without disallowing inline assembler and the like for non-pedantic builds.
17883 * Support Borland C++ builder.
17885 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17891 * SHA-1 cleanups and performance enhancements.
17893 *Andy Polyakov <appro@fy.chalmers.se>*
17895 * Sparc v8plus assembler for the bignum library.
17897 *Andy Polyakov <appro@fy.chalmers.se>*
17899 * Accept any -xxx and +xxx compiler options in Configure.
17903 * Update HPUX configuration.
17907 * Add missing `sk_<type>_unshift()` function to safestack.h
17909 *Ralf S. Engelschall*
17911 * New function SSL_CTX_use_certificate_chain_file that sets the
17912 "extra_cert"s in addition to the certificate. (This makes sense
17913 only for "PEM" format files, as chains as a whole are not
17918 * Support verify_depth from the SSL API.
17919 x509_vfy.c had what can be considered an off-by-one-error:
17920 Its depth (which was not part of the external interface)
17921 was actually counting the number of certificates in a chain;
17922 now it really counts the depth.
17926 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17927 instead of X509err, which often resulted in confusing error
17928 messages since the error codes are not globally unique
17929 (e.g. an alleged error in ssl3_accept when a certificate
17930 didn't match the private key).
17932 * New function SSL_CTX_set_session_id_context that allows to set a default
17933 value (so that you don't need SSL_set_session_id_context for each
17934 connection using the SSL_CTX).
17938 * OAEP decoding bug fix.
17942 * Support INSTALL_PREFIX for package builders, as proposed by
17947 * New Configure options "threads" and "no-threads". For systems
17948 where the proper compiler options are known (currently Solaris
17949 and Linux), "threads" is the default.
17953 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17957 * Install various scripts to $(OPENSSLDIR)/misc, not to
17958 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17959 such as /usr/local/bin.
17963 * "make linux-shared" to build shared libraries.
17965 *Niels Poppe <niels@netbox.org>*
17967 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17971 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17972 extension adding in x509 utility.
17976 * Remove NOPROTO sections and error code comments.
17980 * Partial rewrite of the DEF file generator to now parse the ANSI
17985 * New Configure options --prefix=DIR and --openssldir=DIR.
17989 * Complete rewrite of the error code script(s). It is all now handled
17990 by one script at the top level which handles error code gathering,
17991 header rewriting and C source file generation. It should be much better
17992 than the old method: it now uses a modified version of Ulf's parser to
17993 read the ANSI prototypes in all header files (thus the old K&R definitions
17994 aren't needed for error creation any more) and do a better job of
17995 translating function codes into names. The old 'ASN1 error code embedded
17996 in a comment' is no longer necessary and it doesn't use .err files which
17997 have now been deleted. Also the error code call doesn't have to appear all
17998 on one line (which resulted in some large lines...).
18002 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18006 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18007 0 (which usually indicates a closed connection), but continue reading.
18011 * Fix some race conditions.
18015 * Add support for CRL distribution points extension. Add Certificate
18016 Policies and CRL distribution points documentation.
18020 * Move the autogenerated header file parts to crypto/opensslconf.h.
18024 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18025 8 of keying material. Merlin has also confirmed interop with this fix
18026 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18028 *Merlin Hughes <merlin@baltimore.ie>*
18030 * Fix lots of warnings.
18032 *Richard Levitte <levitte@stacken.kth.se>*
18034 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18035 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18037 *Richard Levitte <levitte@stacken.kth.se>*
18039 * Fix problems with sizeof(long) == 8.
18041 *Andy Polyakov <appro@fy.chalmers.se>*
18043 * Change functions to ANSI C.
18047 * Fix typos in error codes.
18049 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18051 * Remove defunct assembler files from Configure.
18055 * SPARC v8 assembler BIGNUM implementation.
18057 *Andy Polyakov <appro@fy.chalmers.se>*
18059 * Support for Certificate Policies extension: both print and set.
