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 alpha 16 [6 May 2021]
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 numerous "printing" functions such as X509_signature_print(),
138 X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been
139 amended such that there may be cosmetic differences between the output
140 observed in 1.1.1 and 3.0. This also applies to the "-text" output from the
141 x509 and crl applications.
145 * Windows thread synchronization uses read/write primitives (SRWLock) when
146 supported by the OS, otherwise CriticalSection continues to be used.
150 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
151 work on read only BIO source/sinks that do not support these functions.
152 This allows piping or redirection of a file BIO using stdin to be buffered
153 into memory. This is used internally in OSSL_DECODER_from_bio().
157 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
158 this function would return one of the values OSSL_STORE_INFO_NAME,
159 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
160 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
161 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
162 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
163 using this function should be amended to handle the changed return value.
167 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
168 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
169 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
170 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
171 contains more than one certificate identifier: This means that all
172 certificates referenced there MUST be part of the validation chain.
176 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
177 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
178 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
179 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
180 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
181 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
182 using the EVP APIs to access these ciphers should instead use more modern
183 ciphers. If that is not possible then these applications should ensure that
184 the legacy provider has been loaded. This can be achieved either
185 programmatically or via configuration. See the provider(7) man page for
190 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
191 RIPEMD-160 have been moved to the legacy provider. Applications using the
192 EVP APIs to access these digests should instead use more modern digests. If
193 that is not possible then these applications should ensure that the legacy
194 provider has been loaded. This can be achieved either programmatically or via
195 configuration. See the provider(7) man page for further details.
199 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
204 * The deprecated functions EVP_PKEY_get0_RSA(),
205 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
206 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
207 well as the similarly named "get1" functions behave slightly differently in
208 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
209 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
210 provider. Calling these functions will only return a handle on the internal
211 key where the EVP_PKEY was constructed using this key in the first place, for
212 example using a function or macro such as EVP_PKEY_assign_RSA(),
213 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
214 then these functions now return a cached copy of the key. Changes to
215 the internal provider key that take place after the first time the cached key
216 is accessed will not be reflected back in the cached copy. Similarly any
217 changes made to the cached copy by application code will not be reflected
218 back in the internal provider key.
220 For the above reasons the keys returned from these functions should typically
221 be treated as read-only. To emphasise this the value returned from
222 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
223 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
224 break some existing code. Applications broken by this change should be
225 modified. The preferred solution is to refactor the code to avoid the use of
226 these deprecated functions. Failing this the code should be modified to use a
227 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
228 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
229 non-const pointer to enable them to be "freed". However they should also be
230 treated as read-only.
234 * A number of functions handling low level keys or engines were deprecated
235 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
236 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
237 EVP_PKEY_get0_siphash(). Applications using engines should instead use
238 providers. Applications getting or setting low level keys in an EVP_PKEY
239 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
240 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
244 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
245 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
246 and require returning octet ptr parameters from providers that
247 would like to support them which complicates provider implementations.
251 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
252 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
253 RAND_set_rand_method(). Provider based random number generators should
254 be used instead via EVP_RAND(3).
258 * The SRP APIs have been deprecated. The old APIs do not work via providers,
259 and there is no EVP interface to them. Unfortunately there is no replacement
260 for these APIs at this time.
264 * Add a compile time option to prevent the caching of provider fetched
265 algorithms. This is enabled by including the no-cached-fetch option
266 at configuration time.
270 * The default algorithms for pkcs12 creation with the PKCS12_create() function
271 were changed to more modern PBKDF2 and AES based algorithms. The default
272 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
273 with the password-based encryption iteration count. The default digest
274 algorithm for the MAC computation was changed to SHA-256. The pkcs12
275 application now supports -legacy option that restores the previous
276 default algorithms to support interoperability with legacy systems.
278 *Tomáš Mráz and Sahana Prasad*
280 * The openssl speed command does not use low-level API calls anymore. This
281 implies some of the performance numbers might not be fully comparable
282 with the previous releases due to higher overhead. This applies
283 particularly to measuring performance on smaller data chunks.
287 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
290 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
292 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
293 Typically if OpenSSL has no EC or DH algorithms then it cannot support
294 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
295 through providers. Therefore third party providers may supply group
296 implementations even where there are no built-in ones. Attempting to create
297 TLS connections in such a build without also disabling TLSv1.3 at run time or
298 using third party provider groups may result in handshake failures. TLSv1.3
299 can be disabled at compile time using the "no-tls1_3" Configure option.
303 * The undocumented function X509_certificate_type() has been deprecated;
304 applications can use X509_get0_pubkey() and X509_get0_signature() to
305 get the same information.
309 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
310 functions. They are identical to BN_rand() and BN_rand_range()
315 * Removed RSA padding mode for SSLv23 (which was only used for
316 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
317 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
322 * Deprecated the obsolete X9.31 RSA key generation related functions
323 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
324 BN_X931_generate_prime_ex().
328 * The default key generation method for the regular 2-prime RSA keys was
329 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
330 Conditions Based on Auxiliary Probable Primes). This method is slower
331 than the original method.
335 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
336 They are replaced with the BN_check_prime() function that avoids possible
337 misuse and always uses at least 64 rounds of the Miller-Rabin
338 primality test. At least 64 rounds of the Miller-Rabin test are now also
339 used for all prime generation, including RSA key generation.
340 This increases key generation time, especially for larger keys.
344 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
345 as they are not useful with non-deprecated functions.
349 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
350 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
351 OCSP_REQ_CTX_i2d() and its special form OCSP_REQ_CTX_set1_req(),
352 OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
353 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
354 were used to collect all necessary data to form a HTTP request, and to
355 perform the HTTP transfer with that request. With OpenSSL 3.0, the
356 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
357 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
358 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
359 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
360 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
361 OSSL_HTTP_REQ_CTX_set_max_response_length().
363 *Rich Salz and Richard Levitte*
365 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`,
366 which are superseded by `X509_load_http()` and `X509_CRL_load_http()`.
370 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
374 * Validation of SM2 keys has been separated from the validation of regular EC
375 keys, allowing to improve the SM2 validation process to reject loaded private
376 keys that are not conforming to the SM2 ISO standard.
377 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
382 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
383 switches: a validation failure triggers an early exit, returning a failure
384 exit status to the parent process.
388 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
389 to ignore unknown ciphers.
393 * The `-cipher-commands` and `-digest-commands` options
394 of the command line utility `list` have been deprecated.
395 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
399 * All of the low level EC_KEY functions have been deprecated including:
401 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
402 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
403 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
404 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
405 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
406 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
407 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
408 EC_KEY_METHOD_get_verify,
409 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
410 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
411 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
412 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
413 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
414 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
415 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
416 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
417 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
418 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
419 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
420 Applications that need to implement an EC_KEY_METHOD need to consider
421 implementation of the functionality in a special provider.
422 For replacement of the functions manipulating the EC_KEY objects
423 see the EVP_PKEY-EC(7) manual page.
425 Additionally functions that read and write EC_KEY objects such as
426 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
427 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
428 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
429 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
430 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
431 have also been deprecated. Applications should instead use the
432 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
434 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
435 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
436 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
437 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
438 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
440 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
442 * Deprecated all the libcrypto and libssl error string loading
443 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
444 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
445 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
446 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
447 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
448 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
449 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
450 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
451 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
452 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
454 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
455 now loads error strings automatically.
459 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
460 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
461 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
462 are to be used by servers requiring ephemeral DH keys. Instead applications
463 should consider using the built-in DH parameters that are available by
464 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
465 necessary then applications can use the alternative functions
466 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
467 replacement for the "callback" functions. The callback was originally useful
468 in order to have different parameters for export and non-export ciphersuites.
469 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
470 functions should be replaced by one of the other methods described above.
474 * The `-crypt` option to the `passwd` command line tool has been removed.
478 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
483 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
485 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
486 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
487 The inverse ciphers use AES decryption for wrapping, and
488 AES encryption for unwrapping.
492 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
493 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
494 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
495 OpenSSL 3.0 these are replaced by the more generic functions
496 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
497 The old versions have been converted to deprecated macros that just call the
502 * The security callback, which can be customised by application code, supports
503 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
504 in the "other" parameter. In most places this is what is passed. All these
505 places occur server side. However there was one client side call of this
506 security operation and it passed a DH object instead. This is incorrect
507 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
508 of the other locations. Therefore this client side call has been changed to
509 pass an EVP_PKEY instead.
513 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
514 interface. Their functionality remains unchanged.
518 * Added new option for 'openssl list', '-providers', which will display the
519 list of loaded providers, their names, version and status. It optionally
520 displays their gettable parameters.
524 * Removed EVP_PKEY_set_alias_type(). This function was previously
525 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
526 type is internally recognised so the workaround is no longer needed.
528 This is a breaking change from previous OpenSSL versions.
532 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
533 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
537 * Changed all "STACK" functions to be macros instead of inline functions. Macro
538 parameters are still checked for type safety at compile time via helper
543 * Remove the RAND_DRBG API
545 The RAND_DRBG API did not fit well into the new provider concept as
546 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
547 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
548 and some of its API calls are rather low-level. This holds in particular
549 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
551 Adding a compatibility layer to continue supporting the RAND_DRBG API as
552 a legacy API for a regular deprecation period turned out to come at the
553 price of complicating the new provider API unnecessarily. Since the
554 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
557 *Paul Dale and Matthias St. Pierre*
559 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
560 as well as actual hostnames.
564 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
565 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
566 conversely, silently ignore DTLS protocol version bounds when configuring
567 TLS-based contexts. The commands can be repeated to set bounds of both
568 types. The same applies with the corresponding "min_protocol" and
569 "max_protocol" command-line switches, in case some application uses both TLS
572 SSL_CTX instances that are created for a fixed protocol version (e.g.
573 `TLSv1_server_method()`) also silently ignore version bounds. Previously
574 attempts to apply bounds to these protocol versions would result in an
575 error. Now only the "version-flexible" SSL_CTX instances are subject to
576 limits in configuration files in command-line options.
580 * Deprecated the `ENGINE` API. Engines should be replaced with providers
585 * Reworked the recorded ERR codes to make better space for system errors.
586 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
587 given code is a system error (true) or an OpenSSL error (false).
591 * Reworked the test perl framework to better allow parallel testing.
593 *Nicola Tuveri and David von Oheimb*
595 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
596 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
600 * 'Configure' has been changed to figure out the configuration target if
601 none is given on the command line. Consequently, the 'config' script is
602 now only a mere wrapper. All documentation is changed to only mention
605 *Rich Salz and Richard Levitte*
607 * Added a library context `OSSL_LIB_CTX` that applications as well as
608 other libraries can use to form a separate context within which
609 libcrypto operations are performed.
611 There are two ways this can be used:
613 - Directly, by passing a library context to functions that take
614 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
616 - Indirectly, by creating a new library context and then assigning
617 it as the new default, with `OSSL_LIB_CTX_set0_default`.
619 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
620 apart from the functions directly related to `OSSL_LIB_CTX`, accept
621 NULL to indicate that the default library context should be used.
623 Library code that changes the default library context using
624 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
625 second call before returning to the caller.
627 _(Note: the library context was initially called `OPENSSL_CTX` and
628 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
632 * Handshake now fails if Extended Master Secret extension is dropped
637 * Dropped interactive mode from the `openssl` program. From now on,
638 running it without arguments is equivalent to `openssl help`.
642 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()` since their
643 return values were confusing: Unlike other `_cmp()` functions
644 they do not return 0 when their arguments are equal.
645 The new replacement functions `EVP_PKEY_eq()` and `EVP_PKEY_parameters_eq()`
648 *David von Oheimb and Shane Lontis*
650 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
651 `EC_GROUP_get_field_type()`.
655 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
656 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
657 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
658 Applications should rely on the library automatically assigning a suitable
659 EC_METHOD internally upon EC_GROUP construction.
663 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
664 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
665 assigned internally without application intervention.
666 Users of EC_GROUP_new() should switch to a different suitable constructor.
670 * Add CAdES-BES signature verification support, mostly derived
671 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
673 *Filipe Raimundo da Silva*
675 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
679 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
680 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
681 is to support encryption and decryption of a digital envelope that is both
682 authenticated and encrypted using AES GCM mode.
686 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
687 functions are not widely used and now OpenSSL automatically perform this
688 conversion when needed.
692 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
693 EC_KEY_precompute_mult(). These functions are not widely used and
694 applications should instead switch to named curves which OpenSSL has
695 hardcoded lookup tables for.
699 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
700 should instead use the L<EC_POINT_mul(3)> function.
704 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
705 that are not applicable to the new provider model. Applications should
706 instead use EVP_default_properties_is_fips_enabled() and
707 EVP_default_properties_enable_fips().
711 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
712 is set, an unexpected EOF is ignored, it pretends a close notify was received
713 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
717 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
718 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
719 used and applications should instead use the
720 L<EC_POINT_set_affine_coordinates(3)> and
721 L<EC_POINT_get_affine_coordinates(3)> functions.
725 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
726 arrays to be more easily constructed via a series of utility functions.
727 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
728 the various push functions and finally convert to a passable OSSL_PARAM
729 array using OSSL_PARAM_BLD_to_param().
733 * The security strength of SHA1 and MD5 based signatures in TLS has been
734 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
735 working at the default security level of 1 and instead requires security
736 level 0. The security level can be changed either using the cipher string
737 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
738 that where the signature algorithms extension is missing from a ClientHello
739 then the handshake will fail in TLS 1.2 at security level 1. This is because,
740 although this extension is optional, failing to provide one means that
741 OpenSSL will fallback to a default set of signature algorithms. This default
742 set requires the availability of SHA1.
746 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
747 contain a provider side internal key.
751 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
752 They are old functions that we don't use, and that you could disable with
753 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
757 * Project text documents not yet having a proper file name extension
758 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
759 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
763 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
764 have been converted to Markdown with the goal to produce documents
765 which not only look pretty when viewed online in the browser, but
766 remain well readable inside a plain text editor.
768 To achieve this goal, a 'minimalistic' Markdown style has been applied
769 which avoids formatting elements that interfere too much with the
770 reading flow in the text file. For example, it
772 * avoids [ATX headings][] and uses [setext headings][] instead
773 (which works for `<h1>` and `<h2>` headings only).
774 * avoids [inline links][] and uses [reference links][] instead.
775 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
777 [ATX headings]: https://github.github.com/gfm/#atx-headings
778 [setext headings]: https://github.github.com/gfm/#setext-headings
779 [inline links]: https://github.github.com/gfm/#inline-link
780 [reference links]: https://github.github.com/gfm/#reference-link
781 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
782 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
784 *Matthias St. Pierre*
786 * The test suite is changed to preserve results of each test recipe.
787 A new directory test-runs/ with subdirectories named like the
788 test recipes are created in the build tree for this purpose.
792 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
793 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
794 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
796 *David von Oheimb, Martin Peylo*
798 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
799 The legacy OCSP-focused and only partly documented API is retained for
800 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
804 * Added `util/check-format.pl`, a tool for checking adherence to the
805 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
806 The checks performed are incomplete and yield some false positives.
807 Still the tool should be useful for detecting most typical glitches.
811 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
812 If domain name resolution yields multiple IP addresses all of them are tried
813 after `connect()` failures.
817 * All of the low level RSA functions have been deprecated including:
819 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
820 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
821 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
822 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
823 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
824 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
825 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
826 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
827 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
828 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
829 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
830 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
831 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
832 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
833 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
834 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
835 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
836 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
837 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
838 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
839 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
840 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
841 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
842 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
843 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
844 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
845 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
846 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
848 Use of these low level functions has been informally discouraged for a long
849 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
850 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
851 L<EVP_PKEY_decrypt(3)>.
853 All of these low level RSA functions have been deprecated without
856 RSA_blinding_off, RSA_blinding_on, RSA_clear_flags, RSA_get_version,
857 RSAPrivateKey_dup, RSAPublicKey_dup, RSA_set_flags, RSA_setup_blinding and
860 All of these RSA flags have been deprecated without replacement:
862 RSA_FLAG_BLINDING, RSA_FLAG_CACHE_PRIVATE, RSA_FLAG_CACHE_PUBLIC,
863 RSA_FLAG_EXT_PKEY, RSA_FLAG_NO_BLINDING, RSA_FLAG_THREAD_SAFE and
864 RSA_METHOD_FLAG_NO_CHECK.
868 * X509 certificates signed using SHA1 are no longer allowed at security
870 In TLS/SSL the default security level is 1. It can be set either
871 using the cipher string with `@SECLEVEL`, or calling
872 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
873 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
875 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
876 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
877 options of the commands.
881 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
882 modified to use PKEY APIs. These commands are now in maintenance mode
883 and no new features will be added to them.
887 * The command line utility rsautl has been deprecated.
888 Instead use the pkeyutl program.
892 * The command line utilities genrsa and rsa have been modified to use PKEY
893 APIs. They now write PKCS#8 keys by default. These commands are now in
894 maintenance mode and no new features will be added to them.
898 * All of the low level DH functions have been deprecated including:
900 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
901 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
902 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
903 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
904 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
905 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
906 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
907 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
908 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
909 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
910 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
911 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
912 DH_meth_get_generate_params and DH_meth_set_generate_params.
914 Use of these low level functions has been informally discouraged for a long
915 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
916 and L<EVP_PKEY_derive(3)>.
918 These low level DH functions have been deprecated without replacement:
920 DH_clear_flags, DH_get_1024_160, DH_get_2048_224, DH_get_2048_256,
921 DH_set_flags and DH_test_flags.
923 The DH_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
924 The DH_FLAG_TYPE_DH and DH_FLAG_TYPE_DHX have been deprecated. Use
925 EVP_PKEY_is_a() to determine the type of a key. There is no replacement for
928 Additionally functions that read and write DH objects such as d2i_DHparams,
929 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
930 functions have also been deprecated. Applications should instead use the
931 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
933 Finally functions that assign or obtain DH objects from an EVP_PKEY such as
934 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
935 `EVP_PKEY_set1_DH()` are also deprecated.
936 Applications should instead either read or write an
937 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
938 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
940 *Paul Dale and Matt Caswell*
942 * All of the low level DSA functions have been deprecated including:
944 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
945 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
946 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
947 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
948 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
949 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
950 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
951 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
952 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
953 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
954 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
955 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
956 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
957 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
958 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
959 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
960 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
962 Use of these low level functions has been informally discouraged for a long
963 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
964 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
966 These low level DSA functions have been deprecated without replacement:
968 DSA_clear_flags, DSA_dup_DH, DSAparams_dup, DSA_set_flags and
971 The DSA_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
973 Finally functions that assign or obtain DSA objects from an EVP_PKEY such as
974 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
975 `EVP_PKEY_set1_DSA()` are also deprecated.
976 Applications should instead either read or write an
977 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
978 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
982 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
983 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. This is a breaking
984 change from previous OpenSSL versions.
986 Unlike in previous OpenSSL versions, this means that applications must not
987 call `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
988 The `EVP_PKEY_set_alias_type` function has now been removed.
990 Parameter and key generation is also reworked to make it possible
991 to generate EVP_PKEY_SM2 parameters and keys. Applications must now generate
992 SM2 keys directly and must not create an EVP_PKEY_EC key first.
996 * Deprecated low level ECDH and ECDSA functions. These include:
998 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
999 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
1002 Use of these low level functions has been informally discouraged for a long
1003 time. Instead applications should use the EVP_PKEY_derive(3),
1004 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
1008 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
1009 and EVP_PKEY_decrypt() instead.
1010 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
1011 and EVP_PKEY_encrypt() instead.
1015 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
1016 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
1017 a new formulation to include all the things it can be used for,
1018 as well as words of caution.
1022 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1023 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
1027 * All of the low level HMAC functions have been deprecated including:
1029 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
1030 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
1031 and HMAC_CTX_get_md.
1033 Use of these low level functions has been informally discouraged for a long
1034 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1035 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1036 and L<EVP_MAC_final(3)>.
1040 * Over two thousand fixes were made to the documentation, including:
1041 - Common options (such as -rand/-writerand, TLS version control, etc)
1042 were refactored and point to newly-enhanced descriptions in openssl.pod.
1043 - Added style conformance for all options (with help from Richard Levitte),
1044 documented all reported missing options, added a CI build to check
1045 that all options are documented and that no unimplemented options
1047 - Documented some internals, such as all use of environment variables.
1048 - Addressed all internal broken L<> references.
1052 * All of the low level CMAC functions have been deprecated including:
1054 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
1055 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
1057 Use of these low level functions has been informally discouraged for a long
1058 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1059 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1060 and L<EVP_MAC_final(3)>.
1064 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
1065 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
1068 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
1069 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
1070 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
1071 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
1072 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
1073 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
1074 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
1075 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
1076 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
1077 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
1079 Use of these low level functions has been informally discouraged
1080 for a long time. Applications should use the EVP_DigestInit_ex(3),
1081 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
1085 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1086 set of functions. The documentation mentioned negative values for some
1087 errors, but this was never the case, so the mention of negative values
1090 Code that followed the documentation and thereby check with something
1091 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1095 * All of the low level cipher functions have been deprecated including:
1097 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
1098 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
1099 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
1100 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
1101 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
1102 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
1103 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
1104 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
1105 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
1106 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
1107 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
1108 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
1109 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
1110 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
1111 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
1112 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
1113 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
1114 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
1115 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
1116 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
1117 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
1118 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
1119 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
1120 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
1121 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
1122 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
1123 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
1124 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
1125 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
1127 Use of these low level functions has been informally discouraged for
1128 a long time. Applications should use the high level EVP APIs, e.g.
1129 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
1130 equivalently named decrypt functions instead.
1132 *Matt Caswell and Paul Dale*
1134 * Removed include/openssl/opensslconf.h.in and replaced it with
1135 include/openssl/configuration.h.in, which differs in not including
1136 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1137 was added to include both.
1139 This allows internal hacks where one might need to modify the set
1140 of configured macros, for example this if deprecated symbols are
1141 still supposed to be available internally:
1143 #include <openssl/configuration.h>
1145 #undef OPENSSL_NO_DEPRECATED
1146 #define OPENSSL_SUPPRESS_DEPRECATED
1148 #include <openssl/macros.h>
1150 This should not be used by applications that use the exported
1151 symbols, as that will lead to linking errors.
1155 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1156 used in exponentiation with 512-bit moduli. No EC algorithms are
1157 affected. Analysis suggests that attacks against 2-prime RSA1024,
1158 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1159 difficult to perform and are not believed likely. Attacks against DH512
1160 are considered just feasible. However, for an attack the target would
1161 have to re-use the DH512 private key, which is not recommended anyway.
1162 Also applications directly using the low level API BN_mod_exp may be
1163 affected if they use BN_FLG_CONSTTIME.
1168 * Most memory-debug features have been deprecated, and the functionality
1169 replaced with no-ops.
1173 * Added documentation for the STACK API.
1177 * Introduced a new method type and API, OSSL_ENCODER, to
1178 represent generic encoders. An implementation is expected to
1179 be able to encode an object associated with a given name (such
1180 as an algorithm name for an asymmetric key) into forms given by
1181 implementation properties.
1183 Encoders are primarily used from inside libcrypto, through
1184 calls to functions like EVP_PKEY_print_private(),
1185 PEM_write_bio_PrivateKey() and similar.
1187 Encoders are specified in such a way that they can be made to
1188 directly handle the provider side portion of an object, if this
1189 provider side part comes from the same provider as the encoder
1190 itself, but can also be made to handle objects in parametrized
1191 form (as an OSSL_PARAM array of data). This allows a provider to
1192 offer generic encoders as a service for any other provider.
1196 * Added a .pragma directive to the syntax of configuration files, to
1197 allow varying behavior in a supported and predictable manner.
1198 Currently added pragma:
1202 This allows dollar signs to be a keyword character unless it's
1203 followed by a opening brace or parenthesis. This is useful for
1204 platforms where dollar signs are commonly used in names, such as
1205 volume names and system directory names on VMS.
1209 * Added functionality to create an EVP_PKEY from user data. This
1210 is effectively the same as creating a RSA, DH or DSA object and
1211 then assigning them to an EVP_PKEY, but directly using algorithm
1212 agnostic EVP functions. A benefit is that this should be future
1213 proof for public key algorithms to come.
1217 * Change the interpretation of the '--api' configuration option to
1218 mean that this is a desired API compatibility level with no
1219 further meaning. The previous interpretation, that this would
1220 also mean to remove all deprecated symbols up to and including
1221 the given version, no requires that 'no-deprecated' is also used
1222 in the configuration.
1224 When building applications, the desired API compatibility level
1225 can be set with the OPENSSL_API_COMPAT macro like before. For
1226 API compatibility version below 3.0, the old style numerical
1227 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1228 For version 3.0 and on, the value is expected to be the decimal
1229 value calculated from the major and minor version like this:
1231 MAJOR * 10000 + MINOR * 100
1235 -DOPENSSL_API_COMPAT=30000 For 3.0
1236 -DOPENSSL_API_COMPAT=30200 For 3.2
1238 To hide declarations that are deprecated up to and including the
1239 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1240 given when building the application as well.
1244 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1245 access to certificate and CRL stores via URIs and OSSL_STORE
1248 This adds the following functions:
1250 - X509_LOOKUP_store()
1251 - X509_STORE_load_file()
1252 - X509_STORE_load_path()
1253 - X509_STORE_load_store()
1254 - SSL_add_store_cert_subjects_to_stack()
1255 - SSL_CTX_set_default_verify_store()
1256 - SSL_CTX_load_verify_file()
1257 - SSL_CTX_load_verify_dir()
1258 - SSL_CTX_load_verify_store()
1262 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1263 The presence of this system service is determined at run-time.
1267 * Added functionality to create an EVP_PKEY context based on data
1268 for methods from providers. This takes an algorithm name and a
1269 property query string and simply stores them, with the intent
1270 that any operation that uses this context will use those strings
1271 to fetch the needed methods implicitly, thereby making the port
1272 of application written for pre-3.0 OpenSSL easier.
1276 * The undocumented function NCONF_WIN32() has been deprecated; for
1277 conversion details see the HISTORY section of doc/man5/config.pod
1281 * Introduced the new functions EVP_DigestSignInit_ex() and
1282 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1283 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1284 pages for further details.
1288 * Over two thousand fixes were made to the documentation, including:
1289 adding missing command flags, better style conformance, documentation
1292 *Rich Salz, Richard Levitte*
1294 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1295 X25519, X448, Ed25519 and Ed448.
1299 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1304 * Deprecated the public definition of `ERR_STATE` as well as the function
1305 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1310 * Added ERR functionality to give callers access to the stored function
1311 names that have replaced the older function code based functions.
1313 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1314 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1315 ERR_peek_error_all() and ERR_peek_last_error_all().
1317 These functions have become deprecated: ERR_get_error_line(),
1318 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1319 ERR_peek_last_error_line_data() and ERR_func_error_string().
1321 Users are recommended to use ERR_get_error_all(), or to pick information
1322 with ERR_peek functions and finish off with getting the error code by using
1327 * Extended testing to be verbose for failing tests only. The make variables
1328 VERBOSE_FAILURE or VF can be used to enable this:
1330 $ make VF=1 test # Unix
1331 $ mms /macro=(VF=1) test ! OpenVMS
1332 $ nmake VF=1 test # Windows
1336 * Added the `-copy_extensions` option to the `x509` command for use with
1337 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1338 all extensions in the request are copied to the certificate or vice versa.
1340 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1342 * Added the `-copy_extensions` option to the `req` command for use with
1343 `-x509`. When given with the `copy` or `copyall` argument,
1344 all extensions in the certification request are copied to the certificate.
1348 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1349 they generate are by default RFC 5280 compliant in the following sense:
1350 There is a subjectKeyIdentifier extension with a hash value of the public key
1351 and for not self-signed certs there is an authorityKeyIdentifier extension
1352 with a keyIdentifier field or issuer information identifying the signing key.
1353 This is done unless some configuration overrides the new default behavior,
1354 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1358 * Added several checks to `X509_verify_cert()` according to requirements in
1359 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1360 (which may be done by using the CLI option `-x509_strict`):
1361 * The basicConstraints of CA certificates must be marked critical.
1362 * CA certificates must explicitly include the keyUsage extension.
1363 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1364 * The issuer name of any certificate must not be empty.
1365 * The subject name of CA certs, certs with keyUsage crlSign,
1366 and certs without subjectAlternativeName must not be empty.
1367 * If a subjectAlternativeName extension is given it must not be empty.
1368 * The signatureAlgorithm field and the cert signature must be consistent.
1369 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1370 must not be marked critical.
1371 * The authorityKeyIdentifier must be given for X.509v3 certs
1372 unless they are self-signed.
1373 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1377 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1378 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1382 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1383 used even when parsing explicit parameters, when loading a encoded key
1384 or calling `EC_GROUP_new_from_ecpkparameters()`/
1385 `EC_GROUP_new_from_ecparameters()`.
1386 This prevents bypass of security hardening and performance gains,
1387 especially for curves with specialized EC_METHODs.
1388 By default, if a key encoded with explicit parameters is loaded and later
1389 encoded, the output is still encoded with explicit parameters, even if
1390 internally a "named" EC_GROUP is used for computation.
1394 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1395 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1396 NULL. After this change, only the cofactor parameter can be NULL. It also
1397 does some minimal sanity checks on the passed order.
1402 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1403 An attack is simple, if the first CMS_recipientInfo is valid but the
1404 second CMS_recipientInfo is chosen ciphertext. If the second
1405 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1406 encryption key will be replaced by garbage, and the message cannot be
1407 decoded, but if the RSA decryption fails, the correct encryption key is
1408 used and the recipient will not notice the attack.
1409 As a work around for this potential attack the length of the decrypted
1410 key must be equal to the cipher default key length, in case the
1411 certifiate is not given and all recipientInfo are tried out.
1412 The old behaviour can be re-enabled in the CMS code by setting the
1413 CMS_DEBUG_DECRYPT flag.
1417 * Early start up entropy quality from the DEVRANDOM seed source has been
1418 improved for older Linux systems. The RAND subsystem will wait for
1419 /dev/random to be producing output before seeding from /dev/urandom.
1420 The seeded state is stored for future library initialisations using
1421 a system global shared memory segment. The shared memory identifier
1422 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1423 the desired value. The default identifier is 114.
1427 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1428 when primes for RSA keys are computed.
1429 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1430 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1431 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1432 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1433 This avoids possible fingerprinting of newly generated RSA modules.
1437 * Correct the extended master secret constant on EBCDIC systems. Without this
1438 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1439 negotiate EMS will fail. Unfortunately this also means that TLS connections
1440 between EBCDIC systems with this fix, and EBCDIC systems without this
1441 fix will fail if they negotiate EMS.
1445 * Changed the library initialisation so that the config file is now loaded
1446 by default. This was already the case for libssl. It now occurs for both
1447 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1448 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1452 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1453 where the former acts as a replacement for `ERR_put_error()`, and the
1454 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1455 `ERR_raise_data()` adds more flexibility by taking a format string and
1456 an arbitrary number of arguments following it, to be processed with
1461 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1462 to check if a named provider is loaded and available. When called, it
1463 will also activate all fallback providers if such are still present.
1467 * Enforce a minimum DH modulus size of 512 bits.
1471 * Changed DH parameters to generate the order q subgroup instead of 2q.
1472 Previously generated DH parameters are still accepted by DH_check
1473 but DH_generate_key works around that by clearing bit 0 of the
1474 private key for those. This avoids leaking bit 0 of the private key.
1478 * Significantly reduce secure memory usage by the randomness pools.
1482 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1487 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1488 algorithms. An implementation of a key exchange algorithm can be obtained
1489 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1490 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1491 the older EVP_PKEY_derive_init() function. See the man pages for the new
1492 functions for further details.
1496 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1500 * Removed the function names from error messages and deprecated the
1503 * Removed NextStep support and the macro OPENSSL_UNISTD
1507 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1508 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1509 Also removed "export var as function" capability; we do not export
1510 variables, only functions.
1514 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1515 an error and 1 indicating success. In previous versions of OpenSSL this
1516 was a void type. If a key was set longer than the maximum possible this
1521 * Support SM2 signing and verification schemes with X509 certificate.
1525 * Use SHA256 as the default digest for TS query in the `ts` app.
1529 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1530 This checks that the salt length is at least 128 bits, the derived key
1531 length is at least 112 bits, and that the iteration count is at least 1000.
1532 For backwards compatibility these checks are disabled by default in the
1533 default provider, but are enabled by default in the fips provider.
1534 To enable or disable these checks use the control
1535 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1539 * Default cipher lists/suites are now available via a function, the
1540 #defines are deprecated.
1544 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1545 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1546 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1550 * Join the directories crypto/x509 and crypto/x509v3
1554 * Added command 'openssl kdf' that uses the EVP_KDF API.
1558 * Added command 'openssl mac' that uses the EVP_MAC API.
1562 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1563 as default directories. Also added the command 'openssl info'
1564 for scripting purposes.
1568 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1569 deprecated. These undocumented functions were never integrated into the EVP
1570 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1571 Bi-directional IGE mode. These modes were never formally standardised and
1572 usage of these functions is believed to be very small. In particular
1573 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1574 is ever used. The security implications are believed to be minimal, but
1575 this issue was never fixed for backwards compatibility reasons. New code
1576 should not use these modes.
1580 * Add prediction resistance to the DRBG reseeding process.
1584 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1585 mandated by IEEE Std 1619-2018.
1589 * Added newline escaping functionality to a filename when using openssl dgst.
1590 This output format is to replicate the output format found in the `*sum`
1591 checksum programs. This aims to preserve backward compatibility.
1593 *Matt Eaton, Richard Levitte, and Paul Dale*
1595 * Removed the heartbeat message in DTLS feature, as it has very
1596 little usage and doesn't seem to fulfill a valuable purpose.
1597 The configuration option is now deprecated.
1601 * Changed the output of 'openssl {digestname} < file' to display the
1602 digest name in its output.
1606 * Added a new generic trace API which provides support for enabling
1607 instrumentation through trace output. This feature is mainly intended
1608 as an aid for developers and is disabled by default. To utilize it,
1609 OpenSSL needs to be configured with the `enable-trace` option.
1611 If the tracing API is enabled, the application can activate trace output
1612 by registering BIOs as trace channels for a number of tracing and debugging
1615 The `openssl` program has been expanded to enable any of the types
1616 available via environment variables defined by the user, and serves as
1617 one possible example on how to use this functionality.
1619 *Richard Levitte & Matthias St. Pierre*
1621 * Added build tests for C++. These are generated files that only do one
1622 thing, to include one public OpenSSL head file each. This tests that
1623 the public header files can be usefully included in a C++ application.
1625 This test isn't enabled by default. It can be enabled with the option
1626 'enable-buildtest-c++'.
1630 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1634 * Add KMAC to EVP_MAC.
1638 * Added property based algorithm implementation selection framework to
1643 * Added SCA hardening for modular field inversion in EC_GROUP through
1644 a new dedicated field_inv() pointer in EC_METHOD.
1645 This also addresses a leakage affecting conversions from projective
1646 to affine coordinates.
1648 *Billy Bob Brumley, Nicola Tuveri*
1650 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1651 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1652 those algorithms that were already supported through the EVP_PKEY API
1653 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1654 and scrypt are now wrappers that call EVP_KDF.
1658 * Build devcrypto engine as a dynamic engine.
1660 *Eneas U de Queiroz*
1662 * Add keyed BLAKE2 to EVP_MAC.
1666 * Fix a bug in the computation of the endpoint-pair shared secret used
1667 by DTLS over SCTP. This breaks interoperability with older versions
1668 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1669 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1670 interoperability with such broken implementations. However, enabling
1671 this switch breaks interoperability with correct implementations.
1673 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1674 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1678 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1682 * Changed the license to the Apache License v2.0.
1686 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1688 - Major releases (indicated by incrementing the MAJOR release number)
1689 may introduce incompatible API/ABI changes.
1690 - Minor releases (indicated by incrementing the MINOR release number)
1691 may introduce new features but retain API/ABI compatibility.