18060 Various additions to support the r2i method this uses.
18064 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18065 return a const string when you are expecting an allocated buffer.
18069 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18070 types DirectoryString and DisplayText.
18074 * Add code to allow r2i extensions to access the configuration database,
18075 add an LHASH database driver and add several ctx helper functions.
18079 * Fix an evil bug in bn_expand2() which caused various BN functions to
18080 fail when they extended the size of a BIGNUM.
18084 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18085 support typesafe stack.
18089 * Fix typo in SSL_[gs]et_options().
18091 *Nils Frostberg <nils@medcom.se>*
18093 * Delete various functions and files that belonged to the (now obsolete)
18094 old X509V3 handling code.
18098 * New Configure option "rsaref".
18102 * Don't auto-generate pem.h.
18106 * Introduce type-safe ASN.1 SETs.
18110 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18112 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18114 * Introduce type-safe STACKs. This will almost certainly break lots of code
18115 that links with OpenSSL (well at least cause lots of warnings), but fear
18116 not: the conversion is trivial, and it eliminates loads of evil casts. A
18117 few STACKed things have been converted already. Feel free to convert more.
18118 In the fullness of time, I'll do away with the STACK type altogether.
18122 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18123 specified in `<certfile>` by updating the entry in the index.txt file.
18124 This way one no longer has to edit the index.txt file manually for
18125 revoking a certificate. The -revoke option does the gory details now.
18127 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18129 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18130 `-text` option at all and this way the `-noout -text` combination was
18131 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18133 *Ralf S. Engelschall*
18135 * Make sure a corresponding plain text error message exists for the
18136 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18137 verify callback function determined that a certificate was revoked.
18139 *Ralf S. Engelschall*
18141 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18142 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18143 all available ciphers including rc5, which was forgotten until now.
18144 In order to let the testing shell script know which algorithms
18145 are available, a new (up to now undocumented) command
18146 `openssl list-cipher-commands` is used.
18150 * Bugfix: s_client occasionally would sleep in select() when
18151 it should have checked SSL_pending() first.
18155 * New functions DSA_do_sign and DSA_do_verify to provide access to
18156 the raw DSA values prior to ASN.1 encoding.
18160 * Tweaks to Configure
18162 *Niels Poppe <niels@netbox.org>*
18164 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18169 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18173 * New config option to avoid instructions that are illegal on the 80386.
18174 The default code is faster, but requires at least a 486.
18178 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18179 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18180 same as SSL2_VERSION anyway.
18184 * New "-showcerts" option for s_client.
18188 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18189 application. Various cleanups and fixes.
18193 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18194 modify error routines to work internally. Add error codes and PBE init
18195 to library startup routines.
18199 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18200 packing functions to asn1 and evp. Changed function names and error
18201 codes along the way.
18205 * PKCS12 integration: and so it begins... First of several patches to
18206 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18207 objects to objects.h
18211 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18212 and display support for Thawte strong extranet extension.
18216 * Add LinuxPPC support.
18218 *Jeff Dubrule <igor@pobox.org>*
18220 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18221 bn_div_words in alpha.s.
18223 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18225 * Make sure the RSA OAEP test is skipped under -DRSAref because
18226 OAEP isn't supported when OpenSSL is built with RSAref.
18228 *Ulf Moeller <ulf@fitug.de>*
18230 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18231 so they no longer are missing under -DNOPROTO.
18233 *Soren S. Jorvang <soren@t.dk>*
18235 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18237 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18238 doesn't work when the session is reused. Coming soon!
18242 * Fix a security hole, that allows sessions to be reused in the wrong
18243 context thus bypassing client cert protection! All software that uses
18244 client certs and session caches in multiple contexts NEEDS PATCHING to
18245 allow session reuse! A fuller solution is in the works.