1692 - Patch releases (indicated by incrementing the PATCH number)
1693 are intended for bug fixes and other improvements of existing
1694 features only (like improving performance or adding documentation)
1695 and retain API/ABI compatibility.
1699 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1703 * Remove the 'dist' target and add a tarball building script. The
1704 'dist' target has fallen out of use, and it shouldn't be
1705 necessary to configure just to create a source distribution.
1709 * Recreate the OS390-Unix config target. It no longer relies on a
1710 special script like it did for OpenSSL pre-1.1.0.
1714 * Instead of having the source directories listed in Configure, add
1715 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1720 * Add GMAC to EVP_MAC.
1724 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1728 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1729 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1730 to facilitate the continued use of MACs through raw private keys in
1731 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1735 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1736 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1740 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1741 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1742 are retained for backwards compatibility.
1746 * AES-XTS mode now enforces that its two keys are different to mitigate
1747 the attacked described in "Efficient Instantiations of Tweakable
1748 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1749 Details of this attack can be obtained from:
1750 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1754 * Rename the object files, i.e. give them other names than in previous
1755 versions. Their names now include the name of the final product, as
1756 well as its type mnemonic (bin, lib, shlib).
1760 * Added new option for 'openssl list', '-objects', which will display the
1761 list of built in objects, i.e. OIDs with names.
1765 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1766 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1771 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1772 improves application performance by removing data copies and providing
1773 applications with zero-copy system calls such as sendfile and splice.
1777 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1778 option is set, openssl cleanses (zeroize) plaintext bytes from
1779 internal buffers after delivering them to the application. Note,
1780 the application is still responsible for cleansing other copies
1781 (e.g.: data received by SSL_read(3)).
1785 * `PKCS12_parse` now maintains the order of the parsed certificates
1786 when outputting them via `*ca` (rather than reversing it).
1790 * Deprecated pthread fork support methods. These were unused so no
1791 replacement is required.
1793 - OPENSSL_fork_prepare()
1794 - OPENSSL_fork_parent()
1795 - OPENSSL_fork_child()
1799 * Added support for FFDHE key exchange in TLS 1.3.
1806 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1808 * Fixed a problem with verifying a certificate chain when using the
1809 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1810 the certificates present in a certificate chain. It is not set by default.
1812 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1813 the chain that have explicitly encoded elliptic curve parameters was added
1814 as an additional strict check.
1816 An error in the implementation of this check meant that the result of a
1817 previous check to confirm that certificates in the chain are valid CA
1818 certificates was overwritten. This effectively bypasses the check
1819 that non-CA certificates must not be able to issue other certificates.
1821 If a "purpose" has been configured then there is a subsequent opportunity
1822 for checks that the certificate is a valid CA. All of the named "purpose"
1823 values implemented in libcrypto perform this check. Therefore, where
1824 a purpose is set the certificate chain will still be rejected even when the
1825 strict flag has been used. A purpose is set by default in libssl client and
1826 server certificate verification routines, but it can be overridden or
1827 removed by an application.
1829 In order to be affected, an application must explicitly set the
1830 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1831 for the certificate verification or, in the case of TLS client or server
1832 applications, override the default purpose.
1837 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1838 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1839 renegotiation ClientHello omits the signature_algorithms extension (where it
1840 was present in the initial ClientHello), but includes a
1841 signature_algorithms_cert extension then a NULL pointer dereference will
1842 result, leading to a crash and a denial of service attack.
1844 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1845 (which is the default configuration). OpenSSL TLS clients are not impacted by
1849 *Peter Kästle and Samuel Sapalski*
1851 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1853 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1854 create a unique hash value based on the issuer and serial number data
1855 contained within an X509 certificate. However it was failing to correctly
1856 handle any errors that may occur while parsing the issuer field (which might
1857 occur if the issuer field is maliciously constructed). This may subsequently
1858 result in a NULL pointer deref and a crash leading to a potential denial of
1864 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1865 padding mode to correctly check for rollback attacks. This is considered a
1866 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1871 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1872 functions. Previously they could overflow the output length argument in some
1873 cases where the input length is close to the maximum permissable length for
1874 an integer on the platform. In such cases the return value from the function
1875 call would be 1 (indicating success), but the output length value would be
1876 negative. This could cause applications to behave incorrectly or crash.
1881 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1882 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1883 could be exploited in a side channel attack to recover the password. Since
1884 the attack is local host only this is outside of the current OpenSSL
1885 threat model and therefore no CVE is assigned.
1887 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1892 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1894 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1895 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1896 If an attacker can control both items being compared then this could lead
1897 to a possible denial of service attack. OpenSSL itself uses the
1898 GENERAL_NAME_cmp function for two purposes:
1899 1) Comparing CRL distribution point names between an available CRL and a
1900 CRL distribution point embedded in an X509 certificate
1901 2) When verifying that a timestamp response token signer matches the
1902 timestamp authority name (exposed via the API functions
1903 TS_RESP_verify_response and TS_RESP_verify_token)
1908 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1910 * Certificates with explicit curve parameters are now disallowed in
1911 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1915 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1916 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1917 conversely, silently ignore DTLS protocol version bounds when configuring
1918 TLS-based contexts. The commands can be repeated to set bounds of both
1919 types. The same applies with the corresponding "min_protocol" and
1920 "max_protocol" command-line switches, in case some application uses both TLS
1923 SSL_CTX instances that are created for a fixed protocol version (e.g.
1924 TLSv1_server_method()) also silently ignore version bounds. Previously
1925 attempts to apply bounds to these protocol versions would result in an
1926 error. Now only the "version-flexible" SSL_CTX instances are subject to
1927 limits in configuration files in command-line options.
1931 * Handshake now fails if Extended Master Secret extension is dropped
1936 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1938 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1940 * Fixed segmentation fault in SSL_check_chain()
1941 Server or client applications that call the SSL_check_chain() function
1942 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1943 dereference as a result of incorrect handling of the
1944 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1945 or unrecognised signature algorithm is received from the peer. This could
1946 be exploited by a malicious peer in a Denial of Service attack.
1951 * Added AES consttime code for no-asm configurations
1952 an optional constant time support for AES was added
1953 when building openssl for no-asm.
1954 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1955 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1956 At this time this feature is by default disabled.
1957 It will be enabled by default in 3.0.
1961 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1963 * Revert the change of EOF detection while reading in libssl to avoid
1964 regressions in applications depending on the current way of reporting
1965 the EOF. As the existing method is not fully accurate the change to
1966 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1967 branch and will be present in the 3.0 release.
1971 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1972 when primes for RSA keys are computed.
1973 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1974 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1975 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1976 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1977 This avoids possible fingerprinting of newly generated RSA modules.
1981 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1983 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1984 while reading in libssl then we would report an error back to the
1985 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1986 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1987 therefore give a hint as to what went wrong.
1991 * Check that ed25519 and ed448 are allowed by the security level. Previously
1992 signature algorithms not using an MD were not being checked that they were
1993 allowed by the security level.
1997 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1998 was not quite right. The behaviour was not consistent between resumption
1999 and normal handshakes, and also not quite consistent with historical
2000 behaviour. The behaviour in various scenarios has been clarified and
2001 it has been updated to make it match historical behaviour as closely as
2006 * *[VMS only]* The header files that the VMS compilers include automatically,
2007 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2008 that the C++ compiler doesn't understand. This is a shortcoming in the
2009 compiler, but can be worked around with `__cplusplus` guards.
2011 C++ applications that use OpenSSL libraries must be compiled using the
2012 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2013 functions. Otherwise, only functions with symbols of less than 31
2014 characters can be used, as the linker will not be able to successfully
2015 resolve symbols with longer names.
2019 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2020 The presence of this system service is determined at run-time.
2024 * Added newline escaping functionality to a filename when using openssl dgst.
2025 This output format is to replicate the output format found in the `*sum`
2026 checksum programs. This aims to preserve backward compatibility.
2028 *Matt Eaton, Richard Levitte, and Paul Dale*
2030 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2035 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2037 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2038 number generator (RNG). This was intended to include protection in the
2039 event of a fork() system call in order to ensure that the parent and child
2040 processes did not share the same RNG state. However this protection was not
2041 being used in the default case.
2043 A partial mitigation for this issue is that the output from a high
2044 precision timer is mixed into the RNG state so the likelihood of a parent
2045 and child process sharing state is significantly reduced.
2047 If an application already calls OPENSSL_init_crypto() explicitly using
2048 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2051 *Matthias St. Pierre*
2053 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2054 used even when parsing explicit parameters, when loading a encoded key
2055 or calling `EC_GROUP_new_from_ecpkparameters()`/
2056 `EC_GROUP_new_from_ecparameters()`.
2057 This prevents bypass of security hardening and performance gains,
2058 especially for curves with specialized EC_METHODs.
2059 By default, if a key encoded with explicit parameters is loaded and later
2060 encoded, the output is still encoded with explicit parameters, even if
2061 internally a "named" EC_GROUP is used for computation.
2065 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2066 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2067 NULL. After this change, only the cofactor parameter can be NULL. It also
2068 does some minimal sanity checks on the passed order.
2073 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2074 An attack is simple, if the first CMS_recipientInfo is valid but the
2075 second CMS_recipientInfo is chosen ciphertext. If the second
2076 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2077 encryption key will be replaced by garbage, and the message cannot be
2078 decoded, but if the RSA decryption fails, the correct encryption key is
2079 used and the recipient will not notice the attack.
2080 As a work around for this potential attack the length of the decrypted
2081 key must be equal to the cipher default key length, in case the
2082 certifiate is not given and all recipientInfo are tried out.
2083 The old behaviour can be re-enabled in the CMS code by setting the
2084 CMS_DEBUG_DECRYPT flag.
2089 * Early start up entropy quality from the DEVRANDOM seed source has been
2090 improved for older Linux systems. The RAND subsystem will wait for
2091 /dev/random to be producing output before seeding from /dev/urandom.
2092 The seeded state is stored for future library initialisations using
2093 a system global shared memory segment. The shared memory identifier
2094 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2095 the desired value. The default identifier is 114.
2099 * Correct the extended master secret constant on EBCDIC systems. Without this
2100 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2101 negotiate EMS will fail. Unfortunately this also means that TLS connections
2102 between EBCDIC systems with this fix, and EBCDIC systems without this
2103 fix will fail if they negotiate EMS.
2107 * Use Windows installation paths in the mingw builds
2109 Mingw isn't a POSIX environment per se, which means that Windows
2110 paths should be used for installation.
2115 * Changed DH_check to accept parameters with order q and 2q subgroups.
2116 With order 2q subgroups the bit 0 of the private key is not secret
2117 but DH_generate_key works around that by clearing bit 0 of the
2118 private key for those. This avoids leaking bit 0 of the private key.
2122 * Significantly reduce secure memory usage by the randomness pools.
2126 * Revert the DEVRANDOM_WAIT feature for Linux systems
2128 The DEVRANDOM_WAIT feature added a select() call to wait for the
2129 /dev/random device to become readable before reading from the
2130 /dev/urandom device.
2132 It turned out that this change had negative side effects on
2133 performance which were not acceptable. After some discussion it
2134 was decided to revert this feature and leave it up to the OS
2135 resp. the platform maintainer to ensure a proper initialization
2136 during early boot time.
2138 *Matthias St. Pierre*
2140 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2142 * Add build tests for C++. These are generated files that only do one
2143 thing, to include one public OpenSSL head file each. This tests that
2144 the public header files can be usefully included in a C++ application.
2146 This test isn't enabled by default. It can be enabled with the option
2147 'enable-buildtest-c++'.
2151 * Enable SHA3 pre-hashing for ECDSA and DSA.
2155 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2156 This changes the size when using the `genpkey` command when no size is given.
2157 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2158 generation commands to use 2048 bits by default.
2162 * Reorganize the manual pages to consistently have RETURN VALUES,
2163 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2164 util/fix-doc-nits accordingly.
2166 *Paul Yang, Joshua Lock*
2168 * Add the missing accessor EVP_PKEY_get0_engine()
2172 * Have commands like `s_client` and `s_server` output the signature scheme
2173 along with other cipher suite parameters when debugging.
2177 * Make OPENSSL_config() error agnostic again.
2181 * Do the error handling in RSA decryption constant time.
2185 * Prevent over long nonces in ChaCha20-Poly1305.
2187 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2188 for every encryption operation. RFC 7539 specifies that the nonce value
2189 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2190 and front pads the nonce with 0 bytes if it is less than 12
2191 bytes. However it also incorrectly allows a nonce to be set of up to 16
2192 bytes. In this case only the last 12 bytes are significant and any
2193 additional leading bytes are ignored.
2195 It is a requirement of using this cipher that nonce values are
2196 unique. Messages encrypted using a reused nonce value are susceptible to
2197 serious confidentiality and integrity attacks. If an application changes
2198 the default nonce length to be longer than 12 bytes and then makes a
2199 change to the leading bytes of the nonce expecting the new value to be a
2200 new unique nonce then such an application could inadvertently encrypt
2201 messages with a reused nonce.
2203 Additionally the ignored bytes in a long nonce are not covered by the
2204 integrity guarantee of this cipher. Any application that relies on the
2205 integrity of these ignored leading bytes of a long nonce may be further
2206 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2207 is safe because no such use sets such a long nonce value. However user
2208 applications that use this cipher directly and set a non-default nonce
2209 length to be longer than 12 bytes may be vulnerable.
2211 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2217 * Add DEVRANDOM_WAIT feature for Linux systems
2219 On older Linux systems where the getrandom() system call is not available,
2220 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2221 Contrary to getrandom(), the /dev/urandom device will not block during
2222 early boot when the kernel CSPRNG has not been seeded yet.
2224 To mitigate this known weakness, use select() to wait for /dev/random to
2225 become readable before reading from /dev/urandom.
2227 * Ensure that SM2 only uses SM3 as digest algorithm
2231 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2233 * Change the info callback signals for the start and end of a post-handshake
2234 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2235 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2236 confused by this and assume that a TLSv1.2 renegotiation has started. This
2237 can break KeyUpdate handling. Instead we no longer signal the start and end
2238 of a post handshake message exchange (although the messages themselves are
2239 still signalled). This could break some applications that were expecting
2240 the old signals. However without this KeyUpdate is not usable for many
2245 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2247 * Timing vulnerability in DSA signature generation
2249 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2250 timing side channel attack. An attacker could use variations in the signing
2251 algorithm to recover the private key.
2253 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2258 * Timing vulnerability in ECDSA signature generation
2260 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2261 timing side channel attack. An attacker could use variations in the signing
2262 algorithm to recover the private key.
2264 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2269 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2270 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2271 of two gigabytes and the error handling improved.
2273 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2274 categorized as a normal bug, not a security issue, because the DRBG reseeds
2275 automatically and is fully functional even without additional randomness
2276 provided by the application.
2278 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2280 * Add a new ClientHello callback. Provides a callback interface that gives
2281 the application the ability to adjust the nascent SSL object at the
2282 earliest stage of ClientHello processing, immediately after extensions have
2283 been collected but before they have been processed. In particular, this
2284 callback can adjust the supported TLS versions in response to the contents
2289 * Add SM2 base algorithm support.
2293 * s390x assembly pack: add (improved) hardware-support for the following
2294 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2295 aes-cfb/cfb8, aes-ecb.
2299 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2300 parameter is no longer accepted, as it leads to a corrupt table. NULL
2301 pem_str is reserved for alias entries only.
2305 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2306 step for prime curves. The new implementation is based on formulae from
2307 differential addition-and-doubling in homogeneous projective coordinates
2308 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2309 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2310 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2311 to work in projective coordinates.
2313 *Billy Bob Brumley, Nicola Tuveri*
2315 * Change generating and checking of primes so that the error rate of not
2316 being prime depends on the intended use based on the size of the input.
2317 For larger primes this will result in more rounds of Miller-Rabin.
2318 The maximal error rate for primes with more than 1080 bits is lowered
2321 *Kurt Roeckx, Annie Yousar*
2323 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2327 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2328 moving between systems, and to avoid confusion when a Windows build is
2329 done with mingw vs with MSVC. For POSIX installs, there's still a
2330 symlink or copy named 'tsget' to avoid that confusion as well.
2334 * Revert blinding in ECDSA sign and instead make problematic addition
2335 length-invariant. Switch even to fixed-length Montgomery multiplication.
2339 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2340 step for binary curves. The new implementation is based on formulae from
2341 differential addition-and-doubling in mixed Lopez-Dahab projective
2342 coordinates, modified to independently blind the operands.
2344 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2346 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2347 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2348 EC_METHODs to implement their own specialized "ladder step", to take
2349 advantage of more favorable coordinate systems or more efficient
2350 differential addition-and-doubling algorithms.
2352 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2354 * Modified the random device based seed sources to keep the relevant
2355 file descriptors open rather than reopening them on each access.
2356 This allows such sources to operate in a chroot() jail without
2357 the associated device nodes being available. This behaviour can be
2358 controlled using RAND_keep_random_devices_open().
2362 * Numerous side-channel attack mitigations have been applied. This may have
2363 performance impacts for some algorithms for the benefit of improved
2364 security. Specific changes are noted in this change log by their respective
2369 * AIX shared library support overhaul. Switch to AIX "natural" way of
2370 handling shared libraries, which means collecting shared objects of
2371 different versions and bitnesses in one common archive. This allows to
2372 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2373 doesn't affect the way 3rd party applications are linked, only how
2374 multi-version installation is managed.
2378 * Make ec_group_do_inverse_ord() more robust and available to other
2379 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2380 mitigations are applied to the fallback BN_mod_inverse().
2381 When using this function rather than BN_mod_inverse() directly, new
2382 EC cryptosystem implementations are then safer-by-default.
2386 * Add coordinate blinding for EC_POINT and implement projective
2387 coordinate blinding for generic prime curves as a countermeasure to
2388 chosen point SCA attacks.
2390 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2392 * Add blinding to ECDSA and DSA signatures to protect against side channel
2393 attacks discovered by Keegan Ryan (NCC Group).
2397 * Enforce checking in the `pkeyutl` command to ensure that the input
2398 length does not exceed the maximum supported digest length when performing
2399 a sign, verify or verifyrecover operation.
2403 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2404 I/O in combination with something like select() or poll() will hang. This
2405 can be turned off again using SSL_CTX_clear_mode().
2406 Many applications do not properly handle non-application data records, and
2407 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2408 around the problems in those applications, but can also break some.
2409 It's recommended to read the manpages about SSL_read(), SSL_write(),
2410 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2411 SSL_CTX_set_read_ahead() again.
2415 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2416 now allow empty (zero character) pass phrases.
2420 * Apply blinding to binary field modular inversion and remove patent
2421 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2425 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2426 binary and prime elliptic curves.
2430 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2431 constant time fixed point multiplication.
2435 * Revise elliptic curve scalar multiplication with timing attack
2436 defenses: ec_wNAF_mul redirects to a constant time implementation
2437 when computing fixed point and variable point multiplication (which
2438 in OpenSSL are mostly used with secret scalars in keygen, sign,
2439 ECDH derive operations).
2440 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2443 * Updated CONTRIBUTING
2447 * Updated DRBG / RAND to request nonce and additional low entropy
2448 randomness from the system.
2450 *Matthias St. Pierre*
2452 * Updated 'openssl rehash' to use OpenSSL consistent default.
2456 * Moved the load of the ssl_conf module to libcrypto, which helps
2457 loading engines that libssl uses before libssl is initialised.
2461 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2465 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2467 *Ingo Schwarze, Rich Salz*
2469 * Added output of accepting IP address and port for 'openssl s_server'
2473 * Added a new API for TLSv1.3 ciphersuites:
2474 SSL_CTX_set_ciphersuites()
2475 SSL_set_ciphersuites()
2479 * Memory allocation failures consistently add an error to the error
2484 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2485 in libcrypto when run as setuid/setgid.
2489 * Load any config file by default when libssl is used.
2493 * Added new public header file <openssl/rand_drbg.h> and documentation
2494 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2496 *Matthias St. Pierre*
2498 * QNX support removed (cannot find contributors to get their approval
2499 for the license change).
2503 * TLSv1.3 replay protection for early data has been implemented. See the
2504 SSL_read_early_data() man page for further details.
2508 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2509 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2510 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2511 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2512 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2513 configuration has been separated out. See the ciphers man page or the
2514 SSL_CTX_set_ciphersuites() man page for more information.
2518 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2519 in responder mode now supports the new "-multi" option, which
2520 spawns the specified number of child processes to handle OCSP
2521 requests. The "-timeout" option now also limits the OCSP
2522 responder's patience to wait to receive the full client request
2523 on a newly accepted connection. Child processes are respawned
2524 as needed, and the CA index file is automatically reloaded
2525 when changed. This makes it possible to run the "ocsp" responder
2526 as a long-running service, making the OpenSSL CA somewhat more
2527 feature-complete. In this mode, most diagnostic messages logged
2528 after entering the event loop are logged via syslog(3) rather than
2533 * Added support for X448 and Ed448. Heavily based on original work by
2538 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2539 objects loaded. This adds the functions OSSL_STORE_expect() and
2540 OSSL_STORE_find() as well as needed tools to construct searches and
2541 get the search data out of them.
2545 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2546 version of OpenSSL should review their configuration settings to ensure
2547 that they are still appropriate for TLSv1.3. For further information see:
2548 <https://wiki.openssl.org/index.php/TLS1.3>
2552 * Grand redesign of the OpenSSL random generator
2554 The default RAND method now utilizes an AES-CTR DRBG according to
2555 NIST standard SP 800-90Ar1. The new random generator is essentially
2556 a port of the default random generator from the OpenSSL FIPS 2.0
2557 object module. It is a hybrid deterministic random bit generator
2558 using an AES-CTR bit stream and which seeds and reseeds itself
2559 automatically using trusted system entropy sources.
2561 Some of its new features are:
2562 - Support for multiple DRBG instances with seed chaining.
2563 - The default RAND method makes use of a DRBG.
2564 - There is a public and private DRBG instance.
2565 - The DRBG instances are fork-safe.
2566 - Keep all global DRBG instances on the secure heap if it is enabled.
2567 - The public and private DRBG instance are per thread for lock free
2570 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2572 * Changed Configure so it only says what it does and doesn't dump
2573 so much data. Instead, ./configdata.pm should be used as a script
2574 to display all sorts of configuration data.
2578 * Added processing of "make variables" to Configure.
2582 * Added SHA512/224 and SHA512/256 algorithm support.
2586 * The last traces of Netware support, first removed in 1.1.0, have
2591 * Get rid of Makefile.shared, and in the process, make the processing
2592 of certain files (rc.obj, or the .def/.map/.opt files produced from
2593 the ordinal files) more visible and hopefully easier to trace and
2594 debug (or make silent).
2598 * Make it possible to have environment variable assignments as
2599 arguments to config / Configure.
2603 * Add multi-prime RSA (RFC 8017) support.
2607 * Add SM3 implemented according to GB/T 32905-2016
2608 *Jack Lloyd <jack.lloyd@ribose.com>,*
2609 *Ronald Tse <ronald.tse@ribose.com>,*
2610 *Erick Borsboom <erick.borsboom@ribose.com>*
2612 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2613 as documented in RFC6066.
2614 Based on a patch from Tomasz Moń
2616 *Filipe Raimundo da Silva*
2618 * Add SM4 implemented according to GB/T 32907-2016.
2619 *Jack Lloyd <jack.lloyd@ribose.com>,*
2620 *Ronald Tse <ronald.tse@ribose.com>,*
2621 *Erick Borsboom <erick.borsboom@ribose.com>*
2623 * Reimplement -newreq-nodes and ERR_error_string_n; the
2624 original author does not agree with the license change.
2628 * Add ARIA AEAD TLS support.
2632 * Some macro definitions to support VS6 have been removed. Visual
2633 Studio 6 has not worked since 1.1.0
2637 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2638 without clearing the errors.
2642 * Add "atfork" functions. If building on a system that without
2643 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2644 requirements. The RAND facility now uses/requires this.
2652 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2653 not possible to disable entirely. However, it's still possible to
2654 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2657 To disable, configure with 'no-ui-console'. 'no-ui' is still
2658 possible to use as an alias. Check at compile time with the
2659 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2660 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2664 * Add a STORE module, which implements a uniform and URI based reader of
2665 stores that can contain keys, certificates, CRLs and numerous other
2666 objects. The main API is loosely based on a few stdio functions,
2667 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2668 OSSL_STORE_error and OSSL_STORE_close.
2669 The implementation uses backends called "loaders" to implement arbitrary
2670 URI schemes. There is one built in "loader" for the 'file' scheme.
2674 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2675 then adjusted to work on FreeBSD 8.4 as well.
2676 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2677 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2681 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2682 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2683 error code calls like this:
2685 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2687 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2688 that can be encoded in C. For the foreseeable future, this will only
2691 *Richard Levitte and Tim Hudson*
2693 * Removed BSD cryptodev engine.
2697 * Add a build target 'build_all_generated', to build all generated files
2698 and only that. This can be used to prepare everything that requires
2699 things like perl for a system that lacks perl and then move everything
2700 to that system and do the rest of the build there.
2704 * In the UI interface, make it possible to duplicate the user data. This
2705 can be used by engines that need to retain the data for a longer time
2706 than just the call where this user data is passed.
2710 * Ignore the '-named_curve auto' value for compatibility of applications
2713 *Tomáš Mráz <tmraz@fedoraproject.org>*
2715 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2716 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2717 alerts across multiple records (some of which could be empty). In practice
2718 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2719 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2720 support this at all. Supporting it adds significant complexity to the
2721 record layer, and its removal is unlikely to cause interoperability
2726 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2727 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2728 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2733 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2734 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2736 *Richard Levitte, Andy Polyakov*
2738 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2743 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2744 platform rather than 'mingw'.
2748 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2749 success if they are asked to add an object which already exists
2750 in the store. This change cascades to other functions which load
2751 certificates and CRLs.
2755 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2756 facilitate stack unwinding even from assembly subroutines.
2760 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2761 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2765 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2766 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2767 which is the minimum version we support.
2771 * Certificate time validation (X509_cmp_time) enforces stricter
2772 compliance with RFC 5280. Fractional seconds and timezone offsets
2773 are no longer allowed.
2777 * Add support for ARIA
2781 * s_client will now send the Server Name Indication (SNI) extension by
2782 default unless the new "-noservername" option is used. The server name is
2783 based on the host provided to the "-connect" option unless overridden by
2784 using "-servername".
2788 * Add support for SipHash
2792 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2793 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2794 prevent issues where no progress is being made and the peer continually
2795 sends unrecognised record types, using up resources processing them.
2799 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2800 using the algorithm defined in
2801 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2805 * Heartbeat support has been removed; the ABI is changed for now.
2807 *Richard Levitte, Rich Salz*
2809 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2813 * The RSA "null" method, which was partially supported to avoid patent
2814 issues, has been replaced to always returns NULL.
2821 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2823 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2824 used even when parsing explicit parameters, when loading a encoded key
2825 or calling `EC_GROUP_new_from_ecpkparameters()`/
2826 `EC_GROUP_new_from_ecparameters()`.
2827 This prevents bypass of security hardening and performance gains,
2828 especially for curves with specialized EC_METHODs.
2829 By default, if a key encoded with explicit parameters is loaded and later
2830 encoded, the output is still encoded with explicit parameters, even if
2831 internally a "named" EC_GROUP is used for computation.
2835 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2836 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2837 NULL. After this change, only the cofactor parameter can be NULL. It also
2838 does some minimal sanity checks on the passed order.
2843 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2844 An attack is simple, if the first CMS_recipientInfo is valid but the
2845 second CMS_recipientInfo is chosen ciphertext. If the second
2846 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2847 encryption key will be replaced by garbage, and the message cannot be
2848 decoded, but if the RSA decryption fails, the correct encryption key is
2849 used and the recipient will not notice the attack.
2850 As a work around for this potential attack the length of the decrypted
2851 key must be equal to the cipher default key length, in case the
2852 certifiate is not given and all recipientInfo are tried out.
2853 The old behaviour can be re-enabled in the CMS code by setting the
2854 CMS_DEBUG_DECRYPT flag.
2859 * Use Windows installation paths in the mingw builds
2861 Mingw isn't a POSIX environment per se, which means that Windows
2862 paths should be used for installation.
2867 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2869 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2870 This changes the size when using the `genpkey` command when no size is given.
2871 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2872 generation commands to use 2048 bits by default.
2876 * Prevent over long nonces in ChaCha20-Poly1305.
2878 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2879 for every encryption operation. RFC 7539 specifies that the nonce value
2880 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2881 and front pads the nonce with 0 bytes if it is less than 12
2882 bytes. However it also incorrectly allows a nonce to be set of up to 16
2883 bytes. In this case only the last 12 bytes are significant and any
2884 additional leading bytes are ignored.
2886 It is a requirement of using this cipher that nonce values are
2887 unique. Messages encrypted using a reused nonce value are susceptible to
2888 serious confidentiality and integrity attacks. If an application changes
2889 the default nonce length to be longer than 12 bytes and then makes a
2890 change to the leading bytes of the nonce expecting the new value to be a
2891 new unique nonce then such an application could inadvertently encrypt
2892 messages with a reused nonce.
2894 Additionally the ignored bytes in a long nonce are not covered by the
2895 integrity guarantee of this cipher. Any application that relies on the
2896 integrity of these ignored leading bytes of a long nonce may be further
2897 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2898 is safe because no such use sets such a long nonce value. However user
2899 applications that use this cipher directly and set a non-default nonce
2900 length to be longer than 12 bytes may be vulnerable.
2902 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2908 * Added SCA hardening for modular field inversion in EC_GROUP through
2909 a new dedicated field_inv() pointer in EC_METHOD.
2910 This also addresses a leakage affecting conversions from projective
2911 to affine coordinates.
2913 *Billy Bob Brumley, Nicola Tuveri*
2915 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2916 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2920 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2924 * Remove the 'dist' target and add a tarball building script. The
2925 'dist' target has fallen out of use, and it shouldn't be
2926 necessary to configure just to create a source distribution.
2930 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2932 * Timing vulnerability in DSA signature generation
2934 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2935 timing side channel attack. An attacker could use variations in the signing
2936 algorithm to recover the private key.
2938 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2943 * Timing vulnerability in ECDSA signature generation
2945 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2946 timing side channel attack. An attacker could use variations in the signing
2947 algorithm to recover the private key.
2949 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2954 * Add coordinate blinding for EC_POINT and implement projective
2955 coordinate blinding for generic prime curves as a countermeasure to
2956 chosen point SCA attacks.
2958 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2960 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2962 * Client DoS due to large DH parameter
2964 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2965 malicious server can send a very large prime value to the client. This will
2966 cause the client to spend an unreasonably long period of time generating a
2967 key for this prime resulting in a hang until the client has finished. This
2968 could be exploited in a Denial Of Service attack.
2970 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2975 * Cache timing vulnerability in RSA Key Generation
2977 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2978 a cache timing side channel attack. An attacker with sufficient access to
2979 mount cache timing attacks during the RSA key generation process could
2980 recover the private key.
2982 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2983 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2988 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2989 parameter is no longer accepted, as it leads to a corrupt table. NULL
2990 pem_str is reserved for alias entries only.
2994 * Revert blinding in ECDSA sign and instead make problematic addition
2995 length-invariant. Switch even to fixed-length Montgomery multiplication.
2999 * Change generating and checking of primes so that the error rate of not
3000 being prime depends on the intended use based on the size of the input.
3001 For larger primes this will result in more rounds of Miller-Rabin.
3002 The maximal error rate for primes with more than 1080 bits is lowered
3005 *Kurt Roeckx, Annie Yousar*
3007 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3011 * Add blinding to ECDSA and DSA signatures to protect against side channel
3012 attacks discovered by Keegan Ryan (NCC Group).
3016 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3017 now allow empty (zero character) pass phrases.
3021 * Certificate time validation (X509_cmp_time) enforces stricter
3022 compliance with RFC 5280. Fractional seconds and timezone offsets
3023 are no longer allowed.
3027 * Fixed a text canonicalisation bug in CMS
3029 Where a CMS detached signature is used with text content the text goes
3030 through a canonicalisation process first prior to signing or verifying a
3031 signature. This process strips trailing space at the end of lines, converts
3032 line terminators to CRLF and removes additional trailing line terminators
3033 at the end of a file. A bug in the canonicalisation process meant that
3034 some characters, such as form-feed, were incorrectly treated as whitespace
3035 and removed. This is contrary to the specification (RFC5485). This fix
3036 could mean that detached text data signed with an earlier version of
3037 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3038 signed with a fixed OpenSSL may fail to verify with an earlier version of
3039 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3040 and use the "-binary" flag (for the "cms" command line application) or set
3041 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3045 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3047 * Constructed ASN.1 types with a recursive definition could exceed the stack
3049 Constructed ASN.1 types with a recursive definition (such as can be found
3050 in PKCS7) could eventually exceed the stack given malicious input with
3051 excessive recursion. This could result in a Denial Of Service attack. There
3052 are no such structures used within SSL/TLS that come from untrusted sources
3053 so this is considered safe.
3055 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3061 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3063 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3064 effectively reduced to only comparing the least significant bit of each
3065 byte. This allows an attacker to forge messages that would be considered as
3066 authenticated in an amount of tries lower than that guaranteed by the
3067 security claims of the scheme. The module can only be compiled by the
3068 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3070 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3076 * Add a build target 'build_all_generated', to build all generated files
3077 and only that. This can be used to prepare everything that requires
3078 things like perl for a system that lacks perl and then move everything
3079 to that system and do the rest of the build there.
3083 * Backport SSL_OP_NO_RENGOTIATION
3085 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3086 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3087 changes this is no longer possible in 1.1.0. Therefore the new
3088 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3089 1.1.0 to provide equivalent functionality.
3091 Note that if an application built against 1.1.0h headers (or above) is run
3092 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3093 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3097 * Removed the OS390-Unix config target. It relied on a script that doesn't
3102 * rsaz_1024_mul_avx2 overflow bug on x86_64
3104 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3105 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3106 Analysis suggests that attacks against RSA and DSA as a result of this
3107 defect would be very difficult to perform and are not believed likely.
3108 Attacks against DH1024 are considered just feasible, because most of the
3109 work necessary to deduce information about a private key may be performed
3110 offline. The amount of resources required for such an attack would be
3111 significant. However, for an attack on TLS to be meaningful, the server
3112 would have to share the DH1024 private key among multiple clients, which is
3113 no longer an option since CVE-2016-0701.
3115 This only affects processors that support the AVX2 but not ADX extensions
3116 like Intel Haswell (4th generation).
3118 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3119 was originally found via the OSS-Fuzz project.
3124 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3126 * bn_sqrx8x_internal carry bug on x86_64
3128 There is a carry propagating bug in the x86_64 Montgomery squaring
3129 procedure. No EC algorithms are affected. Analysis suggests that attacks
3130 against RSA and DSA as a result of this defect would be very difficult to
3131 perform and are not believed likely. Attacks against DH are considered just
3132 feasible (although very difficult) because most of the work necessary to
3133 deduce information about a private key may be performed offline. The amount
3134 of resources required for such an attack would be very significant and
3135 likely only accessible to a limited number of attackers. An attacker would
3136 additionally need online access to an unpatched system using the target
3137 private key in a scenario with persistent DH parameters and a private
3138 key that is shared between multiple clients.
3140 This only affects processors that support the BMI1, BMI2 and ADX extensions
3141 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3143 This issue was reported to OpenSSL by the OSS-Fuzz project.
3148 * Malformed X.509 IPAddressFamily could cause OOB read
3150 If an X.509 certificate has a malformed IPAddressFamily extension,
3151 OpenSSL could do a one-byte buffer overread. The most likely result
3152 would be an erroneous display of the certificate in text format.
3154 This issue was reported to OpenSSL by the OSS-Fuzz project.
3159 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3161 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3162 platform rather than 'mingw'.
3166 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3167 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3168 which is the minimum version we support.
3172 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3174 * Encrypt-Then-Mac renegotiation crash
3176 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3177 negotiated where it was not in the original handshake (or vice-versa) then
3178 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3179 and servers are affected.