18247 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18249 * Some more source tree cleanups (removed obsolete files
18250 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18251 permission on "config" script to be executable) and a fix for the INSTALL
18254 *Ulf Moeller <ulf@fitug.de>*
18256 * Remove some legacy and erroneous uses of malloc, free instead of
18259 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18261 * Make rsa_oaep_test return non-zero on error.
18263 *Ulf Moeller <ulf@fitug.de>*
18265 * Add support for native Solaris shared libraries. Configure
18266 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18267 if someone would make that last step automatic.
18269 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18271 * ctx_size was not built with the right compiler during "make links". Fixed.
18275 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18276 except NULL ciphers". This means the default cipher list will no longer
18277 enable NULL ciphers. They need to be specifically enabled e.g. with
18278 the string "DEFAULT:eNULL".
18282 * Fix to RSA private encryption routines: if p < q then it would
18283 occasionally produce an invalid result. This will only happen with
18284 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18288 * Be less restrictive and allow also `perl util/perlpath.pl
18289 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18290 because this way one can also use an interpreter named `perl5` (which is
18291 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18292 installed as `perl`).
18294 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18296 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18298 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18300 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18301 advapi32.lib to Win32 build and change the pem test comparison
18302 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18303 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18304 and crypto/des/ede_cbcm_enc.c.
18308 * DES quad checksum was broken on big-endian architectures. Fixed.
18312 * Comment out two functions in bio.h that aren't implemented. Fix up the
18313 Win32 test batch file so it (might) work again. The Win32 test batch file
18314 is horrible: I feel ill....
18318 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18319 in e_os.h. Audit of header files to check ANSI and non ANSI
18320 sections: 10 functions were absent from non ANSI section and not exported
18321 from Windows DLLs. Fixed up libeay.num for new functions.
18325 * Make `openssl version` output lines consistent.
18327 *Ralf S. Engelschall*
18329 * Fix Win32 symbol export lists for BIO functions: Added
18330 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18331 to ms/libeay{16,32}.def.
18333 *Ralf S. Engelschall*
18335 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18336 fine under Unix and passes some trivial tests I've now added. But the
18337 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18338 added to make sure no one expects that this stuff really works in the
18339 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18340 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18343 *Ralf S. Engelschall*
18345 * Fix the generation of two part addresses in perl.
18347 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18349 * Add config entry for Linux on MIPS.
18351 *John Tobey <jtobey@channel1.com>*
18353 * Make links whenever Configure is run, unless we are on Windoze.
18357 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18358 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18363 * Add a useful kludge to allow package maintainers to specify compiler and
18364 other platforms details on the command line without having to patch the
18365 Configure script every time: One now can use
18366 `perl Configure <id>:<details>`,
18367 i.e. platform ids are allowed to have details appended
18368 to them (separated by colons). This is treated as there would be a static
18369 pre-configured entry in Configure's %table under key `<id>` with value
18370 `<details>` and `perl Configure <id>` is called. So, when you want to
18371 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18372 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18373 now, which overrides the FreeBSD-elf entry on-the-fly.
18375 *Ralf S. Engelschall*
18377 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18381 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18382 on the `perl Configure ...` command line. This way one can compile
18383 OpenSSL libraries with Position Independent Code (PIC) which is needed
18384 for linking it into DSOs.
18386 *Ralf S. Engelschall*
18388 * Remarkably, export ciphers were totally broken and no-one had noticed!
18393 * Cleaned up the LICENSE document: The official contact for any license
18394 questions now is the OpenSSL core team under openssl-core@openssl.org.
18395 And add a paragraph about the dual-license situation to make sure people
18396 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18397 to the OpenSSL toolkit.
18399 *Ralf S. Engelschall*
18401 * General source tree makefile cleanups: Made `making xxx in yyy...`
18402 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18403 Additionally cleaned up the `make links` target: Remove unnecessary
18404 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18405 to speed processing and no longer clutter the display with confusing
18406 stuff. Instead only the actually done links are displayed.
18408 *Ralf S. Engelschall*
18410 * Permit null encryption ciphersuites, used for authentication only. It used
18411 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18412 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18417 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18418 signed attributes when verifying signatures (this would break them),
18419 the detached data encoding was wrong and public keys obtained using
18420 X509_get_pubkey() weren't freed.
18424 * Add text documentation for the BUFFER functions. Also added a work around
18425 to a Win95 console bug. This was triggered by the password read stuff: the
18426 last character typed gets carried over to the next fread(). If you were
18427 generating a new cert request using 'req' for example then the last
18428 character of the passphrase would be CR which would then enter the first
18433 * Added the new 'Includes OpenSSL Cryptography Software' button as
18434 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18435 button and can be used by applications based on OpenSSL to show the
18436 relationship to the OpenSSL project.
18438 *Ralf S. Engelschall*
18440 * Remove confusing variables in function signatures in files
18441 ssl/ssl_lib.c and ssl/ssl.h.
18443 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18445 * Don't install bss_file.c under PREFIX/include/
18447 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18449 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18450 functions that return function pointers and has support for NT specific
18451 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18452 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18453 unsigned to signed types: this was killing the Win32 compile.
18457 * Add new certificate file to stack functions,
18458 SSL_add_dir_cert_subjects_to_stack() and
18459 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18460 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18461 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18462 This means that Apache-SSL and similar packages don't have to mess around
18463 to add as many CAs as they want to the preferred list.
18467 * Experiment with doxygen documentation. Currently only partially applied to
18469 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18470 openssl.doxy as the configuration file.
18474 * Get rid of remaining C++-style comments which strict C compilers hate.
18476 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18478 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18479 compiled in by default: it has problems with large keys.
18483 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18484 DH private keys and/or callback functions which directly correspond to
18485 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18486 is needed for applications which have to configure certificates on a
18487 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18489 For the RSA certificate situation is makes no difference, but
18490 for the DSA certificate situation this fixes the "no shared cipher"
18491 problem where the OpenSSL cipher selection procedure failed because the
18492 temporary keys were not overtaken from the context and the API provided
18493 no way to reconfigure them.
18494 The new functions now let applications reconfigure the stuff and they
18495 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18496 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18497 non-public-API function ssl_cert_instantiate() is used as a helper
18498 function and also to reduce code redundancy inside ssl_rsa.c.
18500 *Ralf S. Engelschall*
18502 * Move s_server -dcert and -dkey options out of the undocumented feature
18503 area because they are useful for the DSA situation and should be
18504 recognized by the users.
18506 *Ralf S. Engelschall*
18508 * Fix the cipher decision scheme for export ciphers: the export bits are
18509 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18510 SSL_EXP_MASK. So, the original variable has to be used instead of the
18511 already masked variable.
18513 *Richard Levitte <levitte@stacken.kth.se>*
18515 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18517 *Richard Levitte <levitte@stacken.kth.se>*
18519 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18520 from `int` to `unsigned int` because it is a length and initialized by
18521 EVP_DigestFinal() which expects an `unsigned int *`.
18523 *Richard Levitte <levitte@stacken.kth.se>*
18525 * Don't hard-code path to Perl interpreter on shebang line of Configure
18526 script. Instead use the usual Shell->Perl transition trick.
18528 *Ralf S. Engelschall*
18530 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18531 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18532 -noout -modulus` as it's already the case for `openssl rsa -noout
18533 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18534 currently the public key is printed (a decision which was already done by
18535 `openssl dsa -modulus` in the past) which serves a similar purpose.
18536 Additionally the NO_RSA no longer completely removes the whole -modulus
18537 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18540 *Ralf S. Engelschall*
18542 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18543 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18545 *Arne Ansper <arne@ats.cyber.ee>*
18547 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18548 to be added. Now both 'req' and 'ca' can use new objects defined in the
18553 * Add cool BIO that does syslog (or event log on NT).
18555 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18557 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18558 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18559 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18560 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18564 * Add preliminary config info for new extension code.