3181 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3186 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3188 * Truncated packet could crash via OOB read
3190 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3191 cipher is being used, then a truncated packet can cause that host to
3192 perform an out-of-bounds read, usually resulting in a crash.
3194 This issue was reported to OpenSSL by Robert Święcki of Google.
3199 * Bad (EC)DHE parameters cause a client crash
3201 If a malicious server supplies bad parameters for a DHE or ECDHE key
3202 exchange then this can result in the client attempting to dereference a
3203 NULL pointer leading to a client crash. This could be exploited in a Denial
3206 This issue was reported to OpenSSL by Guido Vranken.
3211 * BN_mod_exp may produce incorrect results on x86_64
3213 There is a carry propagating bug in the x86_64 Montgomery squaring
3214 procedure. No EC algorithms are affected. Analysis suggests that attacks
3215 against RSA and DSA as a result of this defect would be very difficult to
3216 perform and are not believed likely. Attacks against DH are considered just
3217 feasible (although very difficult) because most of the work necessary to
3218 deduce information about a private key may be performed offline. The amount
3219 of resources required for such an attack would be very significant and
3220 likely only accessible to a limited number of attackers. An attacker would
3221 additionally need online access to an unpatched system using the target
3222 private key in a scenario with persistent DH parameters and a private
3223 key that is shared between multiple clients. For example this can occur by
3224 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3225 similar to CVE-2015-3193 but must be treated as a separate problem.
3227 This issue was reported to OpenSSL by the OSS-Fuzz project.
3232 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3234 * ChaCha20/Poly1305 heap-buffer-overflow
3236 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3237 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3238 crash. This issue is not considered to be exploitable beyond a DoS.
3240 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3245 * CMS Null dereference
3247 Applications parsing invalid CMS structures can crash with a NULL pointer
3248 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3249 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3250 structure callback if an attempt is made to free certain invalid encodings.
3251 Only CHOICE structures using a callback which do not handle NULL value are
3254 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3259 * Montgomery multiplication may produce incorrect results
3261 There is a carry propagating bug in the Broadwell-specific Montgomery
3262 multiplication procedure that handles input lengths divisible by, but
3263 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3264 and DH private keys are impossible. This is because the subroutine in
3265 question is not used in operations with the private key itself and an input
3266 of the attacker's direct choice. Otherwise the bug can manifest itself as
3267 transient authentication and key negotiation failures or reproducible
3268 erroneous outcome of public-key operations with specially crafted input.
3269 Among EC algorithms only Brainpool P-512 curves are affected and one
3270 presumably can attack ECDH key negotiation. Impact was not analyzed in
3271 detail, because pre-requisites for attack are considered unlikely. Namely
3272 multiple clients have to choose the curve in question and the server has to
3273 share the private key among them, neither of which is default behaviour.
3274 Even then only clients that chose the curve will be affected.
3276 This issue was publicly reported as transient failures and was not
3277 initially recognized as a security issue. Thanks to Richard Morgan for
3278 providing reproducible case.
3283 * Removed automatic addition of RPATH in shared libraries and executables,
3284 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3288 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3290 * Fix Use After Free for large message sizes
3292 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3293 message larger than approx 16k is received then the underlying buffer to
3294 store the incoming message is reallocated and moved. Unfortunately a
3295 dangling pointer to the old location is left which results in an attempt to
3296 write to the previously freed location. This is likely to result in a
3297 crash, however it could potentially lead to execution of arbitrary code.
3299 This issue only affects OpenSSL 1.1.0a.
3301 This issue was reported to OpenSSL by Robert Święcki.
3306 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3308 * OCSP Status Request extension unbounded memory growth
3310 A malicious client can send an excessively large OCSP Status Request
3311 extension. If that client continually requests renegotiation, sending a
3312 large OCSP Status Request extension each time, then there will be unbounded
3313 memory growth on the server. This will eventually lead to a Denial Of
3314 Service attack through memory exhaustion. Servers with a default
3315 configuration are vulnerable even if they do not support OCSP. Builds using
3316 the "no-ocsp" build time option are not affected.
3318 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3323 * SSL_peek() hang on empty record
3325 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3326 sends an empty record. This could be exploited by a malicious peer in a
3327 Denial Of Service attack.
3329 This issue was reported to OpenSSL by Alex Gaynor.
3334 * Excessive allocation of memory in tls_get_message_header() and
3335 dtls1_preprocess_fragment()
3337 A (D)TLS message includes 3 bytes for its length in the header for the
3338 message. This would allow for messages up to 16Mb in length. Messages of
3339 this length are excessive and OpenSSL includes a check to ensure that a
3340 peer is sending reasonably sized messages in order to avoid too much memory
3341 being consumed to service a connection. A flaw in the logic of version
3342 1.1.0 means that memory for the message is allocated too early, prior to
3343 the excessive message length check. Due to way memory is allocated in
3344 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3345 to service a connection. This could lead to a Denial of Service through
3346 memory exhaustion. However, the excessive message length check still takes
3347 place, and this would cause the connection to immediately fail. Assuming
3348 that the application calls SSL_free() on the failed connection in a timely
3349 manner then the 21Mb of allocated memory will then be immediately freed
3350 again. Therefore the excessive memory allocation will be transitory in
3351 nature. This then means that there is only a security impact if:
3353 1) The application does not call SSL_free() in a timely manner in the event
3354 that the connection fails
3356 2) The application is working in a constrained environment where there is
3357 very little free memory
3359 3) The attacker initiates multiple connection attempts such that there are
3360 multiple connections in a state where memory has been allocated for the
3361 connection; SSL_free() has not yet been called; and there is insufficient
3362 memory to service the multiple requests.
3364 Except in the instance of (1) above any Denial Of Service is likely to be
3365 transitory because as soon as the connection fails the memory is
3366 subsequently freed again in the SSL_free() call. However there is an
3367 increased risk during this period of application crashes due to the lack of
3368 memory - which would then mean a more serious Denial of Service.
3370 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3371 (CVE-2016-6307 and CVE-2016-6308)
3375 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3376 had to be removed. Primary reason is that vendor assembler can't
3377 assemble our modules with -KPIC flag. As result it, assembly
3378 support, was not even available as option. But its lack means
3379 lack of side-channel resistant code, which is incompatible with
3380 security by todays standards. Fortunately gcc is readily available
3381 prepackaged option, which we firmly point at...
3385 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3387 * Windows command-line tool supports UTF-8 opt-in option for arguments
3388 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3389 (to any value) allows Windows user to access PKCS#12 file generated
3390 with Windows CryptoAPI and protected with non-ASCII password, as well
3391 as files generated under UTF-8 locale on Linux also protected with
3396 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3397 have been disabled by default and removed from DEFAULT, just like RC4.
3398 See the RC4 item below to re-enable both.
3402 * The method for finding the storage location for the Windows RAND seed file
3403 has changed. First we check %RANDFILE%. If that is not set then we check
3404 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3405 all else fails we fall back to C:\.
3409 * The EVP_EncryptUpdate() function has had its return type changed from void
3410 to int. A return of 0 indicates and error while a return of 1 indicates
3415 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3416 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3417 off the constant time implementation for RSA, DSA and DH have been made
3418 no-ops and deprecated.
3422 * Windows RAND implementation was simplified to only get entropy by
3423 calling CryptGenRandom(). Various other RAND-related tickets
3426 *Joseph Wylie Yandle, Rich Salz*
3428 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3429 and `OPENSSL_LH_`, respectively. The old names are available
3430 with API compatibility. They new names are now completely documented.
3434 * Unify TYPE_up_ref(obj) methods signature.
3435 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3436 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3437 int (instead of void) like all others TYPE_up_ref() methods.
3438 So now these methods also check the return value of CRYPTO_atomic_add(),
3439 and the validity of object reference counter.
3441 *fdasilvayy@gmail.com*
3443 * With Windows Visual Studio builds, the .pdb files are installed
3444 alongside the installed libraries and executables. For a static
3445 library installation, ossl_static.pdb is the associate compiler
3446 generated .pdb file to be used when linking programs.
3450 * Remove openssl.spec. Packaging files belong with the packagers.
3454 * Automatic Darwin/OSX configuration has had a refresh, it will now
3455 recognise x86_64 architectures automatically. You can still decide
3456 to build for a different bitness with the environment variable
3457 KERNEL_BITS (can be 32 or 64), for example:
3459 KERNEL_BITS=32 ./config
3463 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3464 256 bit AES and HMAC with SHA256.
3468 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3472 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3476 * To enable users to have their own config files and build file templates,
3477 Configure looks in the directory indicated by the environment variable
3478 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3479 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3480 name and is used as is.
3484 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3485 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3486 X509_CERT_FILE_CTX was removed.
3490 * "shared" builds are now the default. To create only static libraries use
3491 the "no-shared" Configure option.
3495 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3496 All of these option have not worked for some while and are fundamental
3501 * Make various cleanup routines no-ops and mark them as deprecated. Most
3502 global cleanup functions are no longer required because they are handled
3503 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3504 Explicitly de-initing can cause problems (e.g. where a library that uses
3505 OpenSSL de-inits, but an application is still using it). The affected
3506 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3507 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3508 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3509 COMP_zlib_cleanup().
3513 * --strict-warnings no longer enables runtime debugging options
3514 such as REF_DEBUG. Instead, debug options are automatically
3515 enabled with '--debug' builds.
3517 *Andy Polyakov, Emilia Käsper*
3519 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3520 have been moved out of the public header files. New functions for managing
3521 these have been added.
3525 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3526 objects have been moved out of the public header files. New
3527 functions for managing these have been added.
3531 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3532 have been moved out of the public header files. New functions for managing
3533 these have been added.
3537 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3538 moved out of the public header files. New functions for managing these
3543 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3547 * Removed the mk1mf build scripts.
3551 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3552 it is always safe to #include a header now.
3556 * Removed the aged BC-32 config and all its supporting scripts
3560 * Removed support for Ultrix, Netware, and OS/2.
3564 * Add support for HKDF.
3566 *Alessandro Ghedini*
3568 * Add support for blake2b and blake2s
3572 * Added support for "pipelining". Ciphers that have the
3573 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3574 encryptions/decryptions simultaneously. There are currently no built-in
3575 ciphers with this property but the expectation is that engines will be able
3576 to offer it to significantly improve throughput. Support has been extended
3577 into libssl so that multiple records for a single connection can be
3578 processed in one go (for >=TLS 1.1).
3582 * Added the AFALG engine. This is an async capable engine which is able to
3583 offload work to the Linux kernel. In this initial version it only supports
3584 AES128-CBC. The kernel must be version 4.1.0 or greater.
3588 * OpenSSL now uses a new threading API. It is no longer necessary to
3589 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3590 are two supported threading models: pthreads and windows threads. It is
3591 also possible to configure OpenSSL at compile time for "no-threads". The
3592 old threading API should no longer be used. The functions have been
3593 replaced with "no-op" compatibility macros.
3595 *Alessandro Ghedini, Matt Caswell*
3597 * Modify behavior of ALPN to invoke callback after SNI/servername
3598 callback, such that updates to the SSL_CTX affect ALPN.
3602 * Add SSL_CIPHER queries for authentication and key-exchange.
3606 * Changes to the DEFAULT cipherlist:
3607 - Prefer (EC)DHE handshakes over plain RSA.
3608 - Prefer AEAD ciphers over legacy ciphers.
3609 - Prefer ECDSA over RSA when both certificates are available.
3610 - Prefer TLSv1.2 ciphers/PRF.
3611 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3616 * Change the ECC default curve list to be this, in order: x25519,
3617 secp256r1, secp521r1, secp384r1.
3621 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3622 disabled by default. They can be re-enabled using the
3623 enable-weak-ssl-ciphers option to Configure.
3627 * If the server has ALPN configured, but supports no protocols that the
3628 client advertises, send a fatal "no_application_protocol" alert.
3629 This behaviour is SHALL in RFC 7301, though it isn't universally
3630 implemented by other servers.
3634 * Add X25519 support.
3635 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3636 for public and private key encoding using the format documented in
3637 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3638 key generation and key derivation.
3640 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3645 * Deprecate SRP_VBASE_get_by_user.
3646 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3647 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3648 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3649 seed, even if the seed is configured.
3651 Users should use SRP_VBASE_get1_by_user instead. Note that in
3652 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3653 also that even though configuring the SRP seed attempts to hide
3654 invalid usernames by continuing the handshake with fake
3655 credentials, this behaviour is not constant time and no strong
3656 guarantees are made that the handshake is indistinguishable from
3657 that of a valid user.
3661 * Configuration change; it's now possible to build dynamic engines
3662 without having to build shared libraries and vice versa. This
3663 only applies to the engines in `engines/`, those in `crypto/engine/`
3664 will always be built into libcrypto (i.e. "static").
3666 Building dynamic engines is enabled by default; to disable, use
3667 the configuration option "disable-dynamic-engine".
3669 The only requirements for building dynamic engines are the
3670 presence of the DSO module and building with position independent
3671 code, so they will also automatically be disabled if configuring
3672 with "disable-dso" or "disable-pic".
3674 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3675 are also taken away from openssl/opensslconf.h, as they are
3680 * Configuration change; if there is a known flag to compile
3681 position independent code, it will always be applied on the
3682 libcrypto and libssl object files, and never on the application
3683 object files. This means other libraries that use routines from
3684 libcrypto / libssl can be made into shared libraries regardless
3685 of how OpenSSL was configured.
3687 If this isn't desirable, the configuration options "disable-pic"
3688 or "no-pic" can be used to disable the use of PIC. This will
3689 also disable building shared libraries and dynamic engines.
3693 * Removed JPAKE code. It was experimental and has no wide use.
3697 * The INSTALL_PREFIX Makefile variable has been renamed to
3698 DESTDIR. That makes for less confusion on what this variable
3699 is for. Also, the configuration option --install_prefix is
3704 * Heartbeat for TLS has been removed and is disabled by default
3705 for DTLS; configure with enable-heartbeats. Code that uses the
3706 old #define's might need to be updated.
3708 *Emilia Käsper, Rich Salz*
3710 * Rename REF_CHECK to REF_DEBUG.
3714 * New "unified" build system
3716 The "unified" build system is aimed to be a common system for all
3717 platforms we support. With it comes new support for VMS.
3719 This system builds supports building in a different directory tree
3720 than the source tree. It produces one Makefile (for unix family
3721 or lookalikes), or one descrip.mms (for VMS).
3723 The source of information to make the Makefile / descrip.mms is
3724 small files called 'build.info', holding the necessary
3725 information for each directory with source to compile, and a
3726 template in Configurations, like unix-Makefile.tmpl or
3729 With this change, the library names were also renamed on Windows
3730 and on VMS. They now have names that are closer to the standard
3731 on Unix, and include the major version number, and in certain
3732 cases, the architecture they are built for. See "Notes on shared
3733 libraries" in INSTALL.
3735 We rely heavily on the perl module Text::Template.
3739 * Added support for auto-initialisation and de-initialisation of the library.
3740 OpenSSL no longer requires explicit init or deinit routines to be called,
3741 except in certain circumstances. See the OPENSSL_init_crypto() and
3742 OPENSSL_init_ssl() man pages for further information.
3746 * The arguments to the DTLSv1_listen function have changed. Specifically the
3747 "peer" argument is now expected to be a BIO_ADDR object.
3749 * Rewrite of BIO networking library. The BIO library lacked consistent
3750 support of IPv6, and adding it required some more extensive
3751 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3752 which hold all types of addresses and chains of address information.
3753 It also introduces a new API, with functions like BIO_socket,
3754 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3755 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3756 have been adapted accordingly.
3760 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3765 * CRIME protection: disable compression by default, even if OpenSSL is
3766 compiled with zlib enabled. Applications can still enable compression
3767 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3768 using the SSL_CONF library to configure compression.
3772 * The signature of the session callback configured with
3773 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3774 was explicitly marked as `const unsigned char*` instead of
3779 * Always DPURIFY. Remove the use of uninitialized memory in the
3780 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3784 * Removed many obsolete configuration items, including
3785 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3786 MD2_CHAR, MD2_INT, MD2_LONG
3788 IDEA_SHORT, IDEA_LONG
3789 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3791 *Rich Salz, with advice from Andy Polyakov*
3793 * Many BN internals have been moved to an internal header file.
3795 *Rich Salz with help from Andy Polyakov*
3797 * Configuration and writing out the results from it has changed.
3798 Files such as Makefile include/openssl/opensslconf.h and are now
3799 produced through general templates, such as Makefile.in and
3800 crypto/opensslconf.h.in and some help from the perl module
3803 Also, the center of configuration information is no longer
3804 Makefile. Instead, Configure produces a perl module in
3805 configdata.pm which holds most of the config data (in the hash
3806 table %config), the target data that comes from the target
3807 configuration in one of the `Configurations/*.conf` files (in
3812 * To clarify their intended purposes, the Configure options
3813 --prefix and --openssldir change their semantics, and become more
3814 straightforward and less interdependent.
3816 --prefix shall be used exclusively to give the location INSTALLTOP
3817 where programs, scripts, libraries, include files and manuals are
3818 going to be installed. The default is now /usr/local.
3820 --openssldir shall be used exclusively to give the default
3821 location OPENSSLDIR where certificates, private keys, CRLs are
3822 managed. This is also where the default openssl.cnf gets
3824 If the directory given with this option is a relative path, the
3825 values of both the --prefix value and the --openssldir value will
3826 be combined to become OPENSSLDIR.
3827 The default for --openssldir is INSTALLTOP/ssl.
3829 Anyone who uses --openssldir to specify where OpenSSL is to be
3830 installed MUST change to use --prefix instead.
3834 * The GOST engine was out of date and therefore it has been removed. An up
3835 to date GOST engine is now being maintained in an external repository.
3836 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3837 support for GOST ciphersuites (these are only activated if a GOST engine
3842 * EGD is no longer supported by default; use enable-egd when
3845 *Ben Kaduk and Rich Salz*
3847 * The distribution now has Makefile.in files, which are used to
3848 create Makefile's when Configure is run. *Configure must be run
3849 before trying to build now.*
3853 * The return value for SSL_CIPHER_description() for error conditions
3858 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3860 Obtaining and performing DNSSEC validation of TLSA records is
3861 the application's responsibility. The application provides
3862 the TLSA records of its choice to OpenSSL, and these are then
3863 used to authenticate the peer.
3865 The TLSA records need not even come from DNS. They can, for
3866 example, be used to implement local end-entity certificate or
3867 trust-anchor "pinning", where the "pin" data takes the form
3868 of TLSA records, which can augment or replace verification
3869 based on the usual WebPKI public certification authorities.
3873 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3874 continues to support deprecated interfaces in default builds.
3875 However, applications are strongly advised to compile their
3876 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3877 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3878 or the 1.1.0 releases.
3880 In environments in which all applications have been ported to
3881 not use any deprecated interfaces OpenSSL's Configure script
3882 should be used with the --api=1.1.0 option to entirely remove
3883 support for the deprecated features from the library and
3884 unconditionally disable them in the installed headers.
3885 Essentially the same effect can be achieved with the "no-deprecated"
3886 argument to Configure, except that this will always restrict
3887 the build to just the latest API, rather than a fixed API
3890 As applications are ported to future revisions of the API,
3891 they should update their compile-time OPENSSL_API_COMPAT define
3892 accordingly, but in most cases should be able to continue to
3893 compile with later releases.
3895 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3896 0x10000000L and 0x00908000L, respectively. However those
3897 versions did not support the OPENSSL_API_COMPAT feature, and
3898 so applications are not typically tested for explicit support
3899 of just the undeprecated features of either release.
3903 * Add support for setting the minimum and maximum supported protocol.
3904 It can bet set via the SSL_set_min_proto_version() and
3905 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3906 MaxProtocol. It's recommended to use the new APIs to disable
3907 protocols instead of disabling individual protocols using
3908 SSL_set_options() or SSL_CONF's Protocol. This change also
3909 removes support for disabling TLS 1.2 in the OpenSSL TLS
3910 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3914 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3918 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3919 and integrates ECDSA and ECDH functionality into EC. Implementations can
3920 now redirect key generation and no longer need to convert to or from
3923 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3924 include the ec.h header file instead.
3928 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3929 ciphers who are no longer supported and drops support the ephemeral RSA key
3930 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3934 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3935 opaque. For HMAC_CTX, the following constructors and destructors
3938 HMAC_CTX *HMAC_CTX_new(void);
3939 void HMAC_CTX_free(HMAC_CTX *ctx);
3941 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3942 destroy such methods has been added. See EVP_MD_meth_new(3) and
3943 EVP_CIPHER_meth_new(3) for documentation.
3946 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3947 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3948 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3949 an already created structure.
3950 2) For consistency with the majority of our object creators and
3951 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3952 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3953 for deprecated builds.
3957 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3958 cryptographic operations to be performed asynchronously as long as an
3959 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3960 further details. Libssl has also had this capability integrated with the
3961 introduction of the new mode SSL_MODE_ASYNC and associated error
3962 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3963 pages. This work was developed in partnership with Intel Corp.
3967 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3968 always enabled now. If you want to disable the support you should
3969 exclude it using the list of supported ciphers. This also means that the
3970 "-no_ecdhe" option has been removed from s_server.
3974 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3975 SSL_{CTX_}set1_curves() which can set a list.
3979 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3980 curve you want to support using SSL_{CTX_}set1_curves().
3984 * State machine rewrite. The state machine code has been significantly
3985 refactored in order to remove much duplication of code and solve issues
3986 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3987 further details). This change does have some associated API changes.
3988 Notably the SSL_state() function has been removed and replaced by
3989 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3990 SSL_set_state() has been removed altogether. The previous handshake states
3991 defined in ssl.h and ssl3.h have also been removed.
3995 * All instances of the string "ssleay" in the public API were replaced
3996 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3997 Some error codes related to internal RSA_eay API's were renamed.
4001 * The demo files in crypto/threads were moved to demo/threads.
4005 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4008 *Matt Caswell, Rich Salz*
4010 * New ASN.1 embed macro.
4012 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4013 structure is not allocated: it is part of the parent. That is instead of
4021 This reduces memory fragmentation and make it impossible to accidentally
4022 set a mandatory field to NULL.
4024 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4025 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4026 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4031 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4035 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4036 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4037 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4038 DES and RC4 ciphersuites.
4042 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4043 This changes the decoding behaviour for some invalid messages,
4044 though the change is mostly in the more lenient direction, and
4045 legacy behaviour is preserved as much as possible.
4049 * Fix no-stdio build.
4050 *David Woodhouse <David.Woodhouse@intel.com> and also*
4051 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4053 * New testing framework
4054 The testing framework has been largely rewritten and is now using
4055 perl and the perl modules Test::Harness and an extended variant of
4056 Test::More called OpenSSL::Test to do its work. All test scripts in
4057 test/ have been rewritten into test recipes, and all direct calls to
4058 executables in test/Makefile have become individual recipes using the
4059 simplified testing OpenSSL::Test::Simple.
4061 For documentation on our testing modules, do:
4063 perldoc test/testlib/OpenSSL/Test/Simple.pm
4064 perldoc test/testlib/OpenSSL/Test.pm
4068 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4069 are used; the latter aborts on memory leaks (usually checked on exit).
4070 Some undocumented "set malloc, etc., hooks" functions were removed
4071 and others were changed. All are now documented.
4075 * In DSA_generate_parameters_ex, if the provided seed is too short,
4078 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4080 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4081 from RFC4279, RFC4785, RFC5487, RFC5489.
4083 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4084 original RSA_PSK patch.
4088 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4089 era flag was never set throughout the codebase (only read). Also removed
4090 SSL3_FLAGS_POP_BUFFER which was only used if
4091 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4095 * Changed the default name options in the "ca", "crl", "req" and "x509"
4096 to be "oneline" instead of "compat".
4100 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4101 not aware of clients that still exhibit this bug, and the workaround
4102 hasn't been working properly for a while.
4106 * The return type of BIO_number_read() and BIO_number_written() as well as
4107 the corresponding num_read and num_write members in the BIO structure has
4108 changed from unsigned long to uint64_t. On platforms where an unsigned
4109 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4114 * Given the pervasive nature of TLS extensions it is inadvisable to run
4115 OpenSSL without support for them. It also means that maintaining
4116 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4117 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4121 * Removed support for the two export grade static DH ciphersuites
4122 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4123 were newly added (along with a number of other static DH ciphersuites) to
4124 1.0.2. However the two export ones have *never* worked since they were
4125 introduced. It seems strange in any case to be adding new export
4126 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4130 * Version negotiation has been rewritten. In particular SSLv23_method(),
4131 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4132 and turned into macros which simply call the new preferred function names
4133 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4134 should use the new names instead. Also as part of this change the ssl23.h
4135 header file has been removed.
4139 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4140 code and the associated standard is no longer considered fit-for-purpose.
4144 * RT2547 was closed. When generating a private key, try to make the
4145 output file readable only by the owner. This behavior change might
4146 be noticeable when interacting with other software.
4148 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4153 * Added HTTP GET support to the ocsp command.
4157 * Changed default digest for the dgst and enc commands from MD5 to
4162 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4166 * Added support for TLS extended master secret from
4167 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4168 initial patch which was a great help during development.
4172 * All libssl internal structures have been removed from the public header
4173 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4174 now redundant). Users should not attempt to access internal structures
4175 directly. Instead they should use the provided API functions.
4179 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4180 Access to deprecated functions can be re-enabled by running config with
4181 "enable-deprecated". In addition applications wishing to use deprecated
4182 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4183 will, by default, disable some transitive includes that previously existed
4184 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4188 * Added support for OCB mode. OpenSSL has been granted a patent license
4189 compatible with the OpenSSL license for use of OCB. Details are available
4190 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4191 for OCB can be removed by calling config with no-ocb.
4195 * SSLv2 support has been removed. It still supports receiving a SSLv2
4196 compatible client hello.
4200 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4201 done while fixing the error code for the key-too-small case.
4203 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4205 * CA.sh has been removed; use CA.pl instead.
4209 * Removed old DES API.
4213 * Remove various unsupported platforms:
4219 Sinix/ReliantUNIX RM400
4224 16-bit platforms such as WIN16
4228 * Clean up OPENSSL_NO_xxx #define's
4229 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4230 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4231 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4232 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4233 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4234 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4235 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4236 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4237 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4238 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4242 * Cleaned up dead code
4243 Remove all but one '#ifdef undef' which is to be looked at.
4247 * Clean up calling of xxx_free routines.
4248 Just like free(), fix most of the xxx_free routines to accept
4249 NULL. Remove the non-null checks from callers. Save much code.
4253 * Add secure heap for storage of private keys (when possible).
4254 Add BIO_s_secmem(), CBIGNUM, etc.
4255 Contributed by Akamai Technologies under our Corporate CLA.
4259 * Experimental support for a new, fast, unbiased prime candidate generator,
4260 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4262 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4264 * New output format NSS in the sess_id command line tool. This allows
4265 exporting the session id and the master key in NSS keylog format.
4267 *Martin Kaiser <martin@kaiser.cx>*
4269 * Harmonize version and its documentation. -f flag is used to display
4272 *mancha <mancha1@zoho.com>*
4274 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4275 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4277 *mancha <mancha1@zoho.com>*
4279 * Fix some double frees. These are not thought to be exploitable.
4281 *mancha <mancha1@zoho.com>*
4283 * A missing bounds check in the handling of the TLS heartbeat extension
4284 can be used to reveal up to 64k of memory to a connected client or
4287 Thanks for Neel Mehta of Google Security for discovering this bug and to
4288 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4289 preparing the fix ([CVE-2014-0160])
4291 *Adam Langley, Bodo Moeller*
4293 * Fix for the attack described in the paper "Recovering OpenSSL
4294 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4295 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4296 <http://eprint.iacr.org/2014/140>
4298 Thanks to Yuval Yarom and Naomi Benger for discovering this
4299 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4301 *Yuval Yarom and Naomi Benger*
4303 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4304 this fixes a limitation in previous versions of OpenSSL.
4308 * Experimental encrypt-then-mac support.
4310 Experimental support for encrypt then mac from
4311 draft-gutmann-tls-encrypt-then-mac-02.txt
4313 To enable it set the appropriate extension number (0x42 for the test
4314 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4316 For non-compliant peers (i.e. just about everything) this should have no
4319 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4323 * Add EVP support for key wrapping algorithms, to avoid problems with
4324 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4325 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4326 algorithms and include tests cases.
4330 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4335 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4336 MGF1 digest and OAEP label.
4340 * Make openssl verify return errors.
4342 *Chris Palmer <palmer@google.com> and Ben Laurie*
4344 * New function ASN1_TIME_diff to calculate the difference between two
4345 ASN1_TIME structures or one structure and the current time.
4349 * Update fips_test_suite to support multiple command line options. New
4350 test to induce all self test errors in sequence and check expected
4355 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4356 sign or verify all in one operation.
4360 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4361 test programs and fips_test_suite. Includes functionality to parse
4362 the minimal script output of fipsalgest.pl directly.
4366 * Add authorisation parameter to FIPS_module_mode_set().
4370 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4374 * Use separate DRBG fields for internal and external flags. New function
4375 FIPS_drbg_health_check() to perform on demand health checking. Add
4376 generation tests to fips_test_suite with reduced health check interval to
4377 demonstrate periodic health checking. Add "nodh" option to
4378 fips_test_suite to skip very slow DH test.
4382 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4387 * More extensive health check for DRBG checking many more failure modes.
4388 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4389 combination: call this in fips_test_suite.
4393 * Add support for canonical generation of DSA parameter 'g'. See
4396 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4397 POST to handle HMAC cases.
4401 * Add functions FIPS_module_version() and FIPS_module_version_text()
4402 to return numerical and string versions of the FIPS module number.
4406 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4407 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4408 outside the validated module in the FIPS capable OpenSSL.
4412 * Minor change to DRBG entropy callback semantics. In some cases
4413 there is no multiple of the block length between min_len and
4414 max_len. Allow the callback to return more than max_len bytes
4415 of entropy but discard any extra: it is the callback's responsibility
4416 to ensure that the extra data discarded does not impact the
4417 requested amount of entropy.
4421 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4422 information in FIPS186-3, SP800-57 and SP800-131A.
4426 * CCM support via EVP. Interface is very similar to GCM case except we
4427 must supply all data in one chunk (i.e. no update, final) and the
4428 message length must be supplied if AAD is used. Add algorithm test
4433 * Initial version of POST overhaul. Add POST callback to allow the status
4434 of POST to be monitored and/or failures induced. Modify fips_test_suite
4435 to use callback. Always run all selftests even if one fails.
4439 * XTS support including algorithm test driver in the fips_gcmtest program.
4440 Note: this does increase the maximum key length from 32 to 64 bytes but
4441 there should be no binary compatibility issues as existing applications
4442 will never use XTS mode.
4446 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4447 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4448 performs algorithm blocking for unapproved PRNG types. Also do not
4449 set PRNG type in FIPS_mode_set(): leave this to the application.
4450 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4451 the standard OpenSSL PRNG: set additional data to a date time vector.
4455 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4456 This shouldn't present any incompatibility problems because applications
4457 shouldn't be using these directly and any that are will need to rethink
4458 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4462 * Extensive self tests and health checking required by SP800-90 DRBG.
4463 Remove strength parameter from FIPS_drbg_instantiate and always
4464 instantiate at maximum supported strength.
4468 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4472 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4476 * New function DH_compute_key_padded() to compute a DH key and pad with
4477 leading zeroes if needed: this complies with SP800-56A et al.
4481 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4482 anything, incomplete, subject to change and largely untested at present.
4486 * Modify fipscanisteronly build option to only build the necessary object
4487 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4491 * Add experimental option FIPSSYMS to give all symbols in
4492 fipscanister.o and FIPS or fips prefix. This will avoid
4493 conflicts with future versions of OpenSSL. Add perl script
4494 util/fipsas.pl to preprocess assembly language source files
4495 and rename any affected symbols.
4499 * Add selftest checks and algorithm block of non-fips algorithms in
4500 FIPS mode. Remove DES2 from selftests.
4504 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4505 return internal method without any ENGINE dependencies. Add new
4506 tiny fips sign and verify functions.
4510 * New build option no-ec2m to disable characteristic 2 code.
4514 * New build option "fipscanisteronly". This only builds fipscanister.o
4515 and (currently) associated fips utilities. Uses the file Makefile.fips
4516 instead of Makefile.org as the prototype.
4520 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4521 Update fips_gcmtest to use IV generator.
4525 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4526 setting output buffer to NULL. The `*Final` function must be
4527 called although it will not retrieve any additional data. The tag
4528 can be set or retrieved with a ctrl. The IV length is by default 12
4529 bytes (96 bits) but can be set to an alternative value. If the IV
4530 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4535 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4536 underlying do_cipher function handles all cipher semantics itself
4537 including padding and finalisation. This is useful if (for example)
4538 an ENGINE cipher handles block padding itself. The behaviour of
4539 do_cipher is subtly changed if this flag is set: the return value
4540 is the number of characters written to the output buffer (zero is
4541 no longer an error code) or a negative error code. Also if the
4542 input buffer is NULL and length 0 finalisation should be performed.
4546 * If a candidate issuer certificate is already part of the constructed
4547 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4551 * Improve forward-security support: add functions
4553 void SSL_CTX_set_not_resumable_session_callback(
4554 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4555 void SSL_set_not_resumable_session_callback(
4556 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4558 for use by SSL/TLS servers; the callback function will be called whenever a
4559 new session is created, and gets to decide whether the session may be
4560 cached to make it resumable (return 0) or not (return 1). (As by the
4561 SSL/TLS protocol specifications, the session_id sent by the server will be
4562 empty to indicate that the session is not resumable; also, the server will
4563 not generate RFC 4507 (RFC 5077) session tickets.)
4565 A simple reasonable callback implementation is to return is_forward_secure.
4566 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4567 by the SSL/TLS server library, indicating whether it can provide forward
4570 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4572 * New -verify_name option in command line utilities to set verification
4577 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4578 Add CMAC pkey methods.
4582 * Experimental renegotiation in s_server -www mode. If the client
4583 browses /reneg connection is renegotiated. If /renegcert it is
4584 renegotiated requesting a certificate.
4588 * Add an "external" session cache for debugging purposes to s_server. This
4589 should help trace issues which normally are only apparent in deployed
4590 multi-process servers.
4594 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4595 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4596 BIO_set_cipher() and some obscure PEM functions were changed so they
4597 can now return an error. The RAND changes required a change to the
4598 RAND_METHOD structure.
4602 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4603 a gcc attribute to warn if the result of a function is ignored. This
4604 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4605 whose return value is often ignored.
4609 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4610 These allow SCTs (signed certificate timestamps) to be requested and
4611 validated when establishing a connection.
4613 *Rob Percival <robpercival@google.com>*
4618 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4620 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4621 used even when parsing explicit parameters, when loading a encoded key
4622 or calling `EC_GROUP_new_from_ecpkparameters()`/
4623 `EC_GROUP_new_from_ecparameters()`.
4624 This prevents bypass of security hardening and performance gains,
4625 especially for curves with specialized EC_METHODs.
4626 By default, if a key encoded with explicit parameters is loaded and later
4627 encoded, the output is still encoded with explicit parameters, even if
4628 internally a "named" EC_GROUP is used for computation.
4632 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4633 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4634 NULL. After this change, only the cofactor parameter can be NULL. It also
4635 does some minimal sanity checks on the passed order.
4640 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4641 An attack is simple, if the first CMS_recipientInfo is valid but the
4642 second CMS_recipientInfo is chosen ciphertext. If the second
4643 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4644 encryption key will be replaced by garbage, and the message cannot be
4645 decoded, but if the RSA decryption fails, the correct encryption key is
4646 used and the recipient will not notice the attack.
4647 As a work around for this potential attack the length of the decrypted
4648 key must be equal to the cipher default key length, in case the
4649 certifiate is not given and all recipientInfo are tried out.