18568 * Make RSA_NO_PADDING really use no padding.
18570 *Ulf Moeller <ulf@fitug.de>*
18572 * Generate errors when private/public key check is done.
18576 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18577 for some CRL extensions and new objects added.
18581 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18582 key usage extension and fuller support for authority key id.
18586 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18587 padding method for RSA, which is recommended for new applications in PKCS
18588 #1 v2.0 (RFC 2437, October 1998).
18589 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18590 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18591 against Bleichbacher's attack on RSA.
18592 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18595 * Updates to the new SSL compression code
18597 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18599 * Fix so that the version number in the master secret, when passed
18600 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18601 (because the server will not accept higher), that the version number
18602 is 0x03,0x01, not 0x03,0x00
18604 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18606 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18607 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18608 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18612 * Support for RAW extensions where an arbitrary extension can be
18613 created by including its DER encoding. See `apps/openssl.cnf` for
18618 * Make sure latest Perl versions don't interpret some generated C array
18619 code as Perl array code in the crypto/err/err_genc.pl script.
18621 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18623 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18624 not many people have the assembler. Various Win32 compilation fixes and
18625 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18626 build instructions.
18630 * Modify configure script 'Configure' to automatically create crypto/date.h
18631 file under Win32 and also build pem.h from pem.org. New script
18632 util/mkfiles.pl to create the MINFO file on environments that can't do a
18633 'make files': perl util/mkfiles.pl >MINFO should work.
18637 * Major rework of DES function declarations, in the pursuit of correctness
18638 and purity. As a result, many evil casts evaporated, and some weirdness,
18639 too. You may find this causes warnings in your code. Zapping your evil
18640 casts will probably fix them. Mostly.
18644 * Fix for a typo in asn1.h. Bug fix to object creation script
18645 obj_dat.pl. It considered a zero in an object definition to mean
18646 "end of object": none of the objects in objects.h have any zeros
18647 so it wasn't spotted.
18649 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18651 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18652 Masking (CBCM). In the absence of test vectors, the best I have been able
18653 to do is check that the decrypt undoes the encrypt, so far. Send me test
18654 vectors if you have them.
18658 * Correct calculation of key length for export ciphers (too much space was
18659 allocated for null ciphers). This has not been tested!
18663 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18664 message is now correct (it understands "crypto" and "ssl" on its
18665 command line). There is also now an "update" option. This will update
18666 the util/ssleay.num and util/libeay.num files with any new functions.
18668 perl util/mkdef.pl crypto ssl update
18669 it will update them.
18673 * Overhauled the Perl interface:
18674 - ported BN stuff to OpenSSL's different BN library
18675 - made the perl/ source tree CVS-aware
18676 - renamed the package from SSLeay to OpenSSL (the files still contain
18677 their history because I've copied them in the repository)
18678 - removed obsolete files (the test scripts will be replaced
18679 by better Test::Harness variants in the future)
18681 *Ralf S. Engelschall*
18683 * First cut for a very conservative source tree cleanup:
18684 1. merge various obsolete readme texts into doc/ssleay.txt
18685 where we collect the old documents and readme texts.
18686 2. remove the first part of files where I'm already sure that we no
18687 longer need them because of three reasons: either they are just temporary
18688 files which were left by Eric or they are preserved original files where
18689 I've verified that the diff is also available in the CVS via "cvs diff
18690 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18691 the crypto/md/ stuff).
18693 *Ralf S. Engelschall*
18695 * More extension code. Incomplete support for subject and issuer alt
18696 name, issuer and authority key id. Change the i2v function parameters
18697 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18698 what that's for :-) Fix to ASN1 macro which messed up
18699 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18703 * Preliminary support for ENUMERATED type. This is largely copied from the
18708 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18710 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18712 * Make sure `make rehash` target really finds the `openssl` program.
18714 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18716 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18717 like to hear about it if this slows down other processors.
18721 * Add CygWin32 platform information to Configure script.
18723 *Alan Batie <batie@aahz.jf.intel.com>*
18725 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18727 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18729 * New program nseq to manipulate netscape certificate sequences
18733 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18738 * Fixes to BN code. Previously the default was to define BN_RECURSION
18739 but the BN code had some problems that would cause failures when
18740 doing certificate verification and some other functions.
18742 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18744 * Add ASN1 and PEM code to support netscape certificate sequences.
18748 * Add ASN1 and PEM code to support netscape certificate sequences.
18752 * Add several PKIX and private extended key usage OIDs.
18756 * Modify the 'ca' program to handle the new extension code. Modify
18757 openssl.cnf for new extension format, add comments.
18761 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18762 and add a sample to openssl.cnf so req -x509 now adds appropriate
18767 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18768 error code, add initial support to X509_print() and x509 application.
18772 * Takes a deep breath and start adding X509 V3 extension support code. Add
18773 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18774 stuff is currently isolated and isn't even compiled yet.
18778 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18779 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18780 Removed the versions check from X509 routines when loading extensions:
18781 this allows certain broken certificates that don't set the version
18782 properly to be processed.
18786 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18787 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18788 can still be regenerated with "make depend".
18792 * Spelling mistake in C version of CAST-128.
18794 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18796 * Changes to the error generation code. The perl script err-code.pl
18797 now reads in the old error codes and retains the old numbers, only
18798 adding new ones if necessary. It also only changes the .err files if new
18799 codes are added. The makefiles have been modified to only insert errors
18800 when needed (to avoid needlessly modifying header files). This is done
18801 by only inserting errors if the .err file is newer than the auto generated
18802 C file. To rebuild all the error codes from scratch (the old behaviour)
18803 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18804 or delete all the .err files.
18808 * CAST-128 was incorrectly implemented for short keys. The C version has
18809 been fixed, but is untested. The assembler versions are also fixed, but
18810 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18811 to regenerate it if needed.
18812 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18813 Hagino <itojun@kame.net>*
18815 * File was opened incorrectly in randfile.c.
18817 *Ulf Möller <ulf@fitug.de>*
18819 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18820 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18821 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18822 al: it's just almost always a UTCTime. Note this patch adds new error
18823 codes so do a "make errors" if there are problems.
18827 * Correct Linux 1 recognition in config.
18829 *Ulf Möller <ulf@fitug.de>*
18831 * Remove pointless MD5 hash when using DSA keys in ca.
18833 *Anonymous <nobody@replay.com>*
18835 * Generate an error if given an empty string as a cert directory. Also
18836 generate an error if handed NULL (previously returned 0 to indicate an
18837 error, but didn't set one).
18839 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18841 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18845 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18846 parameters. This was causing a warning which killed off the Win32 compile.
18850 * Remove C++ style comments from crypto/bn/bn_local.h.
18852 *Neil Costigan <neil.costigan@celocom.com>*
18854 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18855 based on a text string, looking up short and long names and finally
18856 "dot" format. The "dot" format stuff didn't work. Added new function
18857 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18858 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18859 OID is not part of the table.
18863 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18864 X509_LOOKUP_by_alias().
18868 * Sort openssl functions by name.
18872 * Get the `gendsa` command working and add it to the `list` command. Remove
18873 encryption from sample DSA keys (in case anyone is interested the password
18878 * Make *all* `*_free` functions accept a NULL pointer.
18880 *Frans Heymans <fheymans@isaserver.be>*
18882 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18885 *Anonymous <nobody@replay.com>*
18887 * s_server should send the CAfile as acceptable CAs, not its own cert.
18889 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18891 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18893 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18895 * Temp key "for export" tests were wrong in s3_srvr.c.
18897 *Anonymous <nobody@replay.com>*
18899 * Add prototype for temp key callback functions
18900 SSL_CTX_set_tmp_{rsa,dh}_callback().
18904 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18905 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18909 * X509_name_add_entry() freed the wrong thing after an error.