4650 The old behaviour can be re-enabled in the CMS code by setting the
4651 CMS_DEBUG_DECRYPT flag.
4656 * Document issue with installation paths in diverse Windows builds
4658 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4659 binaries and run-time config file.
4664 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4666 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4667 This changes the size when using the `genpkey` command when no size is given.
4668 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4669 generation commands to use 2048 bits by default.
4673 * Add FIPS support for Android Arm 64-bit
4675 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4676 Module in Version 2.0.10. For some reason, the corresponding target
4677 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4678 built with FIPS support on Android Arm 64-bit. This omission has been
4681 *Matthias St. Pierre*
4683 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4685 * 0-byte record padding oracle
4687 If an application encounters a fatal protocol error and then calls
4688 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4689 then OpenSSL can respond differently to the calling application if a 0 byte
4690 record is received with invalid padding compared to if a 0 byte record is
4691 received with an invalid MAC. If the application then behaves differently
4692 based on that in a way that is detectable to the remote peer, then this
4693 amounts to a padding oracle that could be used to decrypt data.
4695 In order for this to be exploitable "non-stitched" ciphersuites must be in
4696 use. Stitched ciphersuites are optimised implementations of certain
4697 commonly used ciphersuites. Also the application must call SSL_shutdown()
4698 twice even if a protocol error has occurred (applications should not do
4699 this but some do anyway).
4701 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4702 Aviram, with additional investigation by Steven Collison and Andrew
4703 Hourselt. It was reported to OpenSSL on 10th December 2018.
4708 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4712 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4714 * Microarchitecture timing vulnerability in ECC scalar multiplication
4716 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4717 shown to be vulnerable to a microarchitecture timing side channel attack.
4718 An attacker with sufficient access to mount local timing attacks during
4719 ECDSA signature generation could recover the private key.
4721 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4722 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4728 * Timing vulnerability in DSA signature generation
4730 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4731 timing side channel attack. An attacker could use variations in the signing
4732 algorithm to recover the private key.
4734 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4739 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4740 Module, accidentally introduced while backporting security fixes from the
4741 development branch and hindering the use of ECC in FIPS mode.
4745 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4747 * Client DoS due to large DH parameter
4749 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4750 malicious server can send a very large prime value to the client. This will
4751 cause the client to spend an unreasonably long period of time generating a
4752 key for this prime resulting in a hang until the client has finished. This
4753 could be exploited in a Denial Of Service attack.
4755 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4760 * Cache timing vulnerability in RSA Key Generation
4762 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4763 a cache timing side channel attack. An attacker with sufficient access to
4764 mount cache timing attacks during the RSA key generation process could
4765 recover the private key.
4767 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4768 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4773 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4774 parameter is no longer accepted, as it leads to a corrupt table. NULL
4775 pem_str is reserved for alias entries only.
4779 * Revert blinding in ECDSA sign and instead make problematic addition
4780 length-invariant. Switch even to fixed-length Montgomery multiplication.
4784 * Change generating and checking of primes so that the error rate of not
4785 being prime depends on the intended use based on the size of the input.
4786 For larger primes this will result in more rounds of Miller-Rabin.
4787 The maximal error rate for primes with more than 1080 bits is lowered
4790 *Kurt Roeckx, Annie Yousar*
4792 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4796 * Add blinding to ECDSA and DSA signatures to protect against side channel
4797 attacks discovered by Keegan Ryan (NCC Group).
4801 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4802 now allow empty (zero character) pass phrases.
4806 * Certificate time validation (X509_cmp_time) enforces stricter
4807 compliance with RFC 5280. Fractional seconds and timezone offsets
4808 are no longer allowed.
4812 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4814 * Constructed ASN.1 types with a recursive definition could exceed the stack
4816 Constructed ASN.1 types with a recursive definition (such as can be found
4817 in PKCS7) could eventually exceed the stack given malicious input with
4818 excessive recursion. This could result in a Denial Of Service attack. There
4819 are no such structures used within SSL/TLS that come from untrusted sources
4820 so this is considered safe.
4822 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4828 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4830 * Read/write after SSL object in error state
4832 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4833 mechanism. The intent was that if a fatal error occurred during a handshake
4834 then OpenSSL would move into the error state and would immediately fail if
4835 you attempted to continue the handshake. This works as designed for the
4836 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4837 SSL_connect()), however due to a bug it does not work correctly if
4838 SSL_read() or SSL_write() is called directly. In that scenario, if the
4839 handshake fails then a fatal error will be returned in the initial function
4840 call. If SSL_read()/SSL_write() is subsequently called by the application
4841 for the same SSL object then it will succeed and the data is passed without
4842 being decrypted/encrypted directly from the SSL/TLS record layer.
4844 In order to exploit this issue an application bug would have to be present
4845 that resulted in a call to SSL_read()/SSL_write() being issued after having
4846 already received a fatal error.
4848 This issue was reported to OpenSSL by David Benjamin (Google).
4853 * rsaz_1024_mul_avx2 overflow bug on x86_64
4855 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4856 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4857 Analysis suggests that attacks against RSA and DSA as a result of this
4858 defect would be very difficult to perform and are not believed likely.
4859 Attacks against DH1024 are considered just feasible, because most of the
4860 work necessary to deduce information about a private key may be performed
4861 offline. The amount of resources required for such an attack would be
4862 significant. However, for an attack on TLS to be meaningful, the server
4863 would have to share the DH1024 private key among multiple clients, which is
4864 no longer an option since CVE-2016-0701.
4866 This only affects processors that support the AVX2 but not ADX extensions
4867 like Intel Haswell (4th generation).
4869 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4870 was originally found via the OSS-Fuzz project.
4875 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4877 * bn_sqrx8x_internal carry bug on x86_64
4879 There is a carry propagating bug in the x86_64 Montgomery squaring
4880 procedure. No EC algorithms are affected. Analysis suggests that attacks
4881 against RSA and DSA as a result of this defect would be very difficult to
4882 perform and are not believed likely. Attacks against DH are considered just
4883 feasible (although very difficult) because most of the work necessary to
4884 deduce information about a private key may be performed offline. The amount
4885 of resources required for such an attack would be very significant and
4886 likely only accessible to a limited number of attackers. An attacker would
4887 additionally need online access to an unpatched system using the target
4888 private key in a scenario with persistent DH parameters and a private
4889 key that is shared between multiple clients.
4891 This only affects processors that support the BMI1, BMI2 and ADX extensions
4892 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4894 This issue was reported to OpenSSL by the OSS-Fuzz project.
4899 * Malformed X.509 IPAddressFamily could cause OOB read
4901 If an X.509 certificate has a malformed IPAddressFamily extension,
4902 OpenSSL could do a one-byte buffer overread. The most likely result
4903 would be an erroneous display of the certificate in text format.
4905 This issue was reported to OpenSSL by the OSS-Fuzz project.
4910 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4912 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4913 platform rather than 'mingw'.
4917 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4919 * Truncated packet could crash via OOB read
4921 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4922 cipher is being used, then a truncated packet can cause that host to
4923 perform an out-of-bounds read, usually resulting in a crash.
4925 This issue was reported to OpenSSL by Robert Święcki of Google.
4930 * BN_mod_exp may produce incorrect results on x86_64
4932 There is a carry propagating bug in the x86_64 Montgomery squaring
4933 procedure. No EC algorithms are affected. Analysis suggests that attacks
4934 against RSA and DSA as a result of this defect would be very difficult to
4935 perform and are not believed likely. Attacks against DH are considered just
4936 feasible (although very difficult) because most of the work necessary to
4937 deduce information about a private key may be performed offline. The amount
4938 of resources required for such an attack would be very significant and
4939 likely only accessible to a limited number of attackers. An attacker would
4940 additionally need online access to an unpatched system using the target
4941 private key in a scenario with persistent DH parameters and a private
4942 key that is shared between multiple clients. For example this can occur by
4943 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4944 similar to CVE-2015-3193 but must be treated as a separate problem.
4946 This issue was reported to OpenSSL by the OSS-Fuzz project.
4951 * Montgomery multiplication may produce incorrect results
4953 There is a carry propagating bug in the Broadwell-specific Montgomery
4954 multiplication procedure that handles input lengths divisible by, but
4955 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4956 and DH private keys are impossible. This is because the subroutine in
4957 question is not used in operations with the private key itself and an input
4958 of the attacker's direct choice. Otherwise the bug can manifest itself as
4959 transient authentication and key negotiation failures or reproducible
4960 erroneous outcome of public-key operations with specially crafted input.
4961 Among EC algorithms only Brainpool P-512 curves are affected and one
4962 presumably can attack ECDH key negotiation. Impact was not analyzed in
4963 detail, because pre-requisites for attack are considered unlikely. Namely
4964 multiple clients have to choose the curve in question and the server has to
4965 share the private key among them, neither of which is default behaviour.
4966 Even then only clients that chose the curve will be affected.
4968 This issue was publicly reported as transient failures and was not
4969 initially recognized as a security issue. Thanks to Richard Morgan for
4970 providing reproducible case.
4975 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4976 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4977 prevent issues where no progress is being made and the peer continually
4978 sends unrecognised record types, using up resources processing them.
4982 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4984 * Missing CRL sanity check
4986 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4987 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4988 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4990 This issue only affects the OpenSSL 1.0.2i
4995 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4997 * OCSP Status Request extension unbounded memory growth
4999 A malicious client can send an excessively large OCSP Status Request
5000 extension. If that client continually requests renegotiation, sending a
5001 large OCSP Status Request extension each time, then there will be unbounded
5002 memory growth on the server. This will eventually lead to a Denial Of
5003 Service attack through memory exhaustion. Servers with a default
5004 configuration are vulnerable even if they do not support OCSP. Builds using
5005 the "no-ocsp" build time option are not affected.
5007 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5012 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5015 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5021 * OOB write in MDC2_Update()
5023 An overflow can occur in MDC2_Update() either if called directly or
5024 through the EVP_DigestUpdate() function using MDC2. If an attacker
5025 is able to supply very large amounts of input data after a previous
5026 call to EVP_EncryptUpdate() with a partial block then a length check
5027 can overflow resulting in a heap corruption.
5029 The amount of data needed is comparable to SIZE_MAX which is impractical
5032 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5037 * Malformed SHA512 ticket DoS
5039 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5040 DoS attack where a malformed ticket will result in an OOB read which will
5043 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5044 a custom server callback and ticket lookup mechanism.
5046 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5051 * OOB write in BN_bn2dec()
5053 The function BN_bn2dec() does not check the return value of BN_div_word().
5054 This can cause an OOB write if an application uses this function with an
5055 overly large BIGNUM. This could be a problem if an overly large certificate
5056 or CRL is printed out from an untrusted source. TLS is not affected because
5057 record limits will reject an oversized certificate before it is parsed.
5059 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5064 * OOB read in TS_OBJ_print_bio()
5066 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5067 the total length the OID text representation would use and not the amount
5068 of data written. This will result in OOB reads when large OIDs are
5071 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5076 * Pointer arithmetic undefined behaviour
5078 Avoid some undefined pointer arithmetic
5080 A common idiom in the codebase is to check limits in the following manner:
5083 Where "p" points to some malloc'd data of SIZE bytes and
5086 "len" here could be from some externally supplied data (e.g. from a TLS
5089 The rules of C pointer arithmetic are such that "p + len" is only well
5090 defined where len <= SIZE. Therefore the above idiom is actually
5091 undefined behaviour.
5093 For example this could cause problems if some malloc implementation
5094 provides an address for "p" such that "p + len" actually overflows for
5095 values of len that are too big and therefore p + len < limit.
5097 This issue was reported to OpenSSL by Guido Vranken
5102 * Constant time flag not preserved in DSA signing
5104 Operations in the DSA signing algorithm should run in constant time in
5105 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5106 implementation means that a non-constant time codepath is followed for
5107 certain operations. This has been demonstrated through a cache-timing
5108 attack to be sufficient for an attacker to recover the private DSA key.
5110 This issue was reported by César Pereida (Aalto University), Billy Brumley
5111 (Tampere University of Technology), and Yuval Yarom (The University of
5112 Adelaide and NICTA).
5117 * DTLS buffered message DoS
5119 In a DTLS connection where handshake messages are delivered out-of-order
5120 those messages that OpenSSL is not yet ready to process will be buffered
5121 for later use. Under certain circumstances, a flaw in the logic means that
5122 those messages do not get removed from the buffer even though the handshake
5123 has been completed. An attacker could force up to approx. 15 messages to
5124 remain in the buffer when they are no longer required. These messages will
5125 be cleared when the DTLS connection is closed. The default maximum size for
5126 a message is 100k. Therefore the attacker could force an additional 1500k
5127 to be consumed per connection. By opening many simulataneous connections an
5128 attacker could cause a DoS attack through memory exhaustion.
5130 This issue was reported to OpenSSL by Quan Luo.
5135 * DTLS replay protection DoS
5137 A flaw in the DTLS replay attack protection mechanism means that records
5138 that arrive for future epochs update the replay protection "window" before
5139 the MAC for the record has been validated. This could be exploited by an
5140 attacker by sending a record for the next epoch (which does not have to
5141 decrypt or have a valid MAC), with a very large sequence number. This means
5142 that all subsequent legitimate packets are dropped causing a denial of
5143 service for a specific DTLS connection.
5145 This issue was reported to OpenSSL by the OCAP audit team.
5150 * Certificate message OOB reads
5152 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5153 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5154 theoretical DoS risk but this has not been observed in practice on common
5157 The messages affected are client certificate, client certificate request
5158 and server certificate. As a result the attack can only be performed
5159 against a client or a server which enables client authentication.
5161 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5166 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5168 * Prevent padding oracle in AES-NI CBC MAC check
5170 A MITM attacker can use a padding oracle attack to decrypt traffic
5171 when the connection uses an AES CBC cipher and the server support
5174 This issue was introduced as part of the fix for Lucky 13 padding
5175 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5176 constant time by making sure that always the same bytes are read and
5177 compared against either the MAC or padding bytes. But it no longer
5178 checked that there was enough data to have both the MAC and padding
5181 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5186 * Fix EVP_EncodeUpdate overflow
5188 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5189 Base64 encoding of binary data. If an attacker is able to supply very large
5190 amounts of input data then a length check can overflow resulting in a heap
5193 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5194 the `PEM_write_bio*` family of functions. These are mainly used within the
5195 OpenSSL command line applications, so any application which processes data
5196 from an untrusted source and outputs it as a PEM file should be considered
5197 vulnerable to this issue. User applications that call these APIs directly
5198 with large amounts of untrusted data may also be vulnerable.
5200 This issue was reported by Guido Vranken.
5205 * Fix EVP_EncryptUpdate overflow
5207 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5208 is able to supply very large amounts of input data after a previous call to
5209 EVP_EncryptUpdate() with a partial block then a length check can overflow
5210 resulting in a heap corruption. Following an analysis of all OpenSSL
5211 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5212 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5213 the first called function after an EVP_EncryptInit(), and therefore that
5214 specific call must be safe. The second form is where the length passed to
5215 EVP_EncryptUpdate() can be seen from the code to be some small value and
5216 therefore there is no possibility of an overflow. Since all instances are
5217 one of these two forms, it is believed that there can be no overflows in
5218 internal code due to this problem. It should be noted that
5219 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5220 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5221 of these calls have also been analysed too and it is believed there are no
5222 instances in internal usage where an overflow could occur.
5224 This issue was reported by Guido Vranken.
5229 * Prevent ASN.1 BIO excessive memory allocation
5231 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5232 a short invalid encoding can cause allocation of large amounts of memory
5233 potentially consuming excessive resources or exhausting memory.
5235 Any application parsing untrusted data through d2i BIO functions is
5236 affected. The memory based functions such as d2i_X509() are *not* affected.
5237 Since the memory based functions are used by the TLS library, TLS
5238 applications are not affected.
5240 This issue was reported by Brian Carpenter.
5247 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5248 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5249 in arbitrary stack data being returned in the buffer.
5251 This issue was reported by Guido Vranken.
5256 * Modify behavior of ALPN to invoke callback after SNI/servername
5257 callback, such that updates to the SSL_CTX affect ALPN.
5261 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5266 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5267 methods are enabled and ssl2 is disabled the methods return NULL.
5271 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5273 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5274 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5275 provide any "EXPORT" or "LOW" strength ciphers.
5279 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5280 is by default disabled at build-time. Builds that are not configured with
5281 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5282 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5283 will need to explicitly call either of:
5285 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5287 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5289 as appropriate. Even if either of those is used, or the application
5290 explicitly uses the version-specific SSLv2_method() or its client and
5291 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5292 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5293 ciphers, and SSLv2 56-bit DES are no longer available.
5298 * Fix a double-free in DSA code
5300 A double free bug was discovered when OpenSSL parses malformed DSA private
5301 keys and could lead to a DoS attack or memory corruption for applications
5302 that receive DSA private keys from untrusted sources. This scenario is
5305 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5311 * Disable SRP fake user seed to address a server memory leak.
5313 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5315 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5316 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5317 was changed to ignore the "fake user" SRP seed, even if the seed
5320 Users should use SRP_VBASE_get1_by_user instead. Note that in
5321 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5322 also that even though configuring the SRP seed attempts to hide
5323 invalid usernames by continuing the handshake with fake
5324 credentials, this behaviour is not constant time and no strong
5325 guarantees are made that the handshake is indistinguishable from
5326 that of a valid user.
5331 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5333 In the BN_hex2bn function the number of hex digits is calculated using an
5334 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5335 large values of `i` this can result in `bn_expand` not allocating any
5336 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5337 field as NULL leading to a subsequent NULL ptr deref. For very large values
5338 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5339 In this case memory is allocated to the internal BIGNUM data field, but it
5340 is insufficiently sized leading to heap corruption. A similar issue exists
5341 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5342 is ever called by user applications with very large untrusted hex/dec data.
5343 This is anticipated to be a rare occurrence.
5345 All OpenSSL internal usage of these functions use data that is not expected
5346 to be untrusted, e.g. config file data or application command line
5347 arguments. If user developed applications generate config file data based
5348 on untrusted data then it is possible that this could also lead to security
5349 consequences. This is also anticipated to be rare.
5351 This issue was reported to OpenSSL by Guido Vranken.
5356 * Fix memory issues in `BIO_*printf` functions
5358 The internal `fmtstr` function used in processing a "%s" format string in
5359 the `BIO_*printf` functions could overflow while calculating the length of a
5360 string and cause an OOB read when printing very long strings.
5362 Additionally the internal `doapr_outch` function can attempt to write to an
5363 OOB memory location (at an offset from the NULL pointer) in the event of a
5364 memory allocation failure. In 1.0.2 and below this could be caused where
5365 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5366 could be in processing a very long "%s" format string. Memory leaks can
5369 The first issue may mask the second issue dependent on compiler behaviour.
5370 These problems could enable attacks where large amounts of untrusted data
5371 is passed to the `BIO_*printf` functions. If applications use these functions
5372 in this way then they could be vulnerable. OpenSSL itself uses these
5373 functions when printing out human-readable dumps of ASN.1 data. Therefore
5374 applications that print this data could be vulnerable if the data is from
5375 untrusted sources. OpenSSL command line applications could also be
5376 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5377 as command line arguments.
5379 Libssl is not considered directly vulnerable. Additionally certificates etc
5380 received via remote connections via libssl are also unlikely to be able to
5381 trigger these issues because of message size limits enforced within libssl.
5383 This issue was reported to OpenSSL Guido Vranken.
5388 * Side channel attack on modular exponentiation
5390 A side-channel attack was found which makes use of cache-bank conflicts on
5391 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5392 of RSA keys. The ability to exploit this issue is limited as it relies on
5393 an attacker who has control of code in a thread running on the same
5394 hyper-threaded core as the victim thread which is performing decryptions.
5396 This issue was reported to OpenSSL by Yuval Yarom, The University of
5397 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5398 Nadia Heninger, University of Pennsylvania with more information at
5399 <http://cachebleed.info>.
5404 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5405 if no keysize is specified with default_bits. This fixes an
5406 omission in an earlier change that changed all RSA/DSA key generation
5407 commands to use 2048 bits by default.
5411 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5413 * DH small subgroups
5415 Historically OpenSSL only ever generated DH parameters based on "safe"
5416 primes. More recently (in version 1.0.2) support was provided for
5417 generating X9.42 style parameter files such as those required for RFC 5114
5418 support. The primes used in such files may not be "safe". Where an
5419 application is using DH configured with parameters based on primes that are
5420 not "safe" then an attacker could use this fact to find a peer's private
5421 DH exponent. This attack requires that the attacker complete multiple
5422 handshakes in which the peer uses the same private DH exponent. For example
5423 this could be used to discover a TLS server's private DH exponent if it's
5424 reusing the private DH exponent or it's using a static DH ciphersuite.
5426 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5427 TLS. It is not on by default. If the option is not set then the server
5428 reuses the same private DH exponent for the life of the server process and
5429 would be vulnerable to this attack. It is believed that many popular
5430 applications do set this option and would therefore not be at risk.
5432 The fix for this issue adds an additional check where a "q" parameter is
5433 available (as is the case in X9.42 based parameters). This detects the
5434 only known attack, and is the only possible defense for static DH
5435 ciphersuites. This could have some performance impact.
5437 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5438 default and cannot be disabled. This could have some performance impact.
5440 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5445 * SSLv2 doesn't block disabled ciphers
5447 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5448 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5449 been disabled, provided that the SSLv2 protocol was not also disabled via
5452 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5453 and Sebastian Schinzel.
5458 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5460 * BN_mod_exp may produce incorrect results on x86_64
5462 There is a carry propagating bug in the x86_64 Montgomery squaring
5463 procedure. No EC algorithms are affected. Analysis suggests that attacks
5464 against RSA and DSA as a result of this defect would be very difficult to
5465 perform and are not believed likely. Attacks against DH are considered just
5466 feasible (although very difficult) because most of the work necessary to
5467 deduce information about a private key may be performed offline. The amount
5468 of resources required for such an attack would be very significant and
5469 likely only accessible to a limited number of attackers. An attacker would
5470 additionally need online access to an unpatched system using the target
5471 private key in a scenario with persistent DH parameters and a private
5472 key that is shared between multiple clients. For example this can occur by
5473 default in OpenSSL DHE based SSL/TLS ciphersuites.
5475 This issue was reported to OpenSSL by Hanno Böck.
5480 * Certificate verify crash with missing PSS parameter
5482 The signature verification routines will crash with a NULL pointer
5483 dereference if presented with an ASN.1 signature using the RSA PSS
5484 algorithm and absent mask generation function parameter. Since these
5485 routines are used to verify certificate signature algorithms this can be
5486 used to crash any certificate verification operation and exploited in a
5487 DoS attack. Any application which performs certificate verification is
5488 vulnerable including OpenSSL clients and servers which enable client
5491 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5496 * X509_ATTRIBUTE memory leak
5498 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5499 memory. This structure is used by the PKCS#7 and CMS routines so any
5500 application which reads PKCS#7 or CMS data from untrusted sources is
5501 affected. SSL/TLS is not affected.
5503 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5509 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5510 This changes the decoding behaviour for some invalid messages,
5511 though the change is mostly in the more lenient direction, and
5512 legacy behaviour is preserved as much as possible.
5516 * In DSA_generate_parameters_ex, if the provided seed is too short,
5519 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5521 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5523 * Alternate chains certificate forgery
5525 During certificate verification, OpenSSL will attempt to find an
5526 alternative certificate chain if the first attempt to build such a chain
5527 fails. An error in the implementation of this logic can mean that an
5528 attacker could cause certain checks on untrusted certificates to be
5529 bypassed, such as the CA flag, enabling them to use a valid leaf
5530 certificate to act as a CA and "issue" an invalid certificate.
5532 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5537 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5539 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5540 incompatibility in the handling of HMAC. The previous ABI has now been
5545 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5547 * Malformed ECParameters causes infinite loop
5549 When processing an ECParameters structure OpenSSL enters an infinite loop
5550 if the curve specified is over a specially malformed binary polynomial
5553 This can be used to perform denial of service against any
5554 system which processes public keys, certificate requests or
5555 certificates. This includes TLS clients and TLS servers with
5556 client authentication enabled.
5558 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5563 * Exploitable out-of-bounds read in X509_cmp_time
5565 X509_cmp_time does not properly check the length of the ASN1_TIME
5566 string and can read a few bytes out of bounds. In addition,
5567 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5570 An attacker can use this to craft malformed certificates and CRLs of
5571 various sizes and potentially cause a segmentation fault, resulting in
5572 a DoS on applications that verify certificates or CRLs. TLS clients
5573 that verify CRLs are affected. TLS clients and servers with client
5574 authentication enabled may be affected if they use custom verification
5577 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5578 independently by Hanno Böck.
5583 * PKCS7 crash with missing EnvelopedContent
5585 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5586 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5587 with missing content and trigger a NULL pointer dereference on parsing.
5589 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5590 structures from untrusted sources are affected. OpenSSL clients and
5591 servers are not affected.
5593 This issue was reported to OpenSSL by Michal Zalewski (Google).
5598 * CMS verify infinite loop with unknown hash function
5600 When verifying a signedData message the CMS code can enter an infinite loop
5601 if presented with an unknown hash function OID. This can be used to perform
5602 denial of service against any system which verifies signedData messages using
5604 This issue was reported to OpenSSL by Johannes Bauer.
5609 * Race condition handling NewSessionTicket
5611 If a NewSessionTicket is received by a multi-threaded client when attempting to
5612 reuse a previous ticket then a race condition can occur potentially leading to
5613 a double free of the ticket data.
5618 * Only support 256-bit or stronger elliptic curves with the
5619 'ecdh_auto' setting (server) or by default (client). Of supported
5620 curves, prefer P-256 (both).
5624 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5626 * ClientHello sigalgs DoS fix
5628 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5629 invalid signature algorithms extension a NULL pointer dereference will
5630 occur. This can be exploited in a DoS attack against the server.
5632 This issue was was reported to OpenSSL by David Ramos of Stanford
5636 *Stephen Henson and Matt Caswell*
5638 * Multiblock corrupted pointer fix
5640 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5641 feature only applies on 64 bit x86 architecture platforms that support AES
5642 NI instructions. A defect in the implementation of "multiblock" can cause
5643 OpenSSL's internal write buffer to become incorrectly set to NULL when
5644 using non-blocking IO. Typically, when the user application is using a
5645 socket BIO for writing, this will only result in a failed connection.
5646 However if some other BIO is used then it is likely that a segmentation
5647 fault will be triggered, thus enabling a potential DoS attack.
5649 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5654 * Segmentation fault in DTLSv1_listen fix
5656 The DTLSv1_listen function is intended to be stateless and processes the
5657 initial ClientHello from many peers. It is common for user code to loop
5658 over the call to DTLSv1_listen until a valid ClientHello is received with
5659 an associated cookie. A defect in the implementation of DTLSv1_listen means
5660 that state is preserved in the SSL object from one invocation to the next
5661 that can lead to a segmentation fault. Errors processing the initial
5662 ClientHello can trigger this scenario. An example of such an error could be
5663 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5666 This issue was reported to OpenSSL by Per Allansson.
5671 * Segmentation fault in ASN1_TYPE_cmp fix
5673 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5674 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5675 certificate signature algorithm consistency this can be used to crash any
5676 certificate verification operation and exploited in a DoS attack. Any
5677 application which performs certificate verification is vulnerable including
5678 OpenSSL clients and servers which enable client authentication.
5683 * Segmentation fault for invalid PSS parameters fix
5685 The signature verification routines will crash with a NULL pointer
5686 dereference if presented with an ASN.1 signature using the RSA PSS
5687 algorithm and invalid parameters. Since these routines are used to verify
5688 certificate signature algorithms 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.
5693 This issue was was reported to OpenSSL by Brian Carpenter.
5698 * ASN.1 structure reuse memory corruption fix
5700 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5701 memory corruption via an invalid write. Such reuse is and has been
5702 strongly discouraged and is believed to be rare.
5704 Applications that parse structures containing CHOICE or ANY DEFINED BY
5705 components may be affected. Certificate parsing (d2i_X509 and related
5706 functions) are however not affected. OpenSSL clients and servers are
5712 * PKCS7 NULL pointer dereferences fix
5714 The PKCS#7 parsing code does not handle missing outer ContentInfo
5715 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5716 missing content and trigger a NULL pointer dereference on parsing.
5718 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5719 otherwise parse PKCS#7 structures from untrusted sources are
5720 affected. OpenSSL clients and servers are not affected.
5722 This issue was reported to OpenSSL by Michal Zalewski (Google).
5727 * DoS via reachable assert in SSLv2 servers fix
5729 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5730 servers that both support SSLv2 and enable export cipher suites by sending
5731 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5733 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5734 (OpenSSL development team).
5739 * Empty CKE with client auth and DHE fix
5741 If client auth is used then a server can seg fault in the event of a DHE
5742 ciphersuite being selected and a zero length ClientKeyExchange message
5743 being sent by the client. This could be exploited in a DoS attack.
5748 * Handshake with unseeded PRNG fix
5750 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5751 with an unseeded PRNG. The conditions are:
5752 - The client is on a platform where the PRNG has not been seeded
5753 automatically, and the user has not seeded manually
5754 - A protocol specific client method version has been used (i.e. not
5755 SSL_client_methodv23)
5756 - A ciphersuite is used that does not require additional random data from
5757 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5759 If the handshake succeeds then the client random that has been used will
5760 have been generated from a PRNG with insufficient entropy and therefore the
5761 output may be predictable.
5763 For example using the following command with an unseeded openssl will
5764 succeed on an unpatched platform:
5766 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5771 * Use After Free following d2i_ECPrivatekey error fix
5773 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5774 could cause a use after free condition. This, in turn, could cause a double
5775 free in several private key parsing functions (such as d2i_PrivateKey
5776 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5777 for applications that receive EC private keys from untrusted
5778 sources. This scenario is considered rare.
5780 This issue was discovered by the BoringSSL project and fixed in their
5786 * X509_to_X509_REQ NULL pointer deref fix
5788 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5789 the certificate key is invalid. This function is rarely used in practice.
5791 This issue was discovered by Brian Carpenter.
5796 * Removed the export ciphers from the DEFAULT ciphers
5800 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5802 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5803 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5804 So far those who have to target multiple platforms would compromise
5805 and argue that binary targeting say ARMv5 would still execute on
5806 ARMv8. "Universal" build resolves this compromise by providing
5807 near-optimal performance even on newer platforms.
5811 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5812 (other platforms pending).
5814 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5816 * Add support for the SignedCertificateTimestampList certificate and
5817 OCSP response extensions from RFC6962.
5821 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5822 for corner cases. (Certain input points at infinity could lead to
5823 bogus results, with non-infinity inputs mapped to infinity too.)
5827 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5828 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5829 common cases are optimized and there still is room for further
5830 improvements. Vector Permutation AES for Altivec is also added.
5834 * Add support for little-endian ppc64 Linux target.
5836 *Marcelo Cerri (IBM)*
5838 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5839 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5840 are optimized and there still is room for further improvements.
5841 Both 32- and 64-bit modes are supported.
5843 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5845 * Improved ARMv7 NEON support.
5849 * Support for SPARC Architecture 2011 crypto extensions, first
5850 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5851 SHA256/512, MD5, GHASH and modular exponentiation.
5853 *Andy Polyakov, David Miller*
5855 * Accelerated modular exponentiation for Intel processors, a.k.a.
5858 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5860 * Support for new and upcoming Intel processors, including AVX2,
5861 BMI and SHA ISA extensions. This includes additional "stitched"
5862 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5865 This work was sponsored by Intel Corp.
5869 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5870 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5871 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5875 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5876 this fixes a limitation in previous versions of OpenSSL.
5880 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5881 MGF1 digest and OAEP label.
5885 * Add EVP support for key wrapping algorithms, to avoid problems with
5886 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5887 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5888 algorithms and include tests cases.
5892 * Add functions to allocate and set the fields of an ECDSA_METHOD
5895 *Douglas E. Engert, Steve Henson*
5897 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5898 difference in days and seconds between two tm or ASN1_TIME structures.
5902 * Add -rev test option to s_server to just reverse order of characters
5903 received by client and send back to server. Also prints an abbreviated
5904 summary of the connection parameters.
5908 * New option -brief for s_client and s_server to print out a brief summary
5909 of connection parameters.
5913 * Add callbacks for arbitrary TLS extensions.
5915 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5917 * New option -crl_download in several openssl utilities to download CRLs
5918 from CRLDP extension in certificates.
5922 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5926 * New function X509_CRL_diff to generate a delta CRL from the difference
5927 of two full CRLs. Add support to "crl" utility.
5931 * New functions to set lookup_crls function and to retrieve
5932 X509_STORE from X509_STORE_CTX.
5936 * Print out deprecated issuer and subject unique ID fields in
5941 * Extend OCSP I/O functions so they can be used for simple general purpose
5942 HTTP as well as OCSP. New wrapper function which can be used to download
5943 CRLs using the OCSP API.
5947 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5951 * `SSL_CONF*` functions. These provide a common framework for application
5952 configuration using configuration files or command lines.
5956 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5957 message callback and prints the results. Needs compile time option
5958 "enable-ssl-trace". New options to s_client and s_server to enable
5963 * New ctrl and macro to retrieve supported points extensions.
5964 Print out extension in s_server and s_client.
5968 * New functions to retrieve certificate signature and signature
5973 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5978 * New Suite B modes for TLS code. These use and enforce the requirements
5979 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5980 only use Suite B curves. The Suite B modes can be set by using the
5981 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5985 * New chain verification flags for Suite B levels of security. Check
5986 algorithms are acceptable when flags are set in X509_verify_cert.
5990 * Make tls1_check_chain return a set of flags indicating checks passed
5991 by a certificate chain. Add additional tests to handle client
5992 certificates: checks for matching certificate type and issuer name
5997 * If an attempt is made to use a signature algorithm not in the peer
5998 preference list abort the handshake. If client has no suitable
5999 signature algorithms in response to a certificate request do not
6000 use the certificate.
6004 * If server EC tmp key is not in client preference list abort handshake.
6008 * Add support for certificate stores in CERT structure. This makes it
6009 possible to have different stores per SSL structure or one store in
6010 the parent SSL_CTX. Include distinct stores for certificate chain
6011 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6012 to build and store a certificate chain in CERT structure: returning
6013 an error if the chain cannot be built: this will allow applications
6014 to test if a chain is correctly configured.
6016 Note: if the CERT based stores are not set then the parent SSL_CTX
6017 store is used to retain compatibility with existing behaviour.
6021 * New function ssl_set_client_disabled to set a ciphersuite disabled
6022 mask based on the current session, check mask when sending client
6023 hello and checking the requested ciphersuite.
6027 * New ctrls to retrieve and set certificate types in a certificate
6028 request message. Print out received values in s_client. If certificate
6029 types is not set with custom values set sensible values based on
6030 supported signature algorithms.
6034 * Support for distinct client and server supported signature algorithms.
6038 * Add certificate callback. If set this is called whenever a certificate
6039 is required by client or server. An application can decide which
6040 certificate chain to present based on arbitrary criteria: for example
6041 supported signature algorithms. Add very simple example to s_server.
6042 This fixes many of the problems and restrictions of the existing client
6043 certificate callback: for example you can now clear an existing
6044 certificate and specify the whole chain.
6048 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6049 the certificate can be used for (if anything). Set valid_flags field
6050 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6051 to have similar checks in it.
6053 Add new "cert_flags" field to CERT structure and include a "strict mode".
6054 This enforces some TLS certificate requirements (such as only permitting
6055 certificate signature algorithms contained in the supported algorithms
6056 extension) which some implementations ignore: this option should be used
6057 with caution as it could cause interoperability issues.
6061 * Update and tidy signature algorithm extension processing. Work out
6062 shared signature algorithms based on preferences and peer algorithms
6063 and print them out in s_client and s_server. Abort handshake if no
6064 shared signature algorithms.
6068 * Add new functions to allow customised supported signature algorithms
6069 for SSL and SSL_CTX structures. Add options to s_client and s_server
6074 * New function SSL_certs_clear() to delete all references to certificates
6075 from an SSL structure. Before this once a certificate had been added
6076 it couldn't be removed.