18911 *Arne Ansper <arne@ats.cyber.ee>*
18913 * rsa_eay.c would attempt to free a NULL context.
18915 *Arne Ansper <arne@ats.cyber.ee>*
18917 * BIO_s_socket() had a broken should_retry() on Windoze.
18919 *Arne Ansper <arne@ats.cyber.ee>*
18921 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18923 *Arne Ansper <arne@ats.cyber.ee>*
18925 * Make sure the already existing X509_STORE->depth variable is initialized
18926 in X509_STORE_new(), but document the fact that this variable is still
18927 unused in the certificate verification process.
18929 *Ralf S. Engelschall*
18931 * Fix the various library and `apps/` files to free up pkeys obtained from
18932 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18936 * Fix reference counting in X509_PUBKEY_get(). This makes
18937 demos/maurice/example2.c work, amongst others, probably.
18939 *Steve Henson and Ben Laurie*
18941 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18942 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18943 are no longer created. This way we have a single and consistent command
18944 line interface `openssl <command>`, similar to `cvs <command>`.
18946 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18948 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18949 BIT STRING wrapper always have zero unused bits.
18953 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18957 * Make the top-level INSTALL documentation easier to understand.
18961 * Makefiles updated to exit if an error occurs in a sub-directory
18962 make (including if user presses ^C) [Paul Sutton]
18964 * Make Montgomery context stuff explicit in RSA data structure.
18968 * Fix build order of pem and err to allow for generated pem.h.
18972 * Fix renumbering bug in X509_NAME_delete_entry().
18976 * Enhanced the err-ins.pl script so it makes the error library number
18977 global and can add a library name. This is needed for external ASN1 and
18978 other error libraries.
18982 * Fixed sk_insert which never worked properly.
18986 * Fix ASN1 macros so they can handle indefinite length constructed
18987 EXPLICIT tags. Some non standard certificates use these: they can now
18992 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18993 into a single doc/ssleay.txt bundle. This way the information is still
18994 preserved but no longer messes up this directory. Now it's new room for
18995 the new set of documentation files.
18997 *Ralf S. Engelschall*
18999 * SETs were incorrectly DER encoded. This was a major pain, because they
19000 shared code with SEQUENCEs, which aren't coded the same. This means that
19001 almost everything to do with SETs or SEQUENCEs has either changed name or
19002 number of arguments.
19004 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19006 * Fix test data to work with the above.
19010 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19011 was already fixed by Eric for 0.9.1 it seems.
19013 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19015 * Autodetect FreeBSD3.
19019 * Fix various bugs in Configure. This affects the following platforms:
19023 unixware-2.0-pentium
19028 * Eliminate generated files from CVS. Reorder tests to regenerate files
19029 before they are needed.
19033 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19037 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19039 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19040 changed SSLeay to OpenSSL in version strings.
19042 *Ralf S. Engelschall*
19044 * Some fixups to the top-level documents.
19048 * Fixed the nasty bug where rsaref.h was not found under compile-time
19049 because the symlink to include/ was missing.
19051 *Ralf S. Engelschall*
19053 * Incorporated the popular no-RSA/DSA-only patches
19054 which allow to compile a RSA-free SSLeay.
19056 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19058 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19059 when "ssleay" is still not found.
19061 *Ralf S. Engelschall*
19063 * Added more platforms to Configure: Cray T3E, HPUX 11,
19065 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19067 * Updated the README file.
19069 *Ralf S. Engelschall*
19071 * Added various .cvsignore files in the CVS repository subdirs
19072 to make a "cvs update" really silent.
19074 *Ralf S. Engelschall*
19076 * Recompiled the error-definition header files and added
19077 missing symbols to the Win32 linker tables.
19079 *Ralf S. Engelschall*
19081 * Cleaned up the top-level documents;
19082 o new files: CHANGES and LICENSE
19083 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19084 o merged COPYRIGHT into LICENSE
19085 o removed obsolete TODO file
19086 o renamed MICROSOFT to INSTALL.W32
19088 *Ralf S. Engelschall*
19090 * Removed dummy files from the 0.9.1b source tree:
19092 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19093 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19094 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19095 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19096 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19099 *Ralf S. Engelschall*
19101 * Added various platform portability fixes.
19105 * The Genesis of the OpenSSL rpject:
19106 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19107 Young and Tim J. Hudson created while they were working for C2Net until
19110 *The OpenSSL Project*
19112 ### Changes between 0.9.0b and 0.9.1b [not released]
19114 * Updated a few CA certificates under certs/
19118 * Changed some BIGNUM api stuff.
19122 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19123 DGUX x86, Linux Alpha, etc.
19127 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19128 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19133 * Add -strparse option to asn1pars program which parses nested
19136 *Dr Stephen Henson <shenson@bigfoot.com>*
19138 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19142 * DSA fix for "ca" program.
19146 * Added "-genkey" option to "dsaparam" program.
19150 * Added RIPE MD160 (rmd160) message digest.
19154 * Added -a (all) option to "ssleay version" command.
19158 * Added PLATFORM define which is the id given to Configure.
19162 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19166 * Extended the ASN.1 parser routines.
19170 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19174 * Added a BN_CTX to the BN library.
19178 * Fixed the weak key values in DES library
19182 * Changed API in EVP library for cipher aliases.
19186 * Added support for RC2/64bit cipher.
19190 * Converted the lhash library to the crypto/mem.c functions.
19194 * Added more recognized ASN.1 object ids.
19198 * Added more RSA padding checks for SSL/TLS.
19202 * Added BIO proxy/filter functionality.
19206 * Added extra_certs to SSL_CTX which can be used
19207 send extra CA certificates to the client in the CA cert chain sending
19208 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19212 * Now Fortezza is denied in the authentication phase because
19213 this is key exchange mechanism is not supported by SSLeay at all.
19217 * Additional PKCS1 checks.
19221 * Support the string "TLSv1" for all TLS v1 ciphers.
19225 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19226 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19230 * Fixed a few memory leaks.
19234 * Fixed various code and comment typos.
19238 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19239 bytes sent in the client random.
19241 *Edward Bishop <ebishop@spyglass.com>*
19245 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19246 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19247 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19248 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19249 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19250 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19251 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19252 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19253 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19254 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19255 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19256 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19257 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19258 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19259 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19260 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19261 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19262 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19263 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19264 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19265 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19266 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19267 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19268 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19269 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19270 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19271 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19272 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19273 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19274 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19275 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19276 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19277 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19278 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19279 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19280 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19281 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19282 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19283 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19284 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19285 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19286 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19287 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19288 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19289 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19290 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19291 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19292 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19293 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19294 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19295 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19296 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19297 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19298 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19299 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19300 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19301 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19302 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19303 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19304 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19305 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19306 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19307 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19308 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19309 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19310 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19311 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19312 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19313 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19314 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19315 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19316 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19317 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19318 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19319 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19320 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19321 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19322 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19323 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19324 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19325 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19326 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19327 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19328 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19329 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19330 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19331 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19332 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19333 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19334 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19335 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19336 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19337 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19338 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19339 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19340 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19341 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19342 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19343 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19344 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19345 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19346 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19347 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19348 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19349 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19350 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19351 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19352 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19353 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19354 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19355 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19356 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19357 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19358 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19359 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19360 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19361 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19362 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19363 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19364 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19365 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19366 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19367 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19368 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19369 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19370 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19371 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19372 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19373 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19374 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19375 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19376 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19377 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19378 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19379 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19380 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19381 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19382 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19383 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19384 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19385 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19386 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19387 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19388 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19389 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19390 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19391 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19392 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19393 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19394 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19395 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19396 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19397 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19398 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19399 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19400 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19401 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19402 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19403 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19404 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19405 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19406 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655