6080 * Integrate hostname, email address and IP address checking with certificate
6081 verification. New verify options supporting checking in openssl utility.
6085 * Fixes and wildcard matching support to hostname and email checking
6086 functions. Add manual page.
6088 *Florian Weimer (Red Hat Product Security Team)*
6090 * New functions to check a hostname email or IP address against a
6091 certificate. Add options x509 utility to print results of checks against
6096 * Fix OCSP checking.
6098 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6100 * Initial experimental support for explicitly trusted non-root CAs.
6101 OpenSSL still tries to build a complete chain to a root but if an
6102 intermediate CA has a trust setting included that is used. The first
6103 setting is used: whether to trust (e.g., -addtrust option to the x509
6108 * Add -trusted_first option which attempts to find certificates in the
6109 trusted store even if an untrusted chain is also supplied.
6113 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6114 platform support for Linux and Android.
6118 * Support for linux-x32, ILP32 environment in x86_64 framework.
6122 * Experimental multi-implementation support for FIPS capable OpenSSL.
6123 When in FIPS mode the approved implementations are used as normal,
6124 when not in FIPS mode the internal unapproved versions are used instead.
6125 This means that the FIPS capable OpenSSL isn't forced to use the
6126 (often lower performance) FIPS implementations outside FIPS mode.
6130 * Transparently support X9.42 DH parameters when calling
6131 PEM_read_bio_DHparameters. This means existing applications can handle
6132 the new parameter format automatically.
6136 * Initial experimental support for X9.42 DH parameter format: mainly
6137 to support use of 'q' parameter for RFC5114 parameters.
6141 * Add DH parameters from RFC5114 including test data to dhtest.
6145 * Support for automatic EC temporary key parameter selection. If enabled
6146 the most preferred EC parameters are automatically used instead of
6147 hardcoded fixed parameters. Now a server just has to call:
6148 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6149 support ECDH and use the most appropriate parameters.
6153 * Enhance and tidy EC curve and point format TLS extension code. Use
6154 static structures instead of allocation if default values are used.
6155 New ctrls to set curves we wish to support and to retrieve shared curves.
6156 Print out shared curves in s_server. New options to s_server and s_client
6157 to set list of supported curves.
6161 * New ctrls to retrieve supported signature algorithms and
6162 supported curve values as an array of NIDs. Extend openssl utility
6163 to print out received values.
6167 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6168 between NIDs and the more common NIST names such as "P-256". Enhance
6169 ecparam utility and ECC method to recognise the NIST names for curves.
6173 * Enhance SSL/TLS certificate chain handling to support different
6174 chains for each certificate instead of one chain in the parent SSL_CTX.
6178 * Support for fixed DH ciphersuite client authentication: where both
6179 server and client use DH certificates with common parameters.
6183 * Support for fixed DH ciphersuites: those requiring DH server
6188 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6190 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6191 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6192 X509_CINF_get_signature were reverted post internal team review.
6197 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6199 * OCSP Status Request extension unbounded memory growth
6201 A malicious client can send an excessively large OCSP Status Request
6202 extension. If that client continually requests renegotiation, sending a
6203 large OCSP Status Request extension each time, then there will be unbounded
6204 memory growth on the server. This will eventually lead to a Denial Of
6205 Service attack through memory exhaustion. Servers with a default
6206 configuration are vulnerable even if they do not support OCSP. Builds using
6207 the "no-ocsp" build time option are not affected.
6209 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6214 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6217 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6223 * OOB write in MDC2_Update()
6225 An overflow can occur in MDC2_Update() either if called directly or
6226 through the EVP_DigestUpdate() function using MDC2. If an attacker
6227 is able to supply very large amounts of input data after a previous
6228 call to EVP_EncryptUpdate() with a partial block then a length check
6229 can overflow resulting in a heap corruption.
6231 The amount of data needed is comparable to SIZE_MAX which is impractical
6234 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6239 * Malformed SHA512 ticket DoS
6241 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6242 DoS attack where a malformed ticket will result in an OOB read which will
6245 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6246 a custom server callback and ticket lookup mechanism.
6248 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6253 * OOB write in BN_bn2dec()
6255 The function BN_bn2dec() does not check the return value of BN_div_word().
6256 This can cause an OOB write if an application uses this function with an
6257 overly large BIGNUM. This could be a problem if an overly large certificate
6258 or CRL is printed out from an untrusted source. TLS is not affected because
6259 record limits will reject an oversized certificate before it is parsed.
6261 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6266 * OOB read in TS_OBJ_print_bio()
6268 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6269 the total length the OID text representation would use and not the amount
6270 of data written. This will result in OOB reads when large OIDs are
6273 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6278 * Pointer arithmetic undefined behaviour
6280 Avoid some undefined pointer arithmetic
6282 A common idiom in the codebase is to check limits in the following manner:
6285 Where "p" points to some malloc'd data of SIZE bytes and
6288 "len" here could be from some externally supplied data (e.g. from a TLS
6291 The rules of C pointer arithmetic are such that "p + len" is only well
6292 defined where len <= SIZE. Therefore the above idiom is actually
6293 undefined behaviour.
6295 For example this could cause problems if some malloc implementation
6296 provides an address for "p" such that "p + len" actually overflows for
6297 values of len that are too big and therefore p + len < limit.
6299 This issue was reported to OpenSSL by Guido Vranken
6304 * Constant time flag not preserved in DSA signing
6306 Operations in the DSA signing algorithm should run in constant time in
6307 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6308 implementation means that a non-constant time codepath is followed for
6309 certain operations. This has been demonstrated through a cache-timing
6310 attack to be sufficient for an attacker to recover the private DSA key.
6312 This issue was reported by César Pereida (Aalto University), Billy Brumley
6313 (Tampere University of Technology), and Yuval Yarom (The University of
6314 Adelaide and NICTA).
6319 * DTLS buffered message DoS
6321 In a DTLS connection where handshake messages are delivered out-of-order
6322 those messages that OpenSSL is not yet ready to process will be buffered
6323 for later use. Under certain circumstances, a flaw in the logic means that
6324 those messages do not get removed from the buffer even though the handshake
6325 has been completed. An attacker could force up to approx. 15 messages to
6326 remain in the buffer when they are no longer required. These messages will
6327 be cleared when the DTLS connection is closed. The default maximum size for
6328 a message is 100k. Therefore the attacker could force an additional 1500k
6329 to be consumed per connection. By opening many simulataneous connections an
6330 attacker could cause a DoS attack through memory exhaustion.
6332 This issue was reported to OpenSSL by Quan Luo.
6337 * DTLS replay protection DoS
6339 A flaw in the DTLS replay attack protection mechanism means that records
6340 that arrive for future epochs update the replay protection "window" before
6341 the MAC for the record has been validated. This could be exploited by an
6342 attacker by sending a record for the next epoch (which does not have to
6343 decrypt or have a valid MAC), with a very large sequence number. This means
6344 that all subsequent legitimate packets are dropped causing a denial of
6345 service for a specific DTLS connection.
6347 This issue was reported to OpenSSL by the OCAP audit team.
6352 * Certificate message OOB reads
6354 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6355 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6356 theoretical DoS risk but this has not been observed in practice on common
6359 The messages affected are client certificate, client certificate request
6360 and server certificate. As a result the attack can only be performed
6361 against a client or a server which enables client authentication.
6363 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6368 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6370 * Prevent padding oracle in AES-NI CBC MAC check
6372 A MITM attacker can use a padding oracle attack to decrypt traffic
6373 when the connection uses an AES CBC cipher and the server support
6376 This issue was introduced as part of the fix for Lucky 13 padding
6377 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6378 constant time by making sure that always the same bytes are read and
6379 compared against either the MAC or padding bytes. But it no longer
6380 checked that there was enough data to have both the MAC and padding
6383 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6388 * Fix EVP_EncodeUpdate overflow
6390 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6391 Base64 encoding of binary data. If an attacker is able to supply very large
6392 amounts of input data then a length check can overflow resulting in a heap
6395 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6396 the `PEM_write_bio*` family of functions. These are mainly used within the
6397 OpenSSL command line applications, so any application which processes data
6398 from an untrusted source and outputs it as a PEM file should be considered
6399 vulnerable to this issue. User applications that call these APIs directly
6400 with large amounts of untrusted data may also be vulnerable.
6402 This issue was reported by Guido Vranken.
6407 * Fix EVP_EncryptUpdate overflow
6409 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6410 is able to supply very large amounts of input data after a previous call to
6411 EVP_EncryptUpdate() with a partial block then a length check can overflow
6412 resulting in a heap corruption. Following an analysis of all OpenSSL
6413 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6414 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6415 the first called function after an EVP_EncryptInit(), and therefore that
6416 specific call must be safe. The second form is where the length passed to
6417 EVP_EncryptUpdate() can be seen from the code to be some small value and
6418 therefore there is no possibility of an overflow. Since all instances are
6419 one of these two forms, it is believed that there can be no overflows in
6420 internal code due to this problem. It should be noted that
6421 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6422 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6423 of these calls have also been analysed too and it is believed there are no
6424 instances in internal usage where an overflow could occur.
6426 This issue was reported by Guido Vranken.
6431 * Prevent ASN.1 BIO excessive memory allocation
6433 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6434 a short invalid encoding can casuse allocation of large amounts of memory
6435 potentially consuming excessive resources or exhausting memory.
6437 Any application parsing untrusted data through d2i BIO functions is
6438 affected. The memory based functions such as d2i_X509() are *not* affected.
6439 Since the memory based functions are used by the TLS library, TLS
6440 applications are not affected.
6442 This issue was reported by Brian Carpenter.
6449 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6450 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6451 in arbitrary stack data being returned in the buffer.
6453 This issue was reported by Guido Vranken.
6458 * Modify behavior of ALPN to invoke callback after SNI/servername
6459 callback, such that updates to the SSL_CTX affect ALPN.
6463 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6468 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6469 methods are enabled and ssl2 is disabled the methods return NULL.
6473 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6475 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6476 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6477 provide any "EXPORT" or "LOW" strength ciphers.
6481 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6482 is by default disabled at build-time. Builds that are not configured with
6483 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6484 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6485 will need to explicitly call either of:
6487 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6489 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6491 as appropriate. Even if either of those is used, or the application
6492 explicitly uses the version-specific SSLv2_method() or its client and
6493 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6494 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6495 ciphers, and SSLv2 56-bit DES are no longer available.
6500 * Fix a double-free in DSA code
6502 A double free bug was discovered when OpenSSL parses malformed DSA private
6503 keys and could lead to a DoS attack or memory corruption for applications
6504 that receive DSA private keys from untrusted sources. This scenario is
6507 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6513 * Disable SRP fake user seed to address a server memory leak.
6515 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6517 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6518 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6519 was changed to ignore the "fake user" SRP seed, even if the seed
6522 Users should use SRP_VBASE_get1_by_user instead. Note that in
6523 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6524 also that even though configuring the SRP seed attempts to hide
6525 invalid usernames by continuing the handshake with fake
6526 credentials, this behaviour is not constant time and no strong
6527 guarantees are made that the handshake is indistinguishable from
6528 that of a valid user.
6533 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6535 In the BN_hex2bn function the number of hex digits is calculated using an
6536 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6537 large values of `i` this can result in `bn_expand` not allocating any
6538 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6539 field as NULL leading to a subsequent NULL ptr deref. For very large values
6540 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6541 In this case memory is allocated to the internal BIGNUM data field, but it
6542 is insufficiently sized leading to heap corruption. A similar issue exists
6543 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6544 is ever called by user applications with very large untrusted hex/dec data.
6545 This is anticipated to be a rare occurrence.
6547 All OpenSSL internal usage of these functions use data that is not expected
6548 to be untrusted, e.g. config file data or application command line
6549 arguments. If user developed applications generate config file data based
6550 on untrusted data then it is possible that this could also lead to security
6551 consequences. This is also anticipated to be rare.
6553 This issue was reported to OpenSSL by Guido Vranken.
6558 * Fix memory issues in `BIO_*printf` functions
6560 The internal `fmtstr` function used in processing a "%s" format string in
6561 the `BIO_*printf` functions could overflow while calculating the length of a
6562 string and cause an OOB read when printing very long strings.
6564 Additionally the internal `doapr_outch` function can attempt to write to an
6565 OOB memory location (at an offset from the NULL pointer) in the event of a
6566 memory allocation failure. In 1.0.2 and below this could be caused where
6567 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6568 could be in processing a very long "%s" format string. Memory leaks can
6571 The first issue may mask the second issue dependent on compiler behaviour.
6572 These problems could enable attacks where large amounts of untrusted data
6573 is passed to the `BIO_*printf` functions. If applications use these functions
6574 in this way then they could be vulnerable. OpenSSL itself uses these
6575 functions when printing out human-readable dumps of ASN.1 data. Therefore
6576 applications that print this data could be vulnerable if the data is from
6577 untrusted sources. OpenSSL command line applications could also be
6578 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6579 as command line arguments.
6581 Libssl is not considered directly vulnerable. Additionally certificates etc
6582 received via remote connections via libssl are also unlikely to be able to
6583 trigger these issues because of message size limits enforced within libssl.
6585 This issue was reported to OpenSSL Guido Vranken.
6590 * Side channel attack on modular exponentiation
6592 A side-channel attack was found which makes use of cache-bank conflicts on
6593 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6594 of RSA keys. The ability to exploit this issue is limited as it relies on
6595 an attacker who has control of code in a thread running on the same
6596 hyper-threaded core as the victim thread which is performing decryptions.
6598 This issue was reported to OpenSSL by Yuval Yarom, The University of
6599 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6600 Nadia Heninger, University of Pennsylvania with more information at
6601 <http://cachebleed.info>.
6606 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6607 if no keysize is specified with default_bits. This fixes an
6608 omission in an earlier change that changed all RSA/DSA key generation
6609 commands to use 2048 bits by default.
6613 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6615 * Protection for DH small subgroup attacks
6617 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6618 switched on by default and cannot be disabled. This could have some
6623 * SSLv2 doesn't block disabled ciphers
6625 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6626 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6627 been disabled, provided that the SSLv2 protocol was not also disabled via
6630 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6631 and Sebastian Schinzel.
6636 * Reject DH handshakes with parameters shorter than 1024 bits.
6640 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6642 * Certificate verify crash with missing PSS parameter
6644 The signature verification routines will crash with a NULL pointer
6645 dereference if presented with an ASN.1 signature using the RSA PSS
6646 algorithm and absent mask generation function parameter. Since these
6647 routines are used to verify certificate signature algorithms this can be
6648 used to crash any certificate verification operation and exploited in a
6649 DoS attack. Any application which performs certificate verification is
6650 vulnerable including OpenSSL clients and servers which enable client
6653 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6658 * X509_ATTRIBUTE memory leak
6660 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6661 memory. This structure is used by the PKCS#7 and CMS routines so any
6662 application which reads PKCS#7 or CMS data from untrusted sources is
6663 affected. SSL/TLS is not affected.
6665 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6671 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6672 This changes the decoding behaviour for some invalid messages,
6673 though the change is mostly in the more lenient direction, and
6674 legacy behaviour is preserved as much as possible.
6678 * In DSA_generate_parameters_ex, if the provided seed is too short,
6679 use a random seed, as already documented.
6681 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6683 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6685 * Alternate chains certificate forgery
6687 During certificate verfification, OpenSSL will attempt to find an
6688 alternative certificate chain if the first attempt to build such a chain
6689 fails. An error in the implementation of this logic can mean that an
6690 attacker could cause certain checks on untrusted certificates to be
6691 bypassed, such as the CA flag, enabling them to use a valid leaf
6692 certificate to act as a CA and "issue" an invalid certificate.
6694 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6700 * Race condition handling PSK identify hint
6702 If PSK identity hints are received by a multi-threaded client then
6703 the values are wrongly updated in the parent SSL_CTX structure. This can
6704 result in a race condition potentially leading to a double free of the
6710 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6712 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6713 incompatibility in the handling of HMAC. The previous ABI has now been
6716 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6718 * Malformed ECParameters causes infinite loop
6720 When processing an ECParameters structure OpenSSL enters an infinite loop
6721 if the curve specified is over a specially malformed binary polynomial
6724 This can be used to perform denial of service against any
6725 system which processes public keys, certificate requests or
6726 certificates. This includes TLS clients and TLS servers with
6727 client authentication enabled.
6729 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6734 * Exploitable out-of-bounds read in X509_cmp_time
6736 X509_cmp_time does not properly check the length of the ASN1_TIME
6737 string and can read a few bytes out of bounds. In addition,
6738 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6741 An attacker can use this to craft malformed certificates and CRLs of
6742 various sizes and potentially cause a segmentation fault, resulting in
6743 a DoS on applications that verify certificates or CRLs. TLS clients
6744 that verify CRLs are affected. TLS clients and servers with client
6745 authentication enabled may be affected if they use custom verification
6748 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6749 independently by Hanno Böck.
6754 * PKCS7 crash with missing EnvelopedContent
6756 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6757 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6758 with missing content and trigger a NULL pointer dereference on parsing.
6760 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6761 structures from untrusted sources are affected. OpenSSL clients and
6762 servers are not affected.
6764 This issue was reported to OpenSSL by Michal Zalewski (Google).
6769 * CMS verify infinite loop with unknown hash function
6771 When verifying a signedData message the CMS code can enter an infinite loop
6772 if presented with an unknown hash function OID. This can be used to perform
6773 denial of service against any system which verifies signedData messages using
6775 This issue was reported to OpenSSL by Johannes Bauer.
6780 * Race condition handling NewSessionTicket
6782 If a NewSessionTicket is received by a multi-threaded client when attempting to
6783 reuse a previous ticket then a race condition can occur potentially leading to
6784 a double free of the ticket data.
6789 * Reject DH handshakes with parameters shorter than 768 bits.
6791 *Kurt Roeckx and Emilia Kasper*
6793 * dhparam: generate 2048-bit parameters by default.
6795 *Kurt Roeckx and Emilia Kasper*
6797 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6799 * Segmentation fault in ASN1_TYPE_cmp fix
6801 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6802 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6803 certificate signature algorithm consistency this can be used to crash any
6804 certificate verification operation and exploited in a DoS attack. Any
6805 application which performs certificate verification is vulnerable including
6806 OpenSSL clients and servers which enable client authentication.
6811 * ASN.1 structure reuse memory corruption fix
6813 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6814 memory corruption via an invalid write. Such reuse is and has been
6815 strongly discouraged and is believed to be rare.
6817 Applications that parse structures containing CHOICE or ANY DEFINED BY
6818 components may be affected. Certificate parsing (d2i_X509 and related
6819 functions) are however not affected. OpenSSL clients and servers are
6825 * PKCS7 NULL pointer dereferences fix
6827 The PKCS#7 parsing code does not handle missing outer ContentInfo
6828 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6829 missing content and trigger a NULL pointer dereference on parsing.
6831 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6832 otherwise parse PKCS#7 structures from untrusted sources are
6833 affected. OpenSSL clients and servers are not affected.
6835 This issue was reported to OpenSSL by Michal Zalewski (Google).
6840 * DoS via reachable assert in SSLv2 servers fix
6842 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6843 servers that both support SSLv2 and enable export cipher suites by sending
6844 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6846 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6847 (OpenSSL development team).
6852 * Use After Free following d2i_ECPrivatekey error fix
6854 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6855 could cause a use after free condition. This, in turn, could cause a double
6856 free in several private key parsing functions (such as d2i_PrivateKey
6857 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6858 for applications that receive EC private keys from untrusted
6859 sources. This scenario is considered rare.
6861 This issue was discovered by the BoringSSL project and fixed in their
6867 * X509_to_X509_REQ NULL pointer deref fix
6869 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6870 the certificate key is invalid. This function is rarely used in practice.
6872 This issue was discovered by Brian Carpenter.
6877 * Removed the export ciphers from the DEFAULT ciphers
6881 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6883 * Build fixes for the Windows and OpenVMS platforms
6885 *Matt Caswell and Richard Levitte*
6887 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6889 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6890 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6891 dereference. This could lead to a Denial Of Service attack. Thanks to
6892 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6897 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6898 dtls1_buffer_record function under certain conditions. In particular this
6899 could occur if an attacker sent repeated DTLS records with the same
6900 sequence number but for the next epoch. The memory leak could be exploited
6901 by an attacker in a Denial of Service attack through memory exhaustion.
6902 Thanks to Chris Mueller for reporting this issue.
6907 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6908 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6909 method would be set to NULL which could later result in a NULL pointer
6910 dereference. Thanks to Frank Schmirler for reporting this issue.
6915 * Abort handshake if server key exchange message is omitted for ephemeral
6918 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6919 reporting this issue.
6924 * Remove non-export ephemeral RSA code on client and server. This code
6925 violated the TLS standard by allowing the use of temporary RSA keys in
6926 non-export ciphersuites and could be used by a server to effectively
6927 downgrade the RSA key length used to a value smaller than the server
6928 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6929 INRIA or reporting this issue.
6934 * Fixed issue where DH client certificates are accepted without verification.
6935 An OpenSSL server will accept a DH certificate for client authentication
6936 without the certificate verify message. This effectively allows a client to
6937 authenticate without the use of a private key. This only affects servers
6938 which trust a client certificate authority which issues certificates
6939 containing DH keys: these are extremely rare and hardly ever encountered.
6940 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6946 * Ensure that the session ID context of an SSL is updated when its
6947 SSL_CTX is updated via SSL_set_SSL_CTX.
6949 The session ID context is typically set from the parent SSL_CTX,
6950 and can vary with the CTX.
6954 * Fix various certificate fingerprint issues.
6956 By using non-DER or invalid encodings outside the signed portion of a
6957 certificate the fingerprint can be changed without breaking the signature.
6958 Although no details of the signed portion of the certificate can be changed
6959 this can cause problems with some applications: e.g. those using the
6960 certificate fingerprint for blacklists.
6962 1. Reject signatures with non zero unused bits.
6964 If the BIT STRING containing the signature has non zero unused bits reject
6965 the signature. All current signature algorithms require zero unused bits.
6967 2. Check certificate algorithm consistency.
6969 Check the AlgorithmIdentifier inside TBS matches the one in the
6970 certificate signature. NB: this will result in signature failure
6971 errors for some broken certificates.
6973 Thanks to Konrad Kraszewski from Google for reporting this issue.
6975 3. Check DSA/ECDSA signatures use DER.
6977 Re-encode DSA/ECDSA signatures and compare with the original received
6978 signature. Return an error if there is a mismatch.
6980 This will reject various cases including garbage after signature
6981 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6982 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6983 (negative or with leading zeroes).
6985 Further analysis was conducted and fixes were developed by Stephen Henson
6986 of the OpenSSL core team.
6992 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6993 results on some platforms, including x86_64. This bug occurs at random
6994 with a very low probability, and is not known to be exploitable in any
6995 way, though its exact impact is difficult to determine. Thanks to Pieter
6996 Wuille (Blockstream) who reported this issue and also suggested an initial
6997 fix. Further analysis was conducted by the OpenSSL development team and
6998 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6999 the OpenSSL core team.
7004 * Do not resume sessions on the server if the negotiated protocol
7005 version does not match the session's version. Resuming with a different
7006 version, while not strictly forbidden by the RFC, is of questionable
7007 sanity and breaks all known clients.
7009 *David Benjamin, Emilia Käsper*
7011 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7012 early CCS messages during renegotiation. (Note that because
7013 renegotiation is encrypted, this early CCS was not exploitable.)
7017 * Tighten client-side session ticket handling during renegotiation:
7018 ensure that the client only accepts a session ticket if the server sends
7019 the extension anew in the ServerHello. Previously, a TLS client would
7020 reuse the old extension state and thus accept a session ticket if one was
7021 announced in the initial ServerHello.
7023 Similarly, ensure that the client requires a session ticket if one
7024 was advertised in the ServerHello. Previously, a TLS client would
7025 ignore a missing NewSessionTicket message.
7029 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7033 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7034 sends a carefully crafted handshake message, to cause OpenSSL to fail
7035 to free up to 64k of memory causing a memory leak. This could be
7036 exploited in a Denial Of Service attack. This issue affects OpenSSL
7037 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7038 whether SRTP is used or configured. Implementations of OpenSSL that
7039 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7041 The fix was developed by the OpenSSL team.
7046 * Session Ticket Memory Leak.
7048 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7049 integrity of that ticket is first verified. In the event of a session
7050 ticket integrity check failing, OpenSSL will fail to free memory
7051 causing a memory leak. By sending a large number of invalid session
7052 tickets an attacker could exploit this issue in a Denial Of Service
7058 * Build option no-ssl3 is incomplete.
7060 When OpenSSL is configured with "no-ssl3" as a build option, servers
7061 could accept and complete a SSL 3.0 handshake, and clients could be
7062 configured to send them.
7065 *Akamai and the OpenSSL team*
7067 * Add support for TLS_FALLBACK_SCSV.
7068 Client applications doing fallback retries should call
7069 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7072 *Adam Langley, Bodo Moeller*
7074 * Add additional DigestInfo checks.
7076 Re-encode DigestInto in DER and check against the original when
7077 verifying RSA signature: this will reject any improperly encoded
7078 DigestInfo structures.
7080 Note: this is a precautionary measure and no attacks are currently known.
7084 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7086 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7087 SRP code can be overrun an internal buffer. Add sanity check that
7088 g, A, B < N to SRP code.
7090 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7091 Group for discovering this issue.
7096 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7097 TLS 1.0 instead of higher protocol versions when the ClientHello message
7098 is badly fragmented. This allows a man-in-the-middle attacker to force a
7099 downgrade to TLS 1.0 even if both the server and the client support a
7100 higher protocol version, by modifying the client's TLS records.
7102 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7103 researching this issue.
7108 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7109 to a denial of service attack. A malicious server can crash the client
7110 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7111 ciphersuite and sending carefully crafted handshake messages.
7113 Thanks to Felix Gröbert (Google) for discovering and researching this
7119 * By sending carefully crafted DTLS packets an attacker could cause openssl
7120 to leak memory. This can be exploited through a Denial of Service attack.
7121 Thanks to Adam Langley for discovering and researching this issue.
7126 * An attacker can force openssl to consume large amounts of memory whilst
7127 processing DTLS handshake messages. This can be exploited through a
7128 Denial of Service attack.
7129 Thanks to Adam Langley for discovering and researching this issue.
7134 * An attacker can force an error condition which causes openssl to crash
7135 whilst processing DTLS packets due to memory being freed twice. This
7136 can be exploited through a Denial of Service attack.
7137 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7143 * If a multithreaded client connects to a malicious server using a resumed
7144 session and the server sends an ec point format extension it could write
7145 up to 255 bytes to freed memory.
7147 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7153 * A malicious server can crash an OpenSSL client with a null pointer
7154 dereference (read) by specifying an SRP ciphersuite even though it was not
7155 properly negotiated with the client. This can be exploited through a
7156 Denial of Service attack.
7158 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7159 discovering and researching this issue.
7164 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7165 X509_name_oneline, X509_name_print_ex et al. to leak some information
7166 from the stack. Applications may be affected if they echo pretty printing
7167 output to the attacker.
7169 Thanks to Ivan Fratric (Google) for discovering this issue.
7172 *Emilia Käsper, and Steve Henson*
7174 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7175 for corner cases. (Certain input points at infinity could lead to
7176 bogus results, with non-infinity inputs mapped to infinity too.)
7180 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7182 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7183 handshake can force the use of weak keying material in OpenSSL
7184 SSL/TLS clients and servers.
7186 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7187 researching this issue. ([CVE-2014-0224])
7189 *KIKUCHI Masashi, Steve Henson*
7191 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7192 OpenSSL DTLS client the code can be made to recurse eventually crashing
7195 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7198 *Imre Rad, Steve Henson*
7200 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7201 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7202 client or server. This is potentially exploitable to run arbitrary
7203 code on a vulnerable client or server.
7205 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7207 *Jüri Aedla, Steve Henson*
7209 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7210 are subject to a denial of service attack.
7212 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7213 this issue. ([CVE-2014-3470])
7215 *Felix Gröbert, Ivan Fratric, Steve Henson*
7217 * Harmonize version and its documentation. -f flag is used to display
7220 *mancha <mancha1@zoho.com>*
7222 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7223 in i2d_ECPrivateKey.
7225 *mancha <mancha1@zoho.com>*
7227 * Fix some double frees. These are not thought to be exploitable.
7229 *mancha <mancha1@zoho.com>*
7231 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7233 * A missing bounds check in the handling of the TLS heartbeat extension
7234 can be used to reveal up to 64k of memory to a connected client or
7237 Thanks for Neel Mehta of Google Security for discovering this bug and to
7238 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7239 preparing the fix ([CVE-2014-0160])
7241 *Adam Langley, Bodo Moeller*
7243 * Fix for the attack described in the paper "Recovering OpenSSL
7244 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7245 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7246 <http://eprint.iacr.org/2014/140>
7248 Thanks to Yuval Yarom and Naomi Benger for discovering this
7249 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7251 *Yuval Yarom and Naomi Benger*
7253 * TLS pad extension: draft-agl-tls-padding-03
7255 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7256 TLS client Hello record length value would otherwise be > 255 and
7257 less that 512 pad with a dummy extension containing zeroes so it
7258 is at least 512 bytes long.
7260 *Adam Langley, Steve Henson*
7262 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7264 * Fix for TLS record tampering bug. A carefully crafted invalid
7265 handshake could crash OpenSSL with a NULL pointer exception.
7266 Thanks to Anton Johansson for reporting this issues.
7269 * Keep original DTLS digest and encryption contexts in retransmission
7270 structures so we can use the previous session parameters if they need
7271 to be resent. ([CVE-2013-6450])
7275 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7276 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7277 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7278 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7279 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7280 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7282 *Rob Stradling, Adam Langley*
7284 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7286 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7287 supporting platforms or when small records were transferred.
7289 *Andy Polyakov, Steve Henson*
7291 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7293 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7295 This addresses the flaw in CBC record processing discovered by
7296 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7297 at: <http://www.isg.rhul.ac.uk/tls/>
7299 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7300 Security Group at Royal Holloway, University of London
7301 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7302 Emilia Käsper for the initial patch.
7305 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7307 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7308 ciphersuites which can be exploited in a denial of service attack.
7309 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7310 and detecting this bug and to Wolfgang Ettlinger
7311 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7316 * Return an error when checking OCSP signatures when key is NULL.
7317 This fixes a DoS attack. ([CVE-2013-0166])
7321 * Make openssl verify return errors.
7323 *Chris Palmer <palmer@google.com> and Ben Laurie*
7325 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7326 the right response is stapled. Also change SSL_get_certificate()
7327 so it returns the certificate actually sent.
7328 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7330 *Rob Stradling <rob.stradling@comodo.com>*
7332 * Fix possible deadlock when decoding public keys.
7336 * Don't use TLS 1.0 record version number in initial client hello
7341 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7343 * Sanity check record length before skipping explicit IV in TLS
7344 1.2, 1.1 and DTLS to fix DoS attack.
7346 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7347 fuzzing as a service testing platform.
7352 * Initialise tkeylen properly when encrypting CMS messages.
7353 Thanks to Solar Designer of Openwall for reporting this issue.
7357 * In FIPS mode don't try to use composite ciphers as they are not
7362 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7364 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7365 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7366 mean any application compiled against OpenSSL 1.0.0 headers setting
7367 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7368 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7369 0x10000000L Any application which was previously compiled against
7370 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7371 will need to be recompiled as a result. Letting be results in
7372 inability to disable specifically TLS 1.1 and in client context,
7373 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7377 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7378 disable just protocol X, but all protocols above X *if* there are
7379 protocols *below* X still enabled. In more practical terms it means
7380 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7381 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7382 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7387 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7389 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7390 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7391 in CRYPTO_realloc_clean.
7393 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7394 issue and to Adam Langley <agl@chromium.org> for fixing it.
7397 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7399 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7403 * Workarounds for some broken servers that "hang" if a client hello
7404 record length exceeds 255 bytes.
7406 1. Do not use record version number > TLS 1.0 in initial client
7407 hello: some (but not all) hanging servers will now work.
7408 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7409 the number of ciphers sent in the client hello. This should be
7410 set to an even number, such as 50, for example by passing:
7411 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7412 Most broken servers should now work.
7413 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7414 TLS 1.2 client support entirely.
7418 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7422 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7424 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7425 STRING form instead of a DigestInfo.
7429 * The format used for MDC2 RSA signatures is inconsistent between EVP
7430 and the RSA_sign/RSA_verify functions. This was made more apparent when
7431 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7432 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7433 the correct format in RSA_verify so both forms transparently work.
7437 * Some servers which support TLS 1.0 can choke if we initially indicate
7438 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7439 encrypted premaster secret. As a workaround use the maximum permitted
7440 client version in client hello, this should keep such servers happy
7441 and still work with previous versions of OpenSSL.
7445 * Add support for TLS/DTLS heartbeats.
7447 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7449 * Add support for SCTP.
7451 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7453 * Improved PRNG seeding for VOS.
7455 *Paul Green <Paul.Green@stratus.com>*
7457 * Extensive assembler packs updates, most notably:
7459 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7460 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7461 - x86_64: bit-sliced AES implementation;
7462 - ARM: NEON support, contemporary platforms optimizations;
7463 - s390x: z196 support;
7464 - `*`: GHASH and GF(2^m) multiplication implementations;
7468 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7469 (removal of unnecessary code)
7471 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7473 * Add TLS key material exporter from RFC 5705.
7477 * Add DTLS-SRTP negotiation from RFC 5764.
7481 * Add Next Protocol Negotiation,
7482 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7483 disabled with a no-npn flag to config or Configure. Code donated
7486 *Adam Langley <agl@google.com> and Ben Laurie*
7488 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7489 NIST-P256, NIST-P521, with constant-time single point multiplication on
7490 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7491 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7492 Code made available under Apache License version 2.0.
7494 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7495 line to include this in your build of OpenSSL, and run "make depend" (or
7496 "make update"). This enables the following EC_METHODs:
7498 EC_GFp_nistp224_method()
7499 EC_GFp_nistp256_method()
7500 EC_GFp_nistp521_method()
7502 EC_GROUP_new_by_curve_name() will automatically use these (while
7503 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7506 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7508 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7509 all platforms. Move ssize_t definition from e_os.h to the public
7510 header file e_os2.h as it now appears in public header file cms.h
7514 * New -sigopt option to the ca, req and x509 utilities. Additional
7515 signature parameters can be passed using this option and in
7520 * Add RSA PSS signing function. This will generate and set the
7521 appropriate AlgorithmIdentifiers for PSS based on those in the
7522 corresponding EVP_MD_CTX structure. No application support yet.
7526 * Support for companion algorithm specific ASN1 signing routines.
7527 New function ASN1_item_sign_ctx() signs a pre-initialised
7528 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7529 the appropriate parameters.
7533 * Add new algorithm specific ASN1 verification initialisation function
7534 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7535 handling will be the same no matter what EVP_PKEY_METHOD is used.
7536 Add a PSS handler to support verification of PSS signatures: checked
7537 against a number of sample certificates.
7541 * Add signature printing for PSS. Add PSS OIDs.
7543 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7545 * Add algorithm specific signature printing. An individual ASN1 method
7546 can now print out signatures instead of the standard hex dump.
7548 More complex signatures (e.g. PSS) can print out more meaningful
7549 information. Include DSA version that prints out the signature
7554 * Password based recipient info support for CMS library: implementing
7559 * Split password based encryption into PBES2 and PBKDF2 functions. This
7560 neatly separates the code into cipher and PBE sections and is required
7561 for some algorithms that split PBES2 into separate pieces (such as
7562 password based CMS).
7566 * Session-handling fixes:
7567 - Fix handling of connections that are resuming with a session ID,
7568 but also support Session Tickets.
7569 - Fix a bug that suppressed issuing of a new ticket if the client
7570 presented a ticket with an expired session.
7571 - Try to set the ticket lifetime hint to something reasonable.
7572 - Make tickets shorter by excluding irrelevant information.
7573 - On the client side, don't ignore renewed tickets.
7575 *Adam Langley, Bodo Moeller (Google)*
7577 * Fix PSK session representation.
7581 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7583 This work was sponsored by Intel.
7587 * Add GCM support to TLS library. Some custom code is needed to split
7588 the IV between the fixed (from PRF) and explicit (from TLS record)
7589 portions. This adds all GCM ciphersuites supported by RFC5288 and
7590 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7591 add a special AESGCM string for GCM only.
7595 * Expand range of ctrls for AES GCM. Permit setting invocation
7596 field on decrypt and retrieval of invocation field only on encrypt.
7600 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7601 As required by RFC5289 these ciphersuites cannot be used if for
7602 versions of TLS earlier than 1.2.
7606 * For FIPS capable OpenSSL interpret a NULL default public key method
7607 as unset and return the appropriate default but do *not* set the default.
7608 This means we can return the appropriate method in applications that
7609 switch between FIPS and non-FIPS modes.
7613 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7614 ENGINE is used then we cannot handle that in the FIPS module so we
7615 keep original code iff non-FIPS operations are allowed.
7619 * Add -attime option to openssl utilities.
7621 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7623 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7627 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7628 FIPS EC methods unconditionally for now.
7632 * New build option no-ec2m to disable characteristic 2 code.
7636 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7637 all cases can be covered as some introduce binary incompatibilities.
7641 * Redirect RSA operations to FIPS module including keygen,
7642 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7646 * Add similar low level API blocking to ciphers.
7650 * Low level digest APIs are not approved in FIPS mode: any attempt
7651 to use these will cause a fatal error. Applications that *really* want
7652 to use them can use the `private_*` version instead.
7656 * Redirect cipher operations to FIPS module for FIPS builds.
7660 * Redirect digest operations to FIPS module for FIPS builds.
7664 * Update build system to add "fips" flag which will link in fipscanister.o
7665 for static and shared library builds embedding a signature if needed.
7669 * Output TLS supported curves in preference order instead of numerical
7670 order. This is currently hardcoded for the highest order curves first.
7671 This should be configurable so applications can judge speed vs strength.
7675 * Add TLS v1.2 server support for client authentication.
7679 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7684 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7685 FIPS modules versions.
7689 * Add TLS v1.2 client side support for client authentication. Keep cache
7690 of handshake records longer as we don't know the hash algorithm to use
7691 until after the certificate request message is received.
7695 * Initial TLS v1.2 client support. Add a default signature algorithms
7696 extension including all the algorithms we support. Parse new signature
7697 format in client key exchange. Relax some ECC signing restrictions for
7698 TLS v1.2 as indicated in RFC5246.
7702 * Add server support for TLS v1.2 signature algorithms extension. Switch
7703 to new signature format when needed using client digest preference.
7704 All server ciphersuites should now work correctly in TLS v1.2. No client
7705 support yet and no support for client certificates.
7709 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7710 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7711 ciphersuites. At present only RSA key exchange ciphersuites work with
7712 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7713 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7714 and version checking.
7718 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7719 with this defined it will not be affected by any changes to ssl internal
7720 structures. Add several utility functions to allow openssl application
7721 to work with OPENSSL_NO_SSL_INTERN defined.
7725 * A long standing patch to add support for SRP from EdelWeb (Peter
7726 Sylvester and Christophe Renou) was integrated.
7727 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7728 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7731 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7735 * Permit abbreviated handshakes when renegotiating using the function
7736 SSL_renegotiate_abbreviated().
7738 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7740 * Add call to ENGINE_register_all_complete() to
7741 ENGINE_load_builtin_engines(), so some implementations get used
7742 automatically instead of needing explicit application support.
7746 * Add support for TLS key exporter as described in RFC5705.
7748 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7750 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7751 a few changes are required:
7753 Add SSL_OP_NO_TLSv1_1 flag.
7754 Add TLSv1_1 methods.
7755 Update version checking logic to handle version 1.1.
7756 Add explicit IV handling (ported from DTLS code).
7757 Add command line options to s_client/s_server.
7764 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7766 * X509_ATTRIBUTE memory leak
7768 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7769 memory. This structure is used by the PKCS#7 and CMS routines so any
7770 application which reads PKCS#7 or CMS data from untrusted sources is
7771 affected. SSL/TLS is not affected.
7773 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7779 * Race condition handling PSK identify hint
7781 If PSK identity hints are received by a multi-threaded client then
7782 the values are wrongly updated in the parent SSL_CTX structure. This can
7783 result in a race condition potentially leading to a double free of the
7789 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7791 * Malformed ECParameters causes infinite loop
7793 When processing an ECParameters structure OpenSSL enters an infinite loop
7794 if the curve specified is over a specially malformed binary polynomial
7797 This can be used to perform denial of service against any
7798 system which processes public keys, certificate requests or
7799 certificates. This includes TLS clients and TLS servers with
7800 client authentication enabled.
7802 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7807 * Exploitable out-of-bounds read in X509_cmp_time
7809 X509_cmp_time does not properly check the length of the ASN1_TIME
7810 string and can read a few bytes out of bounds. In addition,
7811 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7814 An attacker can use this to craft malformed certificates and CRLs of
7815 various sizes and potentially cause a segmentation fault, resulting in
7816 a DoS on applications that verify certificates or CRLs. TLS clients
7817 that verify CRLs are affected. TLS clients and servers with client
7818 authentication enabled may be affected if they use custom verification
7821 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7822 independently by Hanno Böck.
7827 * PKCS7 crash with missing EnvelopedContent
7829 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7830 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7831 with missing content and trigger a NULL pointer dereference on parsing.
7833 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7834 structures from untrusted sources are affected. OpenSSL clients and
7835 servers are not affected.
7837 This issue was reported to OpenSSL by Michal Zalewski (Google).
7842 * CMS verify infinite loop with unknown hash function
7844 When verifying a signedData message the CMS code can enter an infinite loop
7845 if presented with an unknown hash function OID. This can be used to perform
7846 denial of service against any system which verifies signedData messages using
7848 This issue was reported to OpenSSL by Johannes Bauer.
7853 * Race condition handling NewSessionTicket
7855 If a NewSessionTicket is received by a multi-threaded client when attempting to
7856 reuse a previous ticket then a race condition can occur potentially leading to
7857 a double free of the ticket data.
7862 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7864 * Segmentation fault in ASN1_TYPE_cmp fix
7866 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7867 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7868 certificate signature algorithm consistency this can be used to crash any
7869 certificate verification operation and exploited in a DoS attack. Any
7870 application which performs certificate verification is vulnerable including
7871 OpenSSL clients and servers which enable client authentication.
7876 * ASN.1 structure reuse memory corruption fix
7878 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7879 memory corruption via an invalid write. Such reuse is and has been
7880 strongly discouraged and is believed to be rare.
7882 Applications that parse structures containing CHOICE or ANY DEFINED BY
7883 components may be affected. Certificate parsing (d2i_X509 and related
7884 functions) are however not affected. OpenSSL clients and servers are
7890 * PKCS7 NULL pointer dereferences fix
7892 The PKCS#7 parsing code does not handle missing outer ContentInfo
7893 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7894 missing content and trigger a NULL pointer dereference on parsing.
7896 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7897 otherwise parse PKCS#7 structures from untrusted sources are
7898 affected. OpenSSL clients and servers are not affected.
7900 This issue was reported to OpenSSL by Michal Zalewski (Google).
7905 * DoS via reachable assert in SSLv2 servers fix
7907 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7908 servers that both support SSLv2 and enable export cipher suites by sending
7909 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7911 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7912 (OpenSSL development team).
7917 * Use After Free following d2i_ECPrivatekey error fix
7919 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7920 could cause a use after free condition. This, in turn, could cause a double
7921 free in several private key parsing functions (such as d2i_PrivateKey
7922 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7923 for applications that receive EC private keys from untrusted
7924 sources. This scenario is considered rare.
7926 This issue was discovered by the BoringSSL project and fixed in their
7932 * X509_to_X509_REQ NULL pointer deref fix
7934 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7935 the certificate key is invalid. This function is rarely used in practice.
7937 This issue was discovered by Brian Carpenter.
7942 * Removed the export ciphers from the DEFAULT ciphers
7946 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7948 * Build fixes for the Windows and OpenVMS platforms
7950 *Matt Caswell and Richard Levitte*
7952 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7954 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7955 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7956 dereference. This could lead to a Denial Of Service attack. Thanks to
7957 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7962 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7963 dtls1_buffer_record function under certain conditions. In particular this
7964 could occur if an attacker sent repeated DTLS records with the same
7965 sequence number but for the next epoch. The memory leak could be exploited
7966 by an attacker in a Denial of Service attack through memory exhaustion.
7967 Thanks to Chris Mueller for reporting this issue.
7972 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7973 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7974 method would be set to NULL which could later result in a NULL pointer
7975 dereference. Thanks to Frank Schmirler for reporting this issue.
7980 * Abort handshake if server key exchange message is omitted for ephemeral
7983 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7984 reporting this issue.
7989 * Remove non-export ephemeral RSA code on client and server. This code
7990 violated the TLS standard by allowing the use of temporary RSA keys in
7991 non-export ciphersuites and could be used by a server to effectively
7992 downgrade the RSA key length used to a value smaller than the server
7993 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7994 INRIA or reporting this issue.
7999 * Fixed issue where DH client certificates are accepted without verification.
8000 An OpenSSL server will accept a DH certificate for client authentication
8001 without the certificate verify message. This effectively allows a client to
8002 authenticate without the use of a private key. This only affects servers
8003 which trust a client certificate authority which issues certificates
8004 containing DH keys: these are extremely rare and hardly ever encountered.
8005 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8011 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8012 results on some platforms, including x86_64. This bug occurs at random
8013 with a very low probability, and is not known to be exploitable in any
8014 way, though its exact impact is difficult to determine. Thanks to Pieter
8015 Wuille (Blockstream) who reported this issue and also suggested an initial
8016 fix. Further analysis was conducted by the OpenSSL development team and
8017 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8018 the OpenSSL core team.
8023 * Fix various certificate fingerprint issues.
8025 By using non-DER or invalid encodings outside the signed portion of a
8026 certificate the fingerprint can be changed without breaking the signature.
8027 Although no details of the signed portion of the certificate can be changed
8028 this can cause problems with some applications: e.g. those using the
8029 certificate fingerprint for blacklists.
8031 1. Reject signatures with non zero unused bits.
8033 If the BIT STRING containing the signature has non zero unused bits reject
8034 the signature. All current signature algorithms require zero unused bits.
8036 2. Check certificate algorithm consistency.
8038 Check the AlgorithmIdentifier inside TBS matches the one in the
8039 certificate signature. NB: this will result in signature failure
8040 errors for some broken certificates.
8042 Thanks to Konrad Kraszewski from Google for reporting this issue.
8044 3. Check DSA/ECDSA signatures use DER.
8046 Reencode DSA/ECDSA signatures and compare with the original received
8047 signature. Return an error if there is a mismatch.
8049 This will reject various cases including garbage after signature
8050 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8051 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8052 (negative or with leading zeroes).
8054 Further analysis was conducted and fixes were developed by Stephen Henson
8055 of the OpenSSL core team.
8061 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8063 * Session Ticket Memory Leak.
8065 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8066 integrity of that ticket is first verified. In the event of a session
8067 ticket integrity check failing, OpenSSL will fail to free memory
8068 causing a memory leak. By sending a large number of invalid session
8069 tickets an attacker could exploit this issue in a Denial Of Service
8075 * Build option no-ssl3 is incomplete.
8077 When OpenSSL is configured with "no-ssl3" as a build option, servers
8078 could accept and complete a SSL 3.0 handshake, and clients could be
8079 configured to send them.
8082 *Akamai and the OpenSSL team*
8084 * Add support for TLS_FALLBACK_SCSV.
8085 Client applications doing fallback retries should call
8086 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8089 *Adam Langley, Bodo Moeller*
8091 * Add additional DigestInfo checks.
8093 Reencode DigestInto in DER and check against the original when
8094 verifying RSA signature: this will reject any improperly encoded
8095 DigestInfo structures.
8097 Note: this is a precautionary measure and no attacks are currently known.
8101 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8103 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8104 to a denial of service attack. A malicious server can crash the client
8105 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8106 ciphersuite and sending carefully crafted handshake messages.
8108 Thanks to Felix Gröbert (Google) for discovering and researching this
8114 * By sending carefully crafted DTLS packets an attacker could cause openssl
8115 to leak memory. This can be exploited through a Denial of Service attack.
8116 Thanks to Adam Langley for discovering and researching this issue.
8121 * An attacker can force openssl to consume large amounts of memory whilst
8122 processing DTLS handshake messages. This can be exploited through a
8123 Denial of Service attack.
8124 Thanks to Adam Langley for discovering and researching this issue.
8129 * An attacker can force an error condition which causes openssl to crash
8130 whilst processing DTLS packets due to memory being freed twice. This
8131 can be exploited through a Denial of Service attack.
8132 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8138 * If a multithreaded client connects to a malicious server using a resumed
8139 session and the server sends an ec point format extension it could write
8140 up to 255 bytes to freed memory.
8142 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8148 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8149 X509_name_oneline, X509_name_print_ex et al. to leak some information
8150 from the stack. Applications may be affected if they echo pretty printing
8151 output to the attacker.
8153 Thanks to Ivan Fratric (Google) for discovering this issue.
8156 *Emilia Käsper, and Steve Henson*
8158 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8159 for corner cases. (Certain input points at infinity could lead to
8160 bogus results, with non-infinity inputs mapped to infinity too.)
8164 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8166 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8167 handshake can force the use of weak keying material in OpenSSL
8168 SSL/TLS clients and servers.
8170 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8171 researching this issue. ([CVE-2014-0224])
8173 *KIKUCHI Masashi, Steve Henson*
8175 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8176 OpenSSL DTLS client the code can be made to recurse eventually crashing
8179 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8182 *Imre Rad, Steve Henson*
8184 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8185 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8186 client or server. This is potentially exploitable to run arbitrary
8187 code on a vulnerable client or server.
8189 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8191 *Jüri Aedla, Steve Henson*
8193 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8194 are subject to a denial of service attack.
8196 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8197 this issue. ([CVE-2014-3470])
8199 *Felix Gröbert, Ivan Fratric, Steve Henson*
8201 * Harmonize version and its documentation. -f flag is used to display
8204 *mancha <mancha1@zoho.com>*
8206 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8207 in i2d_ECPrivateKey.
8209 *mancha <mancha1@zoho.com>*
8211 * Fix some double frees. These are not thought to be exploitable.
8213 *mancha <mancha1@zoho.com>*
8215 * Fix for the attack described in the paper "Recovering OpenSSL
8216 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8217 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8218 <http://eprint.iacr.org/2014/140>
8220 Thanks to Yuval Yarom and Naomi Benger for discovering this
8221 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8223 *Yuval Yarom and Naomi Benger*
8225 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8227 * Keep original DTLS digest and encryption contexts in retransmission
8228 structures so we can use the previous session parameters if they need
8229 to be resent. ([CVE-2013-6450])
8233 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8234 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8235 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8236 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8237 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8238 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8240 *Rob Stradling, Adam Langley*
8242 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8244 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8246 This addresses the flaw in CBC record processing discovered by
8247 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8248 at: <http://www.isg.rhul.ac.uk/tls/>
8250 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8251 Security Group at Royal Holloway, University of London
8252 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8253 Emilia Käsper for the initial patch.
8256 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8258 * Return an error when checking OCSP signatures when key is NULL.
8259 This fixes a DoS attack. ([CVE-2013-0166])
8263 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8264 the right response is stapled. Also change SSL_get_certificate()
8265 so it returns the certificate actually sent.
8266 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8267 (This is a backport)
8269 *Rob Stradling <rob.stradling@comodo.com>*
8271 * Fix possible deadlock when decoding public keys.
8275 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8277 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8280 * Sanity check record length before skipping explicit IV in DTLS
8283 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8284 fuzzing as a service testing platform.
8289 * Initialise tkeylen properly when encrypting CMS messages.
8290 Thanks to Solar Designer of Openwall for reporting this issue.
8294 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8296 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8297 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8298 in CRYPTO_realloc_clean.
8300 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8301 issue and to Adam Langley <agl@chromium.org> for fixing it.
8304 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8306 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8308 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8309 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8310 content decryption and always return the same error. Note: this attack
8311 needs on average 2^20 messages so it only affects automated senders. The
8312 old behaviour can be re-enabled in the CMS code by setting the
8313 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8314 an MMA defence is not necessary.
8315 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8316 this issue. ([CVE-2012-0884])
8320 * Fix CVE-2011-4619: make sure we really are receiving a
8321 client hello before rejecting multiple SGC restarts. Thanks to
8322 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8326 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8328 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8329 Thanks to Antonio Martin, Enterprise Secure Access Research and
8330 Development, Cisco Systems, Inc. for discovering this bug and
8331 preparing a fix. ([CVE-2012-0050])
8335 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8337 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8338 of the Vaudenay padding oracle attack on CBC mode encryption
8339 which enables an efficient plaintext recovery attack against
8340 the OpenSSL implementation of DTLS. Their attack exploits timing
8341 differences arising during decryption processing. A research
8342 paper describing this attack can be found at:
8343 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8344 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8345 Security Group at Royal Holloway, University of London
8346 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8347 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8348 for preparing the fix. ([CVE-2011-4108])
8350 *Robin Seggelmann, Michael Tuexen*
8352 * Clear bytes used for block padding of SSL 3.0 records.
8355 *Adam Langley (Google)*
8357 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8358 Kadianakis <desnacked@gmail.com> for discovering this issue and
8359 Adam Langley for preparing the fix. ([CVE-2011-4619])
8361 *Adam Langley (Google)*
8363 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8365 *Andrey Kulikov <amdeich@gmail.com>*
8367 * Prevent malformed RFC3779 data triggering an assertion failure.
8368 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8369 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8371 *Rob Austein <sra@hactrn.net>*
8373 * Improved PRNG seeding for VOS.
8375 *Paul Green <Paul.Green@stratus.com>*
8377 * Fix ssl_ciph.c set-up race.
8379 *Adam Langley (Google)*
8381 * Fix spurious failures in ecdsatest.c.
8383 *Emilia Käsper (Google)*
8385 * Fix the BIO_f_buffer() implementation (which was mixing different
8386 interpretations of the `..._len` fields).
8388 *Adam Langley (Google)*
8390 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8391 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8392 threads won't reuse the same blinding coefficients.
8394 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8395 lock to call BN_BLINDING_invert_ex, and avoids one use of
8396 BN_BLINDING_update for each BN_BLINDING structure (previously,
8397 the last update always remained unused).
8399 *Emilia Käsper (Google)*
8401 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8403 *Bob Buckholz (Google)*
8405 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8407 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8408 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8410 *Kaspar Brand <ossl@velox.ch>*
8412 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8413 for multi-threaded use of ECDH. ([CVE-2011-3210])
8415 *Adam Langley (Google)*
8417 * Fix x509_name_ex_d2i memory leak on bad inputs.
8421 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8422 signature public key algorithm by using OID xref utilities instead.
8423 Before this you could only use some ECC ciphersuites with SHA1 only.
8427 * Add protection against ECDSA timing attacks as mentioned in the paper
8428 by Billy Bob Brumley and Nicola Tuveri, see:
8429 <http://eprint.iacr.org/2011/232.pdf>
8431 *Billy Bob Brumley and Nicola Tuveri*
8433 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8435 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8437 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8439 * Fix bug in string printing code: if *any* escaping is enabled we must
8440 escape the escape character (backslash) or the resulting string is
8445 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8447 * Disable code workaround for ancient and obsolete Netscape browsers
8448 and servers: an attacker can use it in a ciphersuite downgrade attack.
8449 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8453 * Fixed J-PAKE implementation error, originally discovered by
8454 Sebastien Martini, further info and confirmation from Stefan
8455 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8459 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8461 * Fix extension code to avoid race conditions which can result in a buffer
8462 overrun vulnerability: resumed sessions must not be modified as they can
8463 be shared by multiple threads. CVE-2010-3864
8467 * Fix WIN32 build system to correctly link an ENGINE directory into
8472 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8474 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8477 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8479 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8481 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8482 context. The operation can be customised via the ctrl mechanism in
8483 case ENGINEs want to include additional functionality.
8487 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8491 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8492 output hashes compatible with older versions of OpenSSL.
8494 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8496 * Fix compression algorithm handling: if resuming a session use the
8497 compression algorithm of the resumed session instead of determining
8498 it from client hello again. Don't allow server to change algorithm.
8502 * Add load_crls() function to commands tidying load_certs() too. Add option
8503 to verify utility to allow additional CRLs to be included.
8507 * Update OCSP request code to permit adding custom headers to the request:
8508 some responders need this.
8512 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8515 *Julia Lawall <julia@diku.dk>*
8517 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8518 needlessly dereferenced structures, used obsolete functions and
8519 didn't handle all updated verify codes correctly.
8523 * Disable MD2 in the default configuration.
8527 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8528 indicate the initial BIO being pushed or popped. This makes it possible
8529 to determine whether the BIO is the one explicitly called or as a result
8530 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8531 it handles reference counts correctly and doesn't zero out the I/O bio
8532 when it is not being explicitly popped. WARNING: applications which
8533 included workarounds for the old buggy behaviour will need to be modified
8534 or they could free up already freed BIOs.
8538 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8539 renaming to all platforms (within the 0.9.8 branch, this was
8540 done conditionally on Netware platforms to avoid a name clash).
8542 *Guenter <lists@gknw.net>*
8544 * Add ECDHE and PSK support to DTLS.
8546 *Michael Tuexen <tuexen@fh-muenster.de>*
8548 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8553 * Add "missing" function EVP_MD_flags() (without this the only way to
8554 retrieve a digest flags is by accessing the structure directly. Update
8555 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8556 or cipher is registered as in the "from" argument. Print out all
8557 registered digests in the dgst usage message instead of manually
8558 attempting to work them out.
8562 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8563 this allows the use of compression and extensions. Change default cipher
8564 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8565 by default unless an application cipher string requests it.
8569 * Alter match criteria in PKCS12_parse(). It used to try to use local
8570 key ids to find matching certificates and keys but some PKCS#12 files
8571 don't follow the (somewhat unwritten) rules and this strategy fails.
8572 Now just gather all certificates together and the first private key
8573 then look for the first certificate that matches the key.
8577 * Support use of registered digest and cipher names for dgst and cipher
8578 commands instead of having to add each one as a special case. So now
8585 openssl dgst -sha256 foo
8587 and this works for ENGINE based algorithms too.
8591 * Update Gost ENGINE to support parameter files.
8593 *Victor B. Wagner <vitus@cryptocom.ru>*
8595 * Support GeneralizedTime in ca utility.
8597 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8599 * Enhance the hash format used for certificate directory links. The new
8600 form uses the canonical encoding (meaning equivalent names will work
8601 even if they aren't identical) and uses SHA1 instead of MD5. This form
8602 is incompatible with the older format and as a result c_rehash should
8603 be used to rebuild symbolic links.
8607 * Make PKCS#8 the default write format for private keys, replacing the
8608 traditional format. This form is standardised, more secure and doesn't
8609 include an implicit MD5 dependency.
8613 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8614 committed to OpenSSL should pass this lot as a minimum.
8618 * Add session ticket override functionality for use by EAP-FAST.
8620 *Jouni Malinen <j@w1.fi>*
8622 * Modify HMAC functions to return a value. Since these can be implemented
8623 in an ENGINE errors can occur.
8627 * Type-checked OBJ_bsearch_ex.
8631 * Type-checked OBJ_bsearch. Also some constification necessitated
8632 by type-checking. Still to come: TXT_DB, bsearch(?),
8633 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8638 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8639 seconds to a tm structure directly, instead of going through OS
8640 specific date routines. This avoids any issues with OS routines such
8641 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8642 and X509_time_adj_ex() to cover the extended range. The existing
8643 X509_time_adj() is still usable and will no longer have any date issues.
8647 * Delta CRL support. New use deltas option which will attempt to locate
8648 and search any appropriate delta CRLs available.
8650 This work was sponsored by Google.
8654 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8655 code and add additional score elements. Validate alternate CRL paths
8656 as part of the CRL checking and indicate a new error "CRL path validation
8657 error" in this case. Applications wanting additional details can use
8658 the verify callback and check the new "parent" field. If this is not
8659 NULL CRL path validation is taking place. Existing applications won't
8660 see this because it requires extended CRL support which is off by
8663 This work was sponsored by Google.
8667 * Support for freshest CRL extension.
8669 This work was sponsored by Google.
8673 * Initial indirect CRL support. Currently only supported in the CRLs
8674 passed directly and not via lookup. Process certificate issuer
8675 CRL entry extension and lookup CRL entries by bother issuer name
8676 and serial number. Check and process CRL issuer entry in IDP extension.
8678 This work was sponsored by Google.
8682 * Add support for distinct certificate and CRL paths. The CRL issuer
8683 certificate is validated separately in this case. Only enabled if
8684 an extended CRL support flag is set: this flag will enable additional
8685 CRL functionality in future.
8687 This work was sponsored by Google.
8691 * Add support for policy mappings extension.
8693 This work was sponsored by Google.
8697 * Fixes to pathlength constraint, self issued certificate handling,
8698 policy processing to align with RFC3280 and PKITS tests.
8700 This work was sponsored by Google.
8704 * Support for name constraints certificate extension. DN, email, DNS
8705 and URI types are currently supported.
8707 This work was sponsored by Google.
8711 * To cater for systems that provide a pointer-based thread ID rather
8712 than numeric, deprecate the current numeric thread ID mechanism and
8713 replace it with a structure and associated callback type. This
8714 mechanism allows a numeric "hash" to be extracted from a thread ID in
8715 either case, and on platforms where pointers are larger than 'long',
8716 mixing is done to help ensure the numeric 'hash' is usable even if it
8717 can't be guaranteed unique. The default mechanism is to use "&errno"
8718 as a pointer-based thread ID to distinguish between threads.
8720 Applications that want to provide their own thread IDs should now use
8721 CRYPTO_THREADID_set_callback() to register a callback that will call
8722 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8724 Note that ERR_remove_state() is now deprecated, because it is tied
8725 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8726 to free the current thread's error state should be replaced by
8727 ERR_remove_thread_state(NULL).
8729 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8730 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8731 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8732 application was previously providing a numeric thread callback that
8733 was inappropriate for distinguishing threads, then uniqueness might
8734 have been obtained with &errno that happened immediately in the
8735 intermediate development versions of OpenSSL; this is no longer the
8736 case, the numeric thread callback will now override the automatic use
8739 *Geoff Thorpe, with help from Bodo Moeller*
8741 * Initial support for different CRL issuing certificates. This covers a
8742 simple case where the self issued certificates in the chain exist and
8743 the real CRL issuer is higher in the existing chain.
8745 This work was sponsored by Google.
8749 * Removed effectively defunct crypto/store from the build.
8753 * Revamp of STACK to provide stronger type-checking. Still to come:
8754 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8755 ASN1_STRING, CONF_VALUE.
8759 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8760 RAM on SSL connections. This option can save about 34k per idle SSL.
8764 * Revamp of LHASH to provide stronger type-checking. Still to come:
8765 STACK, TXT_DB, bsearch, qsort.
8769 * Initial support for Cryptographic Message Syntax (aka CMS) based
8770 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8771 support for data, signedData, compressedData, digestedData and
8772 encryptedData, envelopedData types included. Scripts to check against
8773 RFC4134 examples draft and interop and consistency checks of many
8774 content types and variants.
8778 * Add options to enc utility to support use of zlib compression BIO.
8782 * Extend mk1mf to support importing of options and assembly language
8783 files from Configure script, currently only included in VC-WIN32.
8784 The assembly language rules can now optionally generate the source
8785 files from the associated perl scripts.
8789 * Implement remaining functionality needed to support GOST ciphersuites.
8790 Interop testing has been performed using CryptoPro implementations.
8792 *Victor B. Wagner <vitus@cryptocom.ru>*
8794 * s390x assembler pack.
8798 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8803 * Implement Opaque PRF Input TLS extension as specified in
8804 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8805 official specification yet and no extension type assignment by
8806 IANA exists, this extension (for now) will have to be explicitly
8807 enabled when building OpenSSL by providing the extension number
8808 to use. For example, specify an option
8810 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8812 to the "config" or "Configure" script to enable the extension,
8813 assuming extension number 0x9527 (which is a completely arbitrary
8814 and unofficial assignment based on the MD5 hash of the Internet
8815 Draft). Note that by doing so, you potentially lose
8816 interoperability with other TLS implementations since these might
8817 be using the same extension number for other purposes.
8819 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8820 opaque PRF input value to use in the handshake. This will create
8821 an internal copy of the length-'len' string at 'src', and will
8822 return non-zero for success.
8824 To get more control and flexibility, provide a callback function
8827 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8828 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8832 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8835 Callback function 'cb' will be called in handshakes, and is
8836 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8837 Argument 'arg' is for application purposes (the value as given to
8838 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8839 be provided to the callback function). The callback function
8840 has to return non-zero to report success: usually 1 to use opaque
8841 PRF input just if possible, or 2 to enforce use of the opaque PRF
8842 input. In the latter case, the library will abort the handshake
8843 if opaque PRF input is not successfully negotiated.
8845 Arguments 'peerinput' and 'len' given to the callback function
8846 will always be NULL and 0 in the case of a client. A server will
8847 see the client's opaque PRF input through these variables if
8848 available (NULL and 0 otherwise). Note that if the server
8849 provides an opaque PRF input, the length must be the same as the
8850 length of the client's opaque PRF input.
8852 Note that the callback function will only be called when creating
8853 a new session (session resumption can resume whatever was
8854 previously negotiated), and will not be called in SSL 2.0
8855 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8856 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8857 for applications that need to enforce opaque PRF input.
8861 * Update ssl code to support digests other than SHA1+MD5 for handshake
8864 *Victor B. Wagner <vitus@cryptocom.ru>*
8866 * Add RFC4507 support to OpenSSL. This includes the corrections in
8867 RFC4507bis. The encrypted ticket format is an encrypted encoded
8868 SSL_SESSION structure, that way new session features are automatically
8871 If a client application caches session in an SSL_SESSION structure
8872 support is transparent because tickets are now stored in the encoded
8875 The SSL_CTX structure automatically generates keys for ticket
8876 protection in servers so again support should be possible
8877 with no application modification.
8879 If a client or server wishes to disable RFC4507 support then the option
8880 SSL_OP_NO_TICKET can be set.
8882 Add a TLS extension debugging callback to allow the contents of any client
8883 or server extensions to be examined.
8885 This work was sponsored by Google.
8889 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8890 OpenSSL should now compile cleanly on gcc 4.2
8892 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8894 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8895 support including streaming MAC support: this is required for GOST
8896 ciphersuite support.
8898 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8900 * Add option -stream to use PKCS#7 streaming in smime utility. New
8901 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8902 to output in BER and PEM format.
8906 * Experimental support for use of HMAC via EVP_PKEY interface. This
8907 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8908 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8909 ENGINE support for HMAC keys which are unextractable. New -mac and
8910 -macopt options to dgst utility.
8914 * New option -sigopt to dgst utility. Update dgst to use
8915 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8916 alternative signing parameters such as X9.31 or PSS in the dgst
8921 * Change ssl_cipher_apply_rule(), the internal function that does
8922 the work each time a ciphersuite string requests enabling
8923 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8924 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8925 the order of disabled ciphersuites such that those ciphersuites
8926 that most recently went from enabled to disabled not only stay
8927 in order with respect to each other, but also have higher priority
8928 than other disabled ciphersuites the next time ciphersuites are
8931 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8932 the same ciphersuites as with "HIGH" alone, but in a specific
8933 order where the PSK ciphersuites come first (since they are the
8934 most recently disabled ciphersuites when "HIGH" is parsed).
8936 Also, change ssl_create_cipher_list() (using this new
8937 functionality) such that between otherwise identical
8938 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8943 * Change ssl_create_cipher_list() so that it automatically
8944 arranges the ciphersuites in reasonable order before starting
8945 to process the rule string. Thus, the definition for "DEFAULT"
8946 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8947 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8948 This makes it much easier to arrive at a reasonable default order
8949 in applications for which anonymous ciphers are OK (meaning
8950 that you can't actually use DEFAULT).
8952 *Bodo Moeller; suggested by Victor Duchovni*
8954 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8955 processing) into multiple integers instead of setting
8956 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8957 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8958 (These masks as well as the individual bit definitions are hidden
8959 away into the non-exported interface ssl/ssl_locl.h, so this
8960 change to the definition of the SSL_CIPHER structure shouldn't
8961 affect applications.) This give us more bits for each of these
8962 categories, so there is no longer a need to coagulate AES128 and
8963 AES256 into a single algorithm bit, and to coagulate Camellia128
8964 and Camellia256 into a single algorithm bit, which has led to all
8967 Thus, among other things, the kludge introduced in 0.9.7m and
8968 0.9.8e for masking out AES256 independently of AES128 or masking
8969 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8971 With the change, we also introduce new ciphersuite aliases that
8972 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8977 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8978 Use the leftmost N bytes of the signature input if the input is
8979 larger than the prime q (with N being the size in bytes of q).
8983 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8984 it yet and it is largely untested.
8988 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8992 * Initial incomplete changes to avoid need for function casts in OpenSSL
8993 some compilers (gcc 4.2 and later) reject their use. Safestack is
8994 reimplemented. Update ASN1 to avoid use of legacy functions.
8998 * Win32/64 targets are linked with Winsock2.
9002 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9003 to external functions. This can be used to increase CRL handling
9004 efficiency especially when CRLs are very large by (for example) storing
9005 the CRL revoked certificates in a database.
9009 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9010 new CRLs added to a directory can be used. New command line option
9011 -verify_return_error to s_client and s_server. This causes real errors
9012 to be returned by the verify callback instead of carrying on no matter
9013 what. This reflects the way a "real world" verify callback would behave.
9017 * GOST engine, supporting several GOST algorithms and public key formats.
9018 Kindly donated by Cryptocom.
9022 * Partial support for Issuing Distribution Point CRL extension. CRLs
9023 partitioned by DP are handled but no indirect CRL or reason partitioning
9024 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9025 selected via a scoring technique which handles IDP and AKID in CRLs.
9029 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9030 will ultimately be used for all verify operations: this will remove the
9031 X509_STORE dependency on certificate verification and allow alternative
9032 lookup methods. X509_STORE based implementations of these two callbacks.
9036 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9037 Modify get_crl() to find a valid (unexpired) CRL if possible.
9041 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9042 this would be called X509_CRL_cmp() but that name is already used by
9043 a function that just compares CRL issuer names. Cache several CRL
9044 extensions in X509_CRL structure and cache CRLDP in X509.
9048 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9049 this maps equivalent X509_NAME structures into a consistent structure.
9050 Name comparison can then be performed rapidly using memcmp().
9054 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9059 * Allow digests to supply their own micalg string for S/MIME type using
9060 the ctrl EVP_MD_CTRL_MICALG.
9064 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9065 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9066 ctrl. It can then customise the structure before and/or after signing
9071 * New function OBJ_add_sigid() to allow application defined signature OIDs
9072 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9073 to free up any added signature OIDs.
9077 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9078 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9079 digest and cipher tables. New options added to openssl utility:
9080 list-message-digest-algorithms and list-cipher-algorithms.
9084 * Change the array representation of binary polynomials: the list
9085 of degrees of non-zero coefficients is now terminated with -1.
9086 Previously it was terminated with 0, which was also part of the
9087 value; thus, the array representation was not applicable to
9088 polynomials where t^0 has coefficient zero. This change makes
9089 the array representation useful in a more general context.
9093 * Various modifications and fixes to SSL/TLS cipher string
9094 handling. For ECC, the code now distinguishes between fixed ECDH
9095 with RSA certificates on the one hand and with ECDSA certificates
9096 on the other hand, since these are separate ciphersuites. The
9097 unused code for Fortezza ciphersuites has been removed.
9099 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9100 (not "ECDHE"). For consistency with the code for DH
9101 certificates, use of ECDH certificates is now considered ECDH
9102 authentication, not RSA or ECDSA authentication (the latter is
9103 merely the CA's signing algorithm and not actively used in the
9106 The temporary ciphersuite alias "ECCdraft" is no longer
9107 available, and ECC ciphersuites are no longer excluded from "ALL"
9108 and "DEFAULT". The following aliases now exist for RFC 4492
9109 ciphersuites, most of these by analogy with the DH case:
9111 kECDHr - ECDH cert, signed with RSA
9112 kECDHe - ECDH cert, signed with ECDSA
9113 kECDH - ECDH cert (signed with either RSA or ECDSA)
9114 kEECDH - ephemeral ECDH
9115 ECDH - ECDH cert or ephemeral ECDH
9121 AECDH - anonymous ECDH
9122 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9126 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9127 Use correct micalg parameters depending on digest(s) in signed message.
9131 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9132 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9136 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9137 an engine to register a method. Add ENGINE lookups for methods and
9138 functional reference processing.
9142 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9143 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9148 * New -resign option to smime utility. This adds one or more signers
9149 to an existing PKCS#7 signedData structure. Also -md option to use an
9150 alternative message digest algorithm for signing.
9154 * Tidy up PKCS#7 routines and add new functions to make it easier to
9155 create PKCS7 structures containing multiple signers. Update smime
9156 application to support multiple signers.
9160 * New -macalg option to pkcs12 utility to allow setting of an alternative
9165 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9166 Reorganize PBE internals to lookup from a static table using NIDs,
9167 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9168 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9169 PRF which will be automatically used with PBES2.
9173 * Replace the algorithm specific calls to generate keys in "req" with the
9178 * Update PKCS#7 enveloped data routines to use new API. This is now
9179 supported by any public key method supporting the encrypt operation. A
9180 ctrl is added to allow the public key algorithm to examine or modify
9181 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9186 * Add a ctrl to asn1 method to allow a public key algorithm to express
9187 a default digest type to use. In most cases this will be SHA1 but some
9188 algorithms (such as GOST) need to specify an alternative digest. The
9189 return value indicates how strong the preference is 1 means optional and
9190 2 is mandatory (that is it is the only supported type). Modify
9191 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9192 use the default md. Update openssl utilities to use the default digest
9193 type for signing if it is not explicitly indicated.
9197 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9198 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9199 signing method from the key type. This effectively removes the link
9200 between digests and public key types.
9204 * Add an OID cross reference table and utility functions. Its purpose is to
9205 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9206 rsaEncryption. This will allow some of the algorithm specific hackery
9207 needed to use the correct OID to be removed.
9211 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9212 structures for PKCS7_sign(). They are now set up by the relevant public
9217 * Add provisional EC pkey method with support for ECDSA and ECDH.
9221 * Add support for key derivation (agreement) in the API, DH method and
9226 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9227 public and private key formats. As a side effect these add additional
9228 command line functionality not previously available: DSA signatures can be
9229 generated and verified using pkeyutl and DH key support and generation in
9236 *Oliver Tappe <zooey@hirschkaefer.de>*
9238 * New make target "install_html_docs" installs HTML renditions of the
9241 *Oliver Tappe <zooey@hirschkaefer.de>*
9243 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9244 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9245 support key and parameter generation and add initial key generation
9246 functionality for RSA.
9250 * Add functions for main EVP_PKEY_method operations. The undocumented
9251 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9252 `EVP_PKEY_{encrypt,decrypt}_old`.
9256 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9257 key API, doesn't do much yet.
9261 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9262 public key algorithms. New option to openssl utility:
9263 "list-public-key-algorithms" to print out info.
9267 * Implement the Supported Elliptic Curves Extension for
9268 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9272 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9273 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9277 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9278 utilities such as rsa, dsa, dsaparam etc except they process any key
9283 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9284 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9285 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9290 * Initial support for pluggable public key ASN1.
9291 De-spaghettify the public key ASN1 handling. Move public and private
9292 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9293 algorithm specific handling to a single module within the relevant
9294 algorithm directory. Add functions to allow (near) opaque processing
9295 of public and private key structures.
9299 * Implement the Supported Point Formats Extension for
9300 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9304 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9305 for the psk identity [hint] and the psk callback functions to the
9306 SSL_SESSION, SSL and SSL_CTX structure.
9309 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9313 SSL_CTX_use_psk_identity_hint
9314 SSL_get_psk_identity_hint
9315 SSL_get_psk_identity
9316 SSL_use_psk_identity_hint
9318 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9320 * Add RFC 3161 compliant time stamp request creation, response generation
9321 and response verification functionality.
9323 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9325 * Add initial support for TLS extensions, specifically for the server_name
9326 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9327 have new members for a host name. The SSL data structure has an
9328 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9329 stored in that context to allow for session resumption, even after the
9330 SSL has been switched to a new SSL_CTX in reaction to a client's
9331 server_name extension.
9333 New functions (subject to change):
9335 SSL_get_servername()
9336 SSL_get_servername_type()
9339 New CTRL codes and macros (subject to change):
9341 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9342 - SSL_CTX_set_tlsext_servername_callback()
9343 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9344 - SSL_CTX_set_tlsext_servername_arg()
9345 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9347 openssl s_client has a new '-servername ...' option.
9349 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9350 '-key2 ...', '-servername_fatal' (subject to change). This allows
9351 testing the HostName extension for a specific single host name ('-cert'
9352 and '-key' remain fallbacks for handshakes without HostName
9353 negotiation). If the unrecognized_name alert has to be sent, this by
9354 default is a warning; it becomes fatal with the '-servername_fatal'
9357 *Peter Sylvester, Remy Allais, Christophe Renou*
9359 * Whirlpool hash implementation is added.
9363 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9364 bn(64,32). Because of instruction set limitations it doesn't have
9365 any negative impact on performance. This was done mostly in order
9366 to make it possible to share assembler modules, such as bn_mul_mont
9367 implementations, between 32- and 64-bit builds without hassle.
9371 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9372 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9377 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9378 dedicated Montgomery multiplication procedure, is introduced.
9379 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9380 "64-bit" performance on certain 32-bit targets.
9384 * New option SSL_OP_NO_COMP to disable use of compression selectively
9385 in SSL structures. New SSL ctrl to set maximum send fragment size.
9386 Save memory by setting the I/O buffer sizes dynamically instead of
9387 using the maximum available value.
9391 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9392 in addition to the text details.
9396 * Very, very preliminary EXPERIMENTAL support for printing of general
9397 ASN1 structures. This currently produces rather ugly output and doesn't
9398 handle several customised structures at all.
9402 * Integrated support for PVK file format and some related formats such
9403 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9404 these in the 'rsa' and 'dsa' utilities.
9408 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9412 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9413 place for the (very old) "NETSCAPE" format certificates which are now
9414 handled using new ASN1 code equivalents.
9418 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9419 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9420 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9424 * Modify CRL distribution points extension code to print out previously
9425 unsupported fields. Enhance extension setting code to allow setting of
9430 * Add print and set support for Issuing Distribution Point CRL extension.
9434 * Change 'Configure' script to enable Camellia by default.
9441 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9443 * When rejecting SSL/TLS records due to an incorrect version number, never
9444 update s->server with a new major version number. As of
9445 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9446 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9447 the previous behavior could result in a read attempt at NULL when
9448 receiving specific incorrect SSL/TLS records once record payload
9449 protection is active. ([CVE-2010-0740])
9451 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9453 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9454 could be crashed if the relevant tables were not present (e.g. chrooted).
9456 *Tomas Hoger <thoger@redhat.com>*
9458 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9460 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9462 *Martin Olsson, Neel Mehta*
9464 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9465 accommodate for stack sorting, always a write lock!).
9469 * On some versions of WIN32 Heap32Next is very slow. This can cause
9470 excessive delays in the RAND_poll(): over a minute. As a workaround
9471 include a time check in the inner Heap32Next loop too.
9475 * The code that handled flushing of data in SSL/TLS originally used the
9476 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9477 the problem outlined in PR#1949. The fix suggested there however can
9478 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9479 of Apache). So instead simplify the code to flush unconditionally.
9480 This should be fine since flushing with no data to flush is a no op.
9484 * Handle TLS versions 2.0 and later properly and correctly use the
9485 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9486 off ancient servers have a habit of sticking around for a while...
9490 * Modify compression code so it frees up structures without using the
9491 ex_data callbacks. This works around a problem where some applications
9492 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9493 restarting) then use compression (e.g. SSL with compression) later.
9494 This results in significant per-connection memory leaks and
9495 has caused some security issues including CVE-2008-1678 and
9500 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9501 change when encrypting or decrypting.
9505 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9506 connect and renegotiate with servers which do not support RI.
9507 Until RI is more widely deployed this option is enabled by default.
9511 * Add "missing" ssl ctrls to clear options and mode.
9515 * If client attempts to renegotiate and doesn't support RI respond with
9516 a no_renegotiation alert as required by RFC5746. Some renegotiating
9517 TLS clients will continue a connection gracefully when they receive
9518 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9519 waiting for a server hello which it will never receive. Now we treat a
9520 received no_renegotiation alert as a fatal error. This is because
9521 applications requesting a renegotiation might well expect it to succeed
9522 and would have no code in place to handle the server denying it so the
9523 only safe thing to do is to terminate the connection.
9527 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9528 peer supports secure renegotiation and 0 otherwise. Print out peer
9529 renegotiation support in s_client/s_server.
9533 * Replace the highly broken and deprecated SPKAC certification method with
9534 the updated NID creation version. This should correctly handle UTF8.
9538 * Implement RFC5746. Re-enable renegotiation but require the extension
9539 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9540 turns out to be a bad idea. It has been replaced by
9541 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9542 SSL_CTX_set_options(). This is really not recommended unless you
9543 know what you are doing.
9545 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9547 * Fixes to stateless session resumption handling. Use initial_ctx when
9548 issuing and attempting to decrypt tickets in case it has changed during
9549 servername handling. Use a non-zero length session ID when attempting
9550 stateless session resumption: this makes it possible to determine if
9551 a resumption has occurred immediately after receiving server hello
9552 (several places in OpenSSL subtly assume this) instead of later in
9557 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9558 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9559 fixes for a few places where the return code is not checked
9562 *Julia Lawall <julia@diku.dk>*
9564 * Add --strict-warnings option to Configure script to include devteam
9565 warnings in other configurations.
9569 * Add support for --libdir option and LIBDIR variable in makefiles. This
9570 makes it possible to install openssl libraries in locations which
9571 have names other than "lib", for example "/usr/lib64" which some
9574 *Steve Henson, based on patch from Jeremy Utley*
9576 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9577 X690 8.9.12 and can produce some misleading textual output of OIDs.
9579 *Steve Henson, reported by Dan Kaminsky*
9581 * Delete MD2 from algorithm tables. This follows the recommendation in
9582 several standards that it is not used in new applications due to
9583 several cryptographic weaknesses. For binary compatibility reasons
9584 the MD2 API is still compiled in by default.
9588 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9593 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9594 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9597 *Guenter <lists@gknw.net>*
9599 * Fix the server certificate chain building code to use X509_verify_cert(),
9600 it used to have an ad-hoc builder which was unable to cope with anything
9601 other than a simple chain.
9603 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9605 * Don't check self signed certificate signatures in X509_verify_cert()
9606 by default (a flag can override this): it just wastes time without
9607 adding any security. As a useful side effect self signed root CAs
9608 with non-FIPS digests are now usable in FIPS mode.
9612 * In dtls1_process_out_of_seq_message() the check if the current message
9613 is already buffered was missing. For every new message was memory
9614 allocated, allowing an attacker to perform an denial of service attack
9615 with sending out of seq handshake messages until there is no memory
9616 left. Additionally every future message was buffered, even if the
9617 sequence number made no sense and would be part of another handshake.
9618 So only messages with sequence numbers less than 10 in advance will be
9619 buffered. ([CVE-2009-1378])
9621 *Robin Seggelmann, discovered by Daniel Mentz*
9623 * Records are buffered if they arrive with a future epoch to be
9624 processed after finishing the corresponding handshake. There is
9625 currently no limitation to this buffer allowing an attacker to perform
9626 a DOS attack with sending records with future epochs until there is no
9627 memory left. This patch adds the pqueue_size() function to determine
9628 the size of a buffer and limits the record buffer to 100 entries.
9631 *Robin Seggelmann, discovered by Daniel Mentz*
9633 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9634 parent structure is freed. ([CVE-2009-1379])
9638 * Handle non-blocking I/O properly in SSL_shutdown() call.
9640 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9642 * Add `2.5.4.*` OIDs
9644 *Ilya O. <vrghost@gmail.com>*
9646 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9648 * Disable renegotiation completely - this fixes a severe security
9649 problem ([CVE-2009-3555]) at the cost of breaking all
9650 renegotiation. Renegotiation can be re-enabled by setting
9651 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9652 run-time. This is really not recommended unless you know what
9657 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9659 * Don't set val to NULL when freeing up structures, it is freed up by
9660 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9661 zeroing past the valid field. ([CVE-2009-0789])
9663 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9665 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9666 checked correctly. This would allow some invalid signed attributes to
9667 appear to verify correctly. ([CVE-2009-0591])
9669 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9671 * Reject UniversalString and BMPString types with invalid lengths. This
9672 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9673 a legal length. ([CVE-2009-0590])
9677 * Set S/MIME signing as the default purpose rather than setting it
9678 unconditionally. This allows applications to override it at the store
9683 * Permit restricted recursion of ASN1 strings. This is needed in practice
9684 to handle some structures.
9688 * Improve efficiency of mem_gets: don't search whole buffer each time
9691 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9693 * New -hex option for openssl rand.
9697 * Print out UTF8String and NumericString when parsing ASN1.
9701 * Support NumericString type for name components.
9705 * Allow CC in the environment to override the automatically chosen
9706 compiler. Note that nothing is done to ensure flags work with the
9711 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9713 * Properly check EVP_VerifyFinal() and similar return values
9716 *Ben Laurie, Bodo Moeller, Google Security Team*
9718 * Enable TLS extensions by default.
9722 * Allow the CHIL engine to be loaded, whether the application is
9723 multithreaded or not. (This does not release the developer from the
9724 obligation to set up the dynamic locking callbacks.)
9726 *Sander Temme <sander@temme.net>*
9728 * Use correct exit code if there is an error in dgst command.
9730 *Steve Henson; problem pointed out by Roland Dirlewanger*
9732 * Tweak Configure so that you need to say "experimental-jpake" to enable
9733 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9737 * Add experimental JPAKE support, including demo authentication in
9738 s_client and s_server.
9742 * Set the comparison function in v3_addr_canonize().
9744 *Rob Austein <sra@hactrn.net>*
9746 * Add support for XMPP STARTTLS in s_client.
9748 *Philip Paeps <philip@freebsd.org>*
9750 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9751 to ensure that even with this option, only ciphersuites in the
9752 server's preference list will be accepted. (Note that the option
9753 applies only when resuming a session, so the earlier behavior was
9754 just about the algorithm choice for symmetric cryptography.)
9758 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9760 * Fix NULL pointer dereference if a DTLS server received
9761 ChangeCipherSpec as first record ([CVE-2009-1386]).
9765 * Fix a state transition in s3_srvr.c and d1_srvr.c
9766 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9770 * The fix in 0.9.8c that supposedly got rid of unsafe
9771 double-checked locking was incomplete for RSA blinding,
9772 addressing just one layer of what turns out to have been
9773 doubly unsafe triple-checked locking.
9775 So now fix this for real by retiring the MONT_HELPER macro
9776 in crypto/rsa/rsa_eay.c.
9778 *Bodo Moeller; problem pointed out by Marius Schilder*
9780 * Various precautionary measures:
9782 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9784 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9785 (NB: This would require knowledge of the secret session ticket key
9786 to exploit, in which case you'd be SOL either way.)
9788 - Change bn_nist.c so that it will properly handle input BIGNUMs
9789 outside the expected range.
9791 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9794 *Neel Mehta, Bodo Moeller*
9796 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9797 the load fails. Useful for distros.
9799 *Ben Laurie and the FreeBSD team*
9801 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9805 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9809 * Expand ENGINE to support engine supplied SSL client certificate functions.
9811 This work was sponsored by Logica.
9815 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9816 keystores. Support for SSL/TLS client authentication too.
9817 Not compiled unless enable-capieng specified to Configure.
9819 This work was sponsored by Logica.
9823 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9824 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9825 attribute creation routines such as certificate requests and PKCS#12
9830 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9832 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9833 handshake which could lead to a client crash as found using the
9834 Codenomicon TLS test suite ([CVE-2008-1672])
9836 *Steve Henson, Mark Cox*
9838 * Fix double free in TLS server name extensions which could lead to
9839 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9843 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9845 Clear the error queue to ensure that error entries left from
9846 older function calls do not interfere with the correct operation.
9848 *Lutz Jaenicke, Erik de Castro Lopo*
9850 * Remove root CA certificates of commercial CAs:
9852 The OpenSSL project does not recommend any specific CA and does not
9853 have any policy with respect to including or excluding any CA.
9854 Therefore it does not make any sense to ship an arbitrary selection
9855 of root CA certificates with the OpenSSL software.
9859 * RSA OAEP patches to fix two separate invalid memory reads.
9860 The first one involves inputs when 'lzero' is greater than
9861 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9862 before the beginning of from). The second one involves inputs where
9863 the 'db' section contains nothing but zeroes (there is a one-byte
9864 invalid read after the end of 'db').
9866 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9868 * Partial backport from 0.9.9-dev:
9870 Introduce bn_mul_mont (dedicated Montgomery multiplication
9871 procedure) as a candidate for BIGNUM assembler implementation.
9872 While 0.9.9-dev uses assembler for various architectures, only
9873 x86_64 is available by default here in the 0.9.8 branch, and
9874 32-bit x86 is available through a compile-time setting.
9876 To try the 32-bit x86 assembler implementation, use Configure
9877 option "enable-montasm" (which exists only for this backport).
9879 As "enable-montasm" for 32-bit x86 disclaims code stability
9880 anyway, in this constellation we activate additional code
9881 backported from 0.9.9-dev for further performance improvements,
9882 namely BN_from_montgomery_word. (To enable this otherwise,
9883 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9885 *Andy Polyakov (backport partially by Bodo Moeller)*
9887 * Add TLS session ticket callback. This allows an application to set
9888 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9889 values. This is useful for key rollover for example where several key
9890 sets may exist with different names.
9894 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9895 This was broken until now in 0.9.8 releases, such that the only way
9896 a registered ENGINE could be used (assuming it initialises
9897 successfully on the host) was to explicitly set it as the default
9898 for the relevant algorithms. This is in contradiction with 0.9.7
9899 behaviour and the documentation. With this fix, when an ENGINE is
9900 registered into a given algorithm's table of implementations, the
9901 'uptodate' flag is reset so that auto-discovery will be used next
9902 time a new context for that algorithm attempts to select an
9905 *Ian Lister (tweaked by Geoff Thorpe)*
9907 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9908 implementation in the following ways:
9910 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9913 Lack of BER streaming support means one pass streaming processing is
9914 only supported if data is detached: setting the streaming flag is
9915 ignored for embedded content.
9917 CMS support is disabled by default and must be explicitly enabled
9918 with the enable-cms configuration option.
9922 * Update the GMP engine glue to do direct copies between BIGNUM and
9923 mpz_t when openssl and GMP use the same limb size. Otherwise the
9924 existing "conversion via a text string export" trick is still used.
9926 *Paul Sheer <paulsheer@gmail.com>*
9928 * Zlib compression BIO. This is a filter BIO which compressed and
9929 uncompresses any data passed through it.
9933 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9934 RFC3394 compatible AES key wrapping.
9938 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9939 sets string data without copying. X509_ALGOR_set0() and
9940 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9941 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9942 from an X509_ATTRIBUTE structure optionally checking it occurs only
9943 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9948 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9949 to get the expected BN_FLG_CONSTTIME behavior.
9951 *Bodo Moeller (Google)*
9955 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9956 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9957 - added some more tests to do_tests.pl
9958 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9959 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9960 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9961 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9962 - various changes to netware.pl to enable gcc-cross builds on Win32
9964 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9965 - various changes to fix missing prototype warnings
9966 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9967 - added AES, WHIRLPOOL and CPUID assembler code to build files
9968 - added missing AES assembler make rules to mk1mf.pl
9969 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9971 *Guenter Knauf <eflash@gmx.net>*
9973 * Implement certificate status request TLS extension defined in RFC3546.
9974 A client can set the appropriate parameters and receive the encoded
9975 OCSP response via a callback. A server can query the supplied parameters
9976 and set the encoded OCSP response in the callback. Add simplified examples
9977 to s_client and s_server.
9981 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9984 + Binary incompatibility of ssl_ctx_st structure
9985 + DTLS interoperation with non-compliant servers
9986 + Don't call get_session_cb() without proposed session
9987 + Fix ia64 assembler code
9989 *Andy Polyakov, Steve Henson*
9991 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9993 * DTLS Handshake overhaul. There were longstanding issues with
9994 OpenSSL DTLS implementation, which were making it impossible for
9995 RFC 4347 compliant client to communicate with OpenSSL server.
9996 Unfortunately just fixing these incompatibilities would "cut off"
9997 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9998 server keeps tolerating non RFC compliant syntax. The opposite is
9999 not true, 0.9.8f client can not communicate with earlier server.
10000 This update even addresses CVE-2007-4995.
10004 * Changes to avoid need for function casts in OpenSSL: some compilers
10005 (gcc 4.2 and later) reject their use.
10006 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10009 * Add RFC4507 support to OpenSSL. This includes the corrections in
10010 RFC4507bis. The encrypted ticket format is an encrypted encoded
10011 SSL_SESSION structure, that way new session features are automatically
10014 If a client application caches session in an SSL_SESSION structure
10015 support is transparent because tickets are now stored in the encoded
10018 The SSL_CTX structure automatically generates keys for ticket
10019 protection in servers so again support should be possible
10020 with no application modification.
10022 If a client or server wishes to disable RFC4507 support then the option
10023 SSL_OP_NO_TICKET can be set.
10025 Add a TLS extension debugging callback to allow the contents of any client
10026 or server extensions to be examined.
10028 This work was sponsored by Google.
10032 * Add initial support for TLS extensions, specifically for the server_name
10033 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10034 have new members for a host name. The SSL data structure has an
10035 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10036 stored in that context to allow for session resumption, even after the
10037 SSL has been switched to a new SSL_CTX in reaction to a client's
10038 server_name extension.
10040 New functions (subject to change):
10042 SSL_get_servername()
10043 SSL_get_servername_type()
10046 New CTRL codes and macros (subject to change):
10048 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10049 - SSL_CTX_set_tlsext_servername_callback()
10050 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10051 - SSL_CTX_set_tlsext_servername_arg()
10052 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10054 openssl s_client has a new '-servername ...' option.
10056 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10057 '-key2 ...', '-servername_fatal' (subject to change). This allows
10058 testing the HostName extension for a specific single host name ('-cert'
10059 and '-key' remain fallbacks for handshakes without HostName
10060 negotiation). If the unrecognized_name alert has to be sent, this by
10061 default is a warning; it becomes fatal with the '-servername_fatal'
10064 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10066 * Add AES and SSE2 assembly language support to VC++ build.
10070 * Mitigate attack on final subtraction in Montgomery reduction.
10074 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10075 (which previously caused an internal error).
10079 * Squeeze another 10% out of IGE mode when in != out.
10083 * AES IGE mode speedup.
10085 *Dean Gaudet (Google)*
10087 * Add the Korean symmetric 128-bit cipher SEED (see
10088 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10089 add SEED ciphersuites from RFC 4162:
10091 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10092 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10093 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10094 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10096 To minimize changes between patchlevels in the OpenSSL 0.9.8
10097 series, SEED remains excluded from compilation unless OpenSSL
10098 is configured with 'enable-seed'.
10100 *KISA, Bodo Moeller*
10102 * Mitigate branch prediction attacks, which can be practical if a
10103 single processor is shared, allowing a spy process to extract
10104 information. For detailed background information, see
10105 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10106 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10107 and Necessary Software Countermeasures"). The core of the change
10108 are new versions BN_div_no_branch() and
10109 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10110 respectively, which are slower, but avoid the security-relevant
10111 conditional branches. These are automatically called by BN_div()
10112 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10113 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10114 remove a conditional branch.
10116 BN_FLG_CONSTTIME is the new name for the previous
10117 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10118 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10119 in the exponent causes BN_mod_exp_mont() to use the alternative
10120 implementation in BN_mod_exp_mont_consttime().) The old name
10121 remains as a deprecated alias.
10123 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10124 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10125 constant-time implementations for more than just exponentiation.
10126 Here too the old name is kept as a deprecated alias.
10128 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10129 the BN_BLINDING structure gets an independent copy of the
10130 modulus. This means that the previous `BIGNUM *m` argument to
10131 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10132 essentially becomes `const BIGNUM *m`, although we can't actually
10133 change this in the header file before 0.9.9. It allows
10134 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10135 enable BN_FLG_CONSTTIME.
10137 *Matthew D Wood (Intel Corp)*
10139 * In the SSL/TLS server implementation, be strict about session ID
10140 context matching (which matters if an application uses a single
10141 external cache for different purposes). Previously,
10142 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10143 set. This did ensure strict client verification, but meant that,
10144 with applications using a single external cache for quite
10145 different requirements, clients could circumvent ciphersuite
10146 restrictions for a given session ID context by starting a session
10147 in a different context.
10151 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10152 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10153 authentication-only ciphersuites.
10157 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10158 not complete and could lead to a possible single byte overflow
10159 ([CVE-2007-5135]) [Ben Laurie]
10161 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10163 * Since AES128 and AES256 (and similarly Camellia128 and
10164 Camellia256) share a single mask bit in the logic of
10165 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10166 kludge to work properly if AES128 is available and AES256 isn't
10167 (or if Camellia128 is available and Camellia256 isn't).
10171 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10172 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10173 When a point or a seed is encoded in a BIT STRING, we need to
10174 prevent the removal of trailing zero bits to get the proper DER
10175 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10176 of a NamedBitList, for which trailing 0 bits need to be removed.)
10180 * Have SSL/TLS server implementation tolerate "mismatched" record
10181 protocol version while receiving ClientHello even if the
10182 ClientHello is fragmented. (The server can't insist on the
10183 particular protocol version it has chosen before the ServerHello
10184 message has informed the client about his choice.)
10188 * Add RFC 3779 support.
10190 *Rob Austein for ARIN, Ben Laurie*
10192 * Load error codes if they are not already present instead of using a
10193 static variable. This allows them to be cleanly unloaded and reloaded.
10194 Improve header file function name parsing.
10198 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10199 or CAPABILITY handshake as required by RFCs.
10201 *Goetz Babin-Ebell*
10203 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10205 * Introduce limits to prevent malicious keys being able to
10206 cause a denial of service. ([CVE-2006-2940])
10208 *Steve Henson, Bodo Moeller*
10210 * Fix ASN.1 parsing of certain invalid structures that can result
10211 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10213 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10214 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10216 * Fix SSL client code which could crash if connecting to a
10217 malicious SSLv2 server. ([CVE-2006-4343])
10219 *Tavis Ormandy and Will Drewry, Google Security Team*
10221 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10222 match only those. Before that, "AES256-SHA" would be interpreted
10223 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10224 the same strength classification in 0.9.7h) as we currently only
10225 have a single AES bit in the ciphersuite description bitmap.
10226 That change, however, also applied to ciphersuite strings such as
10227 "RC4-MD5" that intentionally matched multiple ciphersuites --
10228 namely, SSL 2.0 ciphersuites in addition to the more common ones
10229 from SSL 3.0/TLS 1.0.
10231 So we change the selection algorithm again: Naming an explicit
10232 ciphersuite selects this one ciphersuite, and any other similar
10233 ciphersuite (same bitmap) from *other* protocol versions.
10234 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10235 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10237 Since SSL 2.0 does not have any ciphersuites for which the
10238 128/256 bit distinction would be relevant, this works for now.
10239 The proper fix will be to use different bits for AES128 and
10240 AES256, which would have avoided the problems from the beginning;
10241 however, bits are scarce, so we can only do this in a new release
10242 (not just a patchlevel) when we can change the SSL_CIPHER
10243 definition to split the single 'unsigned long mask' bitmap into
10244 multiple values to extend the available space.
10248 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10250 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10251 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10253 * Add AES IGE and biIGE modes.
10257 * Change the Unix randomness entropy gathering to use poll() when
10258 possible instead of select(), since the latter has some
10259 undesirable limitations.
10261 *Darryl Miles via Richard Levitte and Bodo Moeller*
10263 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10264 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10265 cannot be implicitly activated as part of, e.g., the "AES" alias.
10266 However, please upgrade to OpenSSL 0.9.9[-dev] for
10267 non-experimental use of the ECC ciphersuites to get TLS extension
10268 support, which is required for curve and point format negotiation
10269 to avoid potential handshake problems.
10273 * Disable rogue ciphersuites:
10275 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10276 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10277 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10279 The latter two were purportedly from
10280 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10283 Also deactivate the remaining ciphersuites from
10284 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10285 unofficial, and the ID has long expired.
10289 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10290 dual-core machines) and other potential thread-safety issues.
10294 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10295 versions), which is now available for royalty-free use
10296 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10297 Also, add Camellia TLS ciphersuites from RFC 4132.
10299 To minimize changes between patchlevels in the OpenSSL 0.9.8
10300 series, Camellia remains excluded from compilation unless OpenSSL
10301 is configured with 'enable-camellia'.
10305 * Disable the padding bug check when compression is in use. The padding
10306 bug check assumes the first packet is of even length, this is not
10307 necessarily true if compression is enabled and can result in false
10308 positives causing handshake failure. The actual bug test is ancient
10309 code so it is hoped that implementations will either have fixed it by
10310 now or any which still have the bug do not support compression.
10314 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10316 * When applying a cipher rule check to see if string match is an explicit
10317 cipher suite and only match that one cipher suite if it is.
10321 * Link in manifests for VC++ if needed.
10323 *Austin Ziegler <halostatue@gmail.com>*
10325 * Update support for ECC-based TLS ciphersuites according to
10326 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10327 TLS extensions, which are supported starting with the 0.9.9
10328 branch, not in the OpenSSL 0.9.8 branch).
10332 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10333 opaque EVP_CIPHER_CTX handling.
10337 * Fixes and enhancements to zlib compression code. We now only use
10338 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10339 to conform with the standards mentioned here:
10340 <http://www.zlib.net/DLL_FAQ.txt>
10341 Static zlib linking now works on Windows and the new --with-zlib-include
10342 --with-zlib-lib options to Configure can be used to supply the location
10343 of the headers and library. Gracefully handle case where zlib library
10348 * Several fixes and enhancements to the OID generation code. The old code
10349 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10350 handle numbers larger than ULONG_MAX, truncated printing and had a
10351 non standard OBJ_obj2txt() behaviour.
10355 * Add support for building of engines under engine/ as shared libraries
10356 under VC++ build system.
10360 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10361 Hopefully, we will not see any false combination of paths any more.
10365 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10367 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10368 (part of SSL_OP_ALL). This option used to disable the
10369 countermeasure against man-in-the-middle protocol-version
10370 rollback in the SSL 2.0 server implementation, which is a bad
10371 idea. ([CVE-2005-2969])
10373 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10374 for Information Security, National Institute of Advanced Industrial
10375 Science and Technology [AIST], Japan)*
10377 * Add two function to clear and return the verify parameter flags.
10381 * Keep cipherlists sorted in the source instead of sorting them at
10382 runtime, thus removing the need for a lock.
10386 * Avoid some small subgroup attacks in Diffie-Hellman.
10388 *Nick Mathewson and Ben Laurie*
10390 * Add functions for well-known primes.
10394 * Extended Windows CE support.
10396 *Satoshi Nakamura and Andy Polyakov*
10398 * Initialize SSL_METHOD structures at compile time instead of during
10399 runtime, thus removing the need for a lock.
10403 * Make PKCS7_decrypt() work even if no certificate is supplied by
10404 attempting to decrypt each encrypted key in turn. Add support to
10409 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10411 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10414 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10418 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10419 key into the same file any more.
10423 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10427 * Add -utf8 command line and config file option to 'ca'.
10429 *Stefan <stf@udoma.org*
10431 * Removed the macro des_crypt(), as it seems to conflict with some
10432 libraries. Use DES_crypt().
10436 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10437 involves renaming the source and generated shared-libs for
10438 both. The engines will accept the corrected or legacy ids
10439 ('ncipher' and '4758_cca' respectively) when binding. NB,
10440 this only applies when building 'shared'.
10442 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10444 * Add attribute functions to EVP_PKEY structure. Modify
10445 PKCS12_create() to recognize a CSP name attribute and
10446 use it. Make -CSP option work again in pkcs12 utility.
10450 * Add new functionality to the bn blinding code:
10451 - automatic re-creation of the BN_BLINDING parameters after
10452 a fixed number of uses (currently 32)
10453 - add new function for parameter creation
10454 - introduce flags to control the update behaviour of the
10455 BN_BLINDING parameters
10456 - hide BN_BLINDING structure
10457 Add a second BN_BLINDING slot to the RSA structure to improve
10458 performance when a single RSA object is shared among several
10463 * Add support for DTLS.
10465 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10467 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10468 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10472 * Remove buggy and incomplete DH cert support from
10473 ssl/ssl_rsa.c and ssl/s3_both.c
10477 * Use SHA-1 instead of MD5 as the default digest algorithm for
10478 the `apps/openssl` commands.
10482 * Compile clean with "-Wall -Wmissing-prototypes
10483 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10484 DEBUG_SAFESTACK must also be set.
10488 * Change ./Configure so that certain algorithms can be disabled by default.
10489 The new counterpiece to "no-xxx" is "enable-xxx".
10491 The patented RC5 and MDC2 algorithms will now be disabled unless
10492 "enable-rc5" and "enable-mdc2", respectively, are specified.
10494 (IDEA remains enabled despite being patented. This is because IDEA
10495 is frequently required for interoperability, and there is no license
10496 fee for non-commercial use. As before, "no-idea" can be used to
10497 avoid this algorithm.)
10501 * Add processing of proxy certificates (see RFC 3820). This work was
10502 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10503 EGEE (Enabling Grids for E-science in Europe).
10507 * RC4 performance overhaul on modern architectures/implementations, such
10508 as Intel P4, IA-64 and AMD64.
10512 * New utility extract-section.pl. This can be used specify an alternative
10513 section number in a pod file instead of having to treat each file as
10514 a separate case in Makefile. This can be done by adding two lines to the
10517 =for comment openssl_section:XXX
10519 The blank line is mandatory.
10523 * New arguments -certform, -keyform and -pass for s_client and s_server
10524 to allow alternative format key and certificate files and passphrase
10529 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10530 update associated structures and add various utility functions.
10532 Add new policy related verify parameters, include policy checking in
10533 standard verify code. Enhance 'smime' application with extra parameters
10534 to support policy checking and print out.
10538 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10539 Nehemiah processors. These extensions support AES encryption in hardware
10540 as well as RNG (though RNG support is currently disabled).
10542 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10544 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10548 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10550 *Andy Polyakov and a number of other people*
10552 * Improved PowerPC platform support. Most notably BIGNUM assembler
10553 implementation contributed by IBM.
10555 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10557 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10558 exponent rather than 'unsigned long'. There is a corresponding change to
10559 the new 'rsa_keygen' element of the RSA_METHOD structure.
10561 *Jelte Jansen, Geoff Thorpe*
10563 * Functionality for creating the initial serial number file is now
10564 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10566 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10567 number file to 1, which is bound to cause problems. To avoid
10568 the problems while respecting compatibility between different 0.9.7
10569 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10570 CA.pl for serial number initialization. With the new release 0.9.8,
10571 we can fix the problem directly in the 'ca' utility.)
10575 * Reduced header interdependencies by declaring more opaque objects in
10576 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10577 give fewer recursive includes, which could break lazy source code - so
10578 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10579 developers should define this symbol when building and using openssl to
10580 ensure they track the recommended behaviour, interfaces, [etc], but
10581 backwards-compatible behaviour prevails when this isn't defined.
10585 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10589 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10590 This will generate a random key of the appropriate length based on the
10591 cipher context. The EVP_CIPHER can provide its own random key generation
10592 routine to support keys of a specific form. This is used in the des and
10593 3des routines to generate a key of the correct parity. Update S/MIME
10594 code to use new functions and hence generate correct parity DES keys.
10595 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10596 valid (weak or incorrect parity).
10600 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10601 as looking them up. This is useful when the verified structure may contain
10602 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10603 present unless the new PKCS7_NO_CRL flag is asserted.
10607 * Extend ASN1 oid configuration module. It now additionally accepts the
10610 shortName = some long name, 1.2.3.4
10614 * Reimplemented the BN_CTX implementation. There is now no more static
10615 limitation on the number of variables it can handle nor the depth of the
10616 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10617 information can now expand as required, and rather than having a single
10618 static array of bignums, BN_CTX now uses a linked-list of such arrays
10619 allowing it to expand on demand whilst maintaining the usefulness of
10620 BN_CTX's "bundling".
10624 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10625 to allow all RSA operations to function using a single BN_CTX.
10629 * Preliminary support for certificate policy evaluation and checking. This
10630 is initially intended to pass the tests outlined in "Conformance Testing
10631 of Relying Party Client Certificate Path Processing Logic" v1.07.
10635 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10636 remained unused and not that useful. A variety of other little bignum
10637 tweaks and fixes have also been made continuing on from the audit (see
10642 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10643 associated ASN1, EVP and SSL functions and old ASN1 macros.
10647 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10648 and this should never fail. So the return value from the use of
10649 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10650 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10654 * BN_CTX_get() should return zero-valued bignums, providing the same
10655 initialised value as BN_new().
10657 *Geoff Thorpe, suggested by Ulf Möller*
10659 * Support for inhibitAnyPolicy certificate extension.
10663 * An audit of the BIGNUM code is underway, for which debugging code is
10664 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10665 is considered valid when processing BIGNUMs, and causes execution to
10666 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10667 further steps are taken to deliberately pollute unused data in BIGNUM
10668 structures to try and expose faulty code further on. For now, openssl will
10669 (in its default mode of operation) continue to tolerate the inconsistent
10670 forms that it has tolerated in the past, but authors and packagers should
10671 consider trying openssl and their own applications when compiled with
10672 these debugging symbols defined. It will help highlight potential bugs in
10673 their own code, and will improve the test coverage for OpenSSL itself. At
10674 some point, these tighter rules will become openssl's default to improve
10675 maintainability, though the assert()s and other overheads will remain only
10676 in debugging configurations. See bn.h for more details.
10678 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10680 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10681 that can only be obtained through BN_CTX_new() (which implicitly
10682 initialises it). The presence of this function only made it possible
10683 to overwrite an existing structure (and cause memory leaks).
10687 * Because of the callback-based approach for implementing LHASH as a
10688 template type, lh_insert() adds opaque objects to hash-tables and
10689 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10690 to clean up those corresponding objects before destroying the hash table
10691 (and losing the object pointers). So some over-zealous constifications in
10692 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10693 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10694 prototyped to have "const" restrictions on the object pointers they are
10695 given (and so aren't required to cast them away any more).
10699 * The tmdiff.h API was so ugly and minimal that our own timing utility
10700 (speed) prefers to use its own implementation. The two implementations
10701 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10702 its object type properly exposed (MS_TM) instead of casting to/from
10703 `char *`. This may still change yet if someone realises MS_TM and
10705 aren't necessarily the greatest nomenclatures - but this is what was used
10706 internally to the implementation so I've used that for now.
10710 * Ensure that deprecated functions do not get compiled when
10711 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10712 the self-tests were still using deprecated key-generation functions so
10713 these have been updated also.
10717 * Reorganise PKCS#7 code to separate the digest location functionality
10718 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10719 New function PKCS7_set_digest() to set the digest type for PKCS#7
10720 digestedData type. Add additional code to correctly generate the
10721 digestedData type and add support for this type in PKCS7 initialization
10726 * New function PKCS7_set0_type_other() this initializes a PKCS7
10727 structure of type "other".
10731 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10732 sure the loop does correctly stop and breaking ("division by zero")
10733 modulus operations are not performed. The (pre-generated) prime
10734 table crypto/bn/bn_prime.h was already correct, but it could not be
10735 re-generated on some platforms because of the "division by zero"
10736 situation in the script.
10738 *Ralf S. Engelschall*
10740 * Update support for ECC-based TLS ciphersuites according to
10741 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10742 SHA-1 now is only used for "small" curves (where the
10743 representation of a field element takes up to 24 bytes); for
10744 larger curves, the field element resulting from ECDH is directly
10745 used as premaster secret.
10747 *Douglas Stebila (Sun Microsystems Laboratories)*
10749 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10750 curve secp160r1 to the tests.
10752 *Douglas Stebila (Sun Microsystems Laboratories)*
10754 * Add the possibility to load symbols globally with DSO.
10756 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10758 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10759 control of the error stack.
10763 * Add support for STORE in ENGINE.
10767 * Add the STORE type. The intention is to provide a common interface
10768 to certificate and key stores, be they simple file-based stores, or
10769 HSM-type store, or LDAP stores, or...
10770 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10774 * Add a generic structure called OPENSSL_ITEM. This can be used to
10775 pass a list of arguments to any function as well as provide a way
10776 for a function to pass data back to the caller.
10780 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10781 works like BUF_strdup() but can be used to duplicate a portion of
10782 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10787 * Add the function sk_find_ex() which works like sk_find(), but will
10788 return an index to an element even if an exact match couldn't be
10789 found. The index is guaranteed to point at the element where the
10790 searched-for key would be inserted to preserve sorting order.
10794 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10795 takes an extra flags argument for optional functionality. Currently,
10796 the following flags are defined:
10798 OBJ_BSEARCH_VALUE_ON_NOMATCH
10799 This one gets OBJ_bsearch_ex() to return a pointer to the first
10800 element where the comparing function returns a negative or zero
10803 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10804 This one gets OBJ_bsearch_ex() to return a pointer to the first
10805 element where the comparing function returns zero. This is useful
10806 if there are more than one element where the comparing function
10811 * Make it possible to create self-signed certificates with 'openssl ca'
10812 in such a way that the self-signed certificate becomes part of the
10813 CA database and uses the same mechanisms for serial number generation
10814 as all other certificate signing. The new flag '-selfsign' enables
10815 this functionality. Adapt CA.sh and CA.pl.in.
10819 * Add functionality to check the public key of a certificate request
10820 against a given private. This is useful to check that a certificate
10821 request can be signed by that key (self-signing).
10825 * Make it possible to have multiple active certificates with the same
10826 subject in the CA index file. This is done only if the keyword
10827 'unique_subject' is set to 'no' in the main CA section (default
10828 if 'CA_default') of the configuration file. The value is saved
10829 with the database itself in a separate index attribute file,
10830 named like the index file with '.attr' appended to the name.
10834 * Generate multi-valued AVAs using '+' notation in config files for
10839 * Support for nameConstraints certificate extension.
10843 * Support for policyConstraints certificate extension.
10847 * Support for policyMappings certificate extension.
10851 * Make sure the default DSA_METHOD implementation only uses its
10852 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10853 and change its own handlers to be NULL so as to remove unnecessary
10854 indirection. This lets alternative implementations fallback to the
10855 default implementation more easily.
10859 * Support for directoryName in GeneralName related extensions
10864 * Make it possible to link applications using Makefile.shared.
10865 Make that possible even when linking against static libraries!
10869 * Support for single pass processing for S/MIME signing. This now
10870 means that S/MIME signing can be done from a pipe, in addition
10871 cleartext signing (multipart/signed type) is effectively streaming
10872 and the signed data does not need to be all held in memory.
10874 This is done with a new flag PKCS7_STREAM. When this flag is set
10875 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10876 is done after the data is output (and digests calculated) in
10877 SMIME_write_PKCS7().
10881 * Add full support for -rpath/-R, both in shared libraries and
10882 applications, at least on the platforms where it's known how
10887 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10888 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10889 will now compute a table of multiples of the generator that
10890 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10891 faster (notably in the case of a single point multiplication,
10892 scalar * generator).
10894 *Nils Larsch, Bodo Moeller*
10896 * IPv6 support for certificate extensions. The various extensions
10897 which use the IP:a.b.c.d can now take IPv6 addresses using the
10898 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10903 * Added an ENGINE that implements RSA by performing private key
10904 exponentiations with the GMP library. The conversions to and from
10905 GMP's mpz_t format aren't optimised nor are any montgomery forms
10906 cached, and on x86 it appears OpenSSL's own performance has caught up.
10907 However there are likely to be other architectures where GMP could
10908 provide a boost. This ENGINE is not built in by default, but it can be
10909 specified at Configure time and should be accompanied by the necessary
10910 linker additions, eg;
10911 ./config -DOPENSSL_USE_GMP -lgmp
10915 * "openssl engine" will not display ENGINE/DSO load failure errors when
10916 testing availability of engines with "-t" - the old behaviour is
10917 produced by increasing the feature's verbosity with "-tt".
10921 * ECDSA routines: under certain error conditions uninitialized BN objects
10922 could be freed. Solution: make sure initialization is performed early
10923 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10928 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10929 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10930 software implementations. For DSA and DH, parameter generation can
10931 also be overridden by providing the appropriate method callbacks.
10935 * Change the "progress" mechanism used in key-generation and
10936 primality testing to functions that take a new BN_GENCB pointer in
10937 place of callback/argument pairs. The new API functions have `_ex`
10938 postfixes and the older functions are reimplemented as wrappers for
10939 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10940 declarations of the old functions to help (graceful) attempts to
10941 migrate to the new functions. Also, the new key-generation API
10942 functions operate on a caller-supplied key-structure and return
10943 success/failure rather than returning a key or NULL - this is to
10944 help make "keygen" another member function of RSA_METHOD etc.
10946 Example for using the new callback interface:
10948 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10949 void *my_arg = ...;
10952 BN_GENCB_set(&my_cb, my_callback, my_arg);
10954 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10955 /* For the meaning of a, b in calls to my_callback(), see the
10956 * documentation of the function that calls the callback.
10957 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10958 * my_callback should return 1 if it wants BN_is_prime_ex()
10959 * to continue, or 0 to stop.
10964 * Change the ZLIB compression method to be stateful, and make it
10965 available to TLS with the number defined in
10966 draft-ietf-tls-compression-04.txt.
10970 * Add the ASN.1 structures and functions for CertificatePair, which
10971 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10973 CertificatePair ::= SEQUENCE {
10974 forward [0] Certificate OPTIONAL,
10975 reverse [1] Certificate OPTIONAL,
10976 -- at least one of the pair shall be present -- }
10978 Also implement the PEM functions to read and write certificate
10979 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10981 This needed to be defined, mostly for the sake of the LDAP
10982 attribute crossCertificatePair, but may prove useful elsewhere as
10987 * Make it possible to inhibit symlinking of shared libraries in
10988 Makefile.shared, for Cygwin's sake.
10992 * Extend the BIGNUM API by creating a function
10993 void BN_set_negative(BIGNUM *a, int neg);
10994 and a macro that behave like
10995 int BN_is_negative(const BIGNUM *a);
10997 to avoid the need to access 'a->neg' directly in applications.
11001 * Implement fast modular reduction for pseudo-Mersenne primes
11002 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11003 EC_GROUP_new_curve_GFp() will now automatically use this
11006 *Nils Larsch <nla@trustcenter.de>*
11008 * Add new lock type (CRYPTO_LOCK_BN).
11012 * Change the ENGINE framework to automatically load engines
11013 dynamically from specific directories unless they could be
11014 found to already be built in or loaded. Move all the
11015 current engines except for the cryptodev one to a new
11016 directory engines/.
11017 The engines in engines/ are built as shared libraries if
11018 the "shared" options was given to ./Configure or ./config.
11019 Otherwise, they are inserted in libcrypto.a.
11020 /usr/local/ssl/engines is the default directory for dynamic
11021 engines, but that can be overridden at configure time through
11022 the usual use of --prefix and/or --openssldir, and at run
11023 time with the environment variable OPENSSL_ENGINES.
11025 *Geoff Thorpe and Richard Levitte*
11027 * Add Makefile.shared, a helper makefile to build shared
11028 libraries. Adapt Makefile.org.
11032 * Add version info to Win32 DLLs.
11034 *Peter 'Luna' Runestig" <peter@runestig.com>*
11036 * Add new 'medium level' PKCS#12 API. Certificates and keys
11037 can be added using this API to created arbitrary PKCS#12
11038 files while avoiding the low level API.
11040 New options to PKCS12_create(), key or cert can be NULL and
11041 will then be omitted from the output file. The encryption
11042 algorithm NIDs can be set to -1 for no encryption, the mac
11043 iteration count can be set to 0 to omit the mac.
11045 Enhance pkcs12 utility by making the -nokeys and -nocerts
11046 options work when creating a PKCS#12 file. New option -nomac
11047 to omit the mac, NONE can be set for an encryption algorithm.
11048 New code is modified to use the enhanced PKCS12_create()
11049 instead of the low level API.
11053 * Extend ASN1 encoder to support indefinite length constructed
11054 encoding. This can output sequences tags and octet strings in
11055 this form. Modify pk7_asn1.c to support indefinite length
11056 encoding. This is experimental and needs additional code to
11057 be useful, such as an ASN1 bio and some enhanced streaming
11060 Extend template encode functionality so that tagging is passed
11061 down to the template encoder.
11065 * Let 'openssl req' fail if an argument to '-newkey' is not
11066 recognized instead of using RSA as a default.
11070 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11071 As these are not official, they are not included in "ALL";
11072 the "ECCdraft" ciphersuite group alias can be used to select them.
11074 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11076 * Add ECDH engine support.
11078 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11080 * Add ECDH in new directory crypto/ecdh/.
11082 *Douglas Stebila (Sun Microsystems Laboratories)*
11084 * Let BN_rand_range() abort with an error after 100 iterations
11085 without success (which indicates a broken PRNG).
11089 * Change BN_mod_sqrt() so that it verifies that the input value
11090 is really the square of the return value. (Previously,
11091 BN_mod_sqrt would show GIGO behaviour.)
11095 * Add named elliptic curves over binary fields from X9.62, SECG,
11096 and WAP/WTLS; add OIDs that were still missing.
11098 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11100 * Extend the EC library for elliptic curves over binary fields
11101 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11104 EC_GF2m_simple_method
11108 EC_GROUP_new_curve_GF2m
11109 EC_GROUP_set_curve_GF2m
11110 EC_GROUP_get_curve_GF2m
11111 EC_POINT_set_affine_coordinates_GF2m
11112 EC_POINT_get_affine_coordinates_GF2m
11113 EC_POINT_set_compressed_coordinates_GF2m
11115 Point compression for binary fields is disabled by default for
11116 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11119 As binary polynomials are represented as BIGNUMs, various members
11120 of the EC_GROUP and EC_POINT data structures can be shared
11121 between the implementations for prime fields and binary fields;
11122 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11123 are essentially identical to their `..._GFp` counterparts.
11124 (For simplicity, the `..._GFp` prefix has been dropped from
11125 various internal method names.)
11127 An internal 'field_div' method (similar to 'field_mul' and
11128 'field_sqr') has been added; this is used only for binary fields.
11130 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11132 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11133 through methods ('mul', 'precompute_mult').
11135 The generic implementations (now internally called 'ec_wNAF_mul'
11136 and 'ec_wNAF_precomputed_mult') remain the default if these
11137 methods are undefined.
11139 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11141 * New function EC_GROUP_get_degree, which is defined through
11142 EC_METHOD. For curves over prime fields, this returns the bit
11143 length of the modulus.
11145 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11147 * New functions EC_GROUP_dup, EC_POINT_dup.
11148 (These simply call ..._new and ..._copy).
11150 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11152 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11153 Polynomials are represented as BIGNUMs (where the sign bit is not
11154 used) in the following functions [macros]:
11157 BN_GF2m_sub [= BN_GF2m_add]
11158 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11159 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11160 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11162 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11163 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11164 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11165 BN_GF2m_cmp [= BN_ucmp]
11167 (Note that only the 'mod' functions are actually for fields GF(2^m).
11168 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11170 For some functions, an the irreducible polynomial defining a
11171 field can be given as an 'unsigned int[]' with strictly
11172 decreasing elements giving the indices of those bits that are set;
11173 i.e., p[] represents the polynomial
11174 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11176 p[0] > p[1] > ... > p[k] = 0.
11177 This applies to the following functions:
11180 BN_GF2m_mod_mul_arr
11181 BN_GF2m_mod_sqr_arr
11182 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11183 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11184 BN_GF2m_mod_exp_arr
11185 BN_GF2m_mod_sqrt_arr
11186 BN_GF2m_mod_solve_quad_arr
11190 Conversion can be performed by the following functions:
11195 bntest.c has additional tests for binary polynomial arithmetic.
11197 Two implementations for BN_GF2m_mod_div() are available.
11198 The default algorithm simply uses BN_GF2m_mod_inv() and
11199 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11200 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11201 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11203 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11205 * Add new error code 'ERR_R_DISABLED' that can be used when some
11206 functionality is disabled at compile-time.
11208 *Douglas Stebila <douglas.stebila@sun.com>*
11210 * Change default behaviour of 'openssl asn1parse' so that more
11211 information is visible when viewing, e.g., a certificate:
11213 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11214 mode the content of non-printable OCTET STRINGs is output in a
11215 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11216 avoid the appearance of a printable string.
11218 *Nils Larsch <nla@trustcenter.de>*
11220 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11222 EC_GROUP_set_asn1_flag()
11223 EC_GROUP_get_asn1_flag()
11224 EC_GROUP_set_point_conversion_form()
11225 EC_GROUP_get_point_conversion_form()
11226 These control ASN1 encoding details:
11227 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11228 has been set to OPENSSL_EC_NAMED_CURVE.
11229 - Points are encoded in uncompressed form by default; options for
11230 asn1_for are as for point2oct, namely
11231 POINT_CONVERSION_COMPRESSED
11232 POINT_CONVERSION_UNCOMPRESSED
11233 POINT_CONVERSION_HYBRID
11235 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11237 EC_GROUP_set_seed()
11238 EC_GROUP_get0_seed()
11239 EC_GROUP_get_seed_len()
11240 This is used only for ASN1 purposes (so far).
11242 *Nils Larsch <nla@trustcenter.de>*
11244 * Add 'field_type' member to EC_METHOD, which holds the NID
11245 of the appropriate field type OID. The new function
11246 EC_METHOD_get_field_type() returns this value.
11248 *Nils Larsch <nla@trustcenter.de>*
11251 EC_POINT_point2bn()
11252 EC_POINT_bn2point()
11253 EC_POINT_point2hex()
11254 EC_POINT_hex2point()
11255 providing useful interfaces to EC_POINT_point2oct() and
11256 EC_POINT_oct2point().
11258 *Nils Larsch <nla@trustcenter.de>*
11260 * Change internals of the EC library so that the functions
11261 EC_GROUP_set_generator()
11262 EC_GROUP_get_generator()
11263 EC_GROUP_get_order()
11264 EC_GROUP_get_cofactor()
11265 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11266 to methods, which would lead to unnecessary code duplication when
11267 adding different types of curves.
11269 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11271 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11272 arithmetic, and such that modified wNAFs are generated
11273 (which avoid length expansion in many cases).
11277 * Add a function EC_GROUP_check_discriminant() (defined via
11278 EC_METHOD) that verifies that the curve discriminant is non-zero.
11280 Add a function EC_GROUP_check() that makes some sanity tests
11281 on a EC_GROUP, its generator and order. This includes
11282 EC_GROUP_check_discriminant().
11284 *Nils Larsch <nla@trustcenter.de>*
11286 * Add ECDSA in new directory crypto/ecdsa/.
11288 Add applications 'openssl ecparam' and 'openssl ecdsa'
11289 (these are based on 'openssl dsaparam' and 'openssl dsa').
11291 ECDSA support is also included in various other files across the
11292 library. Most notably,
11293 - 'openssl req' now has a '-newkey ecdsa:file' option;
11294 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11295 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11296 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11297 them suitable for ECDSA where domain parameters must be
11298 extracted before the specific public key;
11299 - ECDSA engine support has been added.
11301 *Nils Larsch <nla@trustcenter.de>*
11303 * Include some named elliptic curves, and add OIDs from X9.62,
11304 SECG, and WAP/WTLS. Each curve can be obtained from the new
11306 EC_GROUP_new_by_curve_name(),
11307 and the list of available named curves can be obtained with
11308 EC_get_builtin_curves().
11309 Also add a 'curve_name' member to EC_GROUP objects, which can be
11311 EC_GROUP_set_curve_name()
11312 EC_GROUP_get_curve_name()
11314 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11316 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11317 was actually never needed) and in BN_mul(). The removal in BN_mul()
11318 required a small change in bn_mul_part_recursive() and the addition
11319 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11320 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11321 bn_sub_words() and bn_add_words() except they take arrays with
11326 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11328 * Cleanse PEM buffers before freeing them since they may contain
11331 *Benjamin Bennett <ben@psc.edu>*
11333 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11334 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11335 authentication-only ciphersuites.
11339 * Since AES128 and AES256 share a single mask bit in the logic of
11340 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11341 kludge to work properly if AES128 is available and AES256 isn't.
11345 * Expand security boundary to match 1.1.1 module.
11349 * Remove redundant features: hash file source, editing of test vectors
11350 modify fipsld to use external fips_premain.c signature.
11354 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11355 run algorithm test programs.
11359 * Make algorithm test programs more tolerant of whitespace.
11363 * Have SSL/TLS server implementation tolerate "mismatched" record
11364 protocol version while receiving ClientHello even if the
11365 ClientHello is fragmented. (The server can't insist on the
11366 particular protocol version it has chosen before the ServerHello
11367 message has informed the client about his choice.)
11371 * Load error codes if they are not already present instead of using a
11372 static variable. This allows them to be cleanly unloaded and reloaded.
11376 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11378 * Introduce limits to prevent malicious keys being able to
11379 cause a denial of service. ([CVE-2006-2940])
11381 *Steve Henson, Bodo Moeller*
11383 * Fix ASN.1 parsing of certain invalid structures that can result
11384 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11386 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11387 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11389 * Fix SSL client code which could crash if connecting to a
11390 malicious SSLv2 server. ([CVE-2006-4343])
11392 *Tavis Ormandy and Will Drewry, Google Security Team*
11394 * Change ciphersuite string processing so that an explicit
11395 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11396 will no longer include "AES128-SHA"), and any other similar
11397 ciphersuite (same bitmap) from *other* protocol versions (so that
11398 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11399 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11400 changes from 0.9.8b and 0.9.8d.
11404 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11406 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11407 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11409 * Change the Unix randomness entropy gathering to use poll() when
11410 possible instead of select(), since the latter has some
11411 undesirable limitations.
11413 *Darryl Miles via Richard Levitte and Bodo Moeller*
11415 * Disable rogue ciphersuites:
11417 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11418 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11419 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11421 The latter two were purportedly from
11422 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11425 Also deactivate the remaining ciphersuites from
11426 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11427 unofficial, and the ID has long expired.
11431 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11432 dual-core machines) and other potential thread-safety issues.
11436 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11438 * Adapt fipsld and the build system to link against the validated FIPS
11439 module in FIPS mode.
11443 * Fixes for VC++ 2005 build under Windows.
11447 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11448 from a Windows bash shell such as MSYS. It is autodetected from the
11449 "config" script when run from a VC++ environment. Modify standard VC++
11450 build to use fipscanister.o from the GNU make build.
11454 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11456 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11457 The value now differs depending on if you build for FIPS or not.
11458 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11459 safely run with a non-FIPSed libcrypto, as it may crash because of
11460 the difference induced by this change.
11464 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11466 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11467 (part of SSL_OP_ALL). This option used to disable the
11468 countermeasure against man-in-the-middle protocol-version
11469 rollback in the SSL 2.0 server implementation, which is a bad
11470 idea. ([CVE-2005-2969])
11472 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11473 for Information Security, National Institute of Advanced Industrial
11474 Science and Technology [AIST, Japan)]*
11476 * Minimal support for X9.31 signatures and PSS padding modes. This is
11477 mainly for FIPS compliance and not fully integrated at this stage.
11481 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11482 the exponentiation using a fixed-length exponent. (Otherwise,
11483 the information leaked through timing could expose the secret key
11484 after many signatures; cf. Bleichenbacher's attack on DSA with
11489 * Make a new fixed-window mod_exp implementation the default for
11490 RSA, DSA, and DH private-key operations so that the sequence of
11491 squares and multiplies and the memory access pattern are
11492 independent of the particular secret key. This will mitigate
11493 cache-timing and potential related attacks.
11495 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11496 and this is automatically used by BN_mod_exp_mont() if the new flag
11497 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11498 will use this BN flag for private exponents unless the flag
11499 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11500 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11502 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11504 * Change the client implementation for SSLv23_method() and
11505 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11506 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11507 (Previously, the SSL 2.0 backwards compatible Client Hello
11508 message format would be used even with SSL_OP_NO_SSLv2.)
11512 * Add support for smime-type MIME parameter in S/MIME messages which some
11517 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11518 a threadsafe manner. Modify rsa code to use new function and add calls
11519 to dsa and dh code (which had race conditions before).
11523 * Include the fixed error library code in the C error file definitions
11524 instead of fixing them up at runtime. This keeps the error code
11525 structures constant.
11529 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11531 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11534 * Fixes for newer kerberos headers. NB: the casts are needed because
11535 the 'length' field is signed on one version and unsigned on another
11536 with no (?) obvious way to tell the difference, without these VC++
11537 complains. Also the "definition" of FAR (blank) is no longer included
11538 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11539 some needed definitions.
11543 * Undo Cygwin change.
11547 * Added support for proxy certificates according to RFC 3820.
11548 Because they may be a security thread to unaware applications,
11549 they must be explicitly allowed in run-time. See
11550 docs/HOWTO/proxy_certificates.txt for further information.
11554 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11556 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11557 server and client random values. Previously
11558 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11559 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11561 This change has negligible security impact because:
11563 1. Server and client random values still have 24 bytes of pseudo random
11566 2. Server and client random values are sent in the clear in the initial
11569 3. The master secret is derived using the premaster secret (48 bytes in
11570 size for static RSA ciphersuites) as well as client server and random
11573 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11576 *Stephen Henson, reported by UK NISCC*
11578 * Use Windows randomness collection on Cygwin.
11582 * Fix hang in EGD/PRNGD query when communication socket is closed
11583 prematurely by EGD/PRNGD.
11585 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11587 * Prompt for pass phrases when appropriate for PKCS12 input format.
11591 * Back-port of selected performance improvements from development
11592 branch, as well as improved support for PowerPC platforms.
11596 * Add lots of checks for memory allocation failure, error codes to indicate
11597 failure and freeing up memory if a failure occurs.
11599 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11601 * Add new -passin argument to dgst.
11605 * Perform some character comparisons of different types in X509_NAME_cmp:
11606 this is needed for some certificates that re-encode DNs into UTF8Strings
11607 (in violation of RFC3280) and can't or won't issue name rollover
11612 * Make an explicit check during certificate validation to see that
11613 the CA setting in each certificate on the chain is correct. As a
11614 side effect always do the following basic checks on extensions,
11615 not just when there's an associated purpose to the check:
11617 - if there is an unhandled critical extension (unless the user
11618 has chosen to ignore this fault)
11619 - if the path length has been exceeded (if one is set at all)
11620 - that certain extensions fit the associated purpose (if one has
11625 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11627 * Avoid a race condition when CRLs are checked in a multi threaded
11628 environment. This would happen due to the reordering of the revoked
11629 entries during signature checking and serial number lookup. Now the
11630 encoding is cached and the serial number sort performed under a lock.
11631 Add new STACK function sk_is_sorted().
11635 * Add Delta CRL to the extension code.
11639 * Various fixes to s3_pkt.c so alerts are sent properly.
11641 *David Holmes <d.holmes@f5.com>*
11643 * Reduce the chances of duplicate issuer name and serial numbers (in
11644 violation of RFC3280) using the OpenSSL certificate creation utilities.
11645 This is done by creating a random 64 bit value for the initial serial
11646 number when a serial number file is created or when a self signed
11647 certificate is created using 'openssl req -x509'. The initial serial
11648 number file is created using 'openssl x509 -next_serial' in CA.pl
11649 rather than being initialized to 1.
11653 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11655 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11656 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11658 *Joe Orton, Steve Henson*
11660 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11663 *Joe Orton, Steve Henson*
11665 * Make it possible to have multiple active certificates with the same
11666 subject in the CA index file. This is done only if the keyword
11667 'unique_subject' is set to 'no' in the main CA section (default
11668 if 'CA_default') of the configuration file. The value is saved
11669 with the database itself in a separate index attribute file,
11670 named like the index file with '.attr' appended to the name.
11674 * X509 verify fixes. Disable broken certificate workarounds when
11675 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11676 keyUsage extension present. Don't accept CRLs with unhandled critical
11677 extensions: since verify currently doesn't process CRL extensions this
11678 rejects a CRL with *any* critical extensions. Add new verify error codes
11683 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11684 A clarification of RFC2560 will require the use of OCTET STRINGs and
11685 some implementations cannot handle the current raw format. Since OpenSSL
11686 copies and compares OCSP nonces as opaque blobs without any attempt at
11687 parsing them this should not create any compatibility issues.
11691 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11692 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11693 this HMAC (and other) operations are several times slower than OpenSSL
11698 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11700 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11702 * Use the correct content when signing type "other".
11706 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11708 * Fix various bugs revealed by running the NISCC test suite:
11710 Stop out of bounds reads in the ASN1 code when presented with
11711 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11713 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11715 If verify callback ignores invalid public key errors don't try to check
11716 certificate signature with the NULL public key.
11720 * New -ignore_err option in ocsp application to stop the server
11721 exiting on the first error in a request.
11725 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11726 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11731 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11732 extra data after the compression methods not only for TLS 1.0
11733 but also for SSL 3.0 (as required by the specification).
11735 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11737 * Change X509_certificate_type() to mark the key as exported/exportable
11738 when it's 512 *bits* long, not 512 bytes.
11742 * Change AES_cbc_encrypt() so it outputs exact multiple of
11743 blocks during encryption.
11747 * Various fixes to base64 BIO and non blocking I/O. On write
11748 flushes were not handled properly if the BIO retried. On read
11749 data was not being buffered properly and had various logic bugs.
11750 This also affects blocking I/O when the data being decoded is a
11755 * Various S/MIME bugfixes and compatibility changes:
11756 output correct application/pkcs7 MIME type if
11757 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11758 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11759 of files as .eml work). Correctly handle very long lines in MIME
11764 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11766 * Countermeasure against the Klima-Pokorny-Rosa extension of
11767 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11768 a protocol version number mismatch like a decryption error
11769 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11773 * Turn on RSA blinding by default in the default implementation
11774 to avoid a timing attack. Applications that don't want it can call
11775 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11776 They would be ill-advised to do so in most cases.
11778 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11780 * Change RSA blinding code so that it works when the PRNG is not
11781 seeded (in this case, the secret RSA exponent is abused as
11782 an unpredictable seed -- if it is not unpredictable, there
11783 is no point in blinding anyway). Make RSA blinding thread-safe
11784 by remembering the creator's thread ID in rsa->blinding and
11785 having all other threads use local one-time blinding factors
11786 (this requires more computation than sharing rsa->blinding, but
11787 avoids excessive locking; and if an RSA object is not shared
11788 between threads, blinding will still be very fast).
11792 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11793 ENGINE as defaults for all supported algorithms irrespective of
11794 the 'flags' parameter. 'flags' is now honoured, so applications
11795 should make sure they are passing it correctly.
11799 * Target "mingw" now allows native Windows code to be generated in
11800 the Cygwin environment as well as with the MinGW compiler.
11804 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11806 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11807 via timing by performing a MAC computation even if incorrect
11808 block cipher padding has been found. This is a countermeasure
11809 against active attacks where the attacker has to distinguish
11810 between bad padding and a MAC verification error. ([CVE-2003-0078])
11812 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11813 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11814 Martin Vuagnoux (EPFL, Ilion)*
11816 * Make the no-err option work as intended. The intention with no-err
11817 is not to have the whole error stack handling routines removed from
11818 libcrypto, it's only intended to remove all the function name and
11819 reason texts, thereby removing some of the footprint that may not
11820 be interesting if those errors aren't displayed anyway.
11822 NOTE: it's still possible for any application or module to have its
11823 own set of error texts inserted. The routines are there, just not
11824 used by default when no-err is given.
11828 * Add support for FreeBSD on IA64.
11830 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11832 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11833 Kerberos function mit_des_cbc_cksum(). Before this change,
11834 the value returned by DES_cbc_cksum() was like the one from
11835 mit_des_cbc_cksum(), except the bytes were swapped.
11837 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11839 * Allow an application to disable the automatic SSL chain building.
11840 Before this a rather primitive chain build was always performed in
11841 ssl3_output_cert_chain(): an application had no way to send the
11842 correct chain if the automatic operation produced an incorrect result.
11844 Now the chain builder is disabled if either:
11846 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11848 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11850 The reasoning behind this is that an application would not want the
11851 auto chain building to take place if extra chain certificates are
11852 present and it might also want a means of sending no additional
11853 certificates (for example the chain has two certificates and the
11858 * Add the possibility to build without the ENGINE framework.
11860 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11862 * Under Win32 gmtime() can return NULL: check return value in
11863 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11867 * DSA routines: under certain error conditions uninitialized BN objects
11868 could be freed. Solution: make sure initialization is performed early
11869 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11870 Nils Larsch <nla@trustcenter.de> via PR#459)
11874 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11875 checked on reconnect on the client side, therefore session resumption
11876 could still fail with a "ssl session id is different" error. This
11877 behaviour is masked when SSL_OP_ALL is used due to
11878 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11879 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11880 followup to PR #377.
11884 * IA-32 assembler support enhancements: unified ELF targets, support
11885 for SCO/Caldera platforms, fix for Cygwin shared build.
11889 * Add support for FreeBSD on sparc64. As a consequence, support for
11890 FreeBSD on non-x86 processors is separate from x86 processors on
11891 the config script, much like the NetBSD support.
11893 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11895 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11897 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11900 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11901 code (06) was taken as the first octet of the session ID and the last
11902 octet was ignored consequently. As a result SSLv2 client side session
11903 caching could not have worked due to the session ID mismatch between
11905 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11910 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11911 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11916 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11917 seems that in spite of existing for more than a year, many application
11918 author have done nothing to provide the necessary callbacks, which
11919 means that this particular engine will not work properly anywhere.
11920 This is a very unfortunate situation which forces us, in the name
11921 of usability, to give the hw_ncipher.c a static lock, which is part
11923 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11924 appear in 0.9.8 or later. We EXPECT application authors to have
11925 dealt properly with this when 0.9.8 is released (unless we actually
11926 make such changes in the libcrypto locking code that changes will
11927 have to be made anyway).
11931 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11932 octets have been read, EOF or an error occurs. Without this change
11933 some truncated ASN1 structures will not produce an error.
11937 * Disable Heimdal support, since it hasn't been fully implemented.
11938 Still give the possibility to force the use of Heimdal, but with
11939 warnings and a request that patches get sent to openssl-dev.
11943 * Add the VC-CE target, introduce the WINCE sysname, and add
11944 INSTALL.WCE and appropriate conditionals to make it build.
11946 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11948 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11949 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11950 edit numbers of the version.
11952 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11954 * Introduce safe string copy and catenation functions
11955 (BUF_strlcpy() and BUF_strlcat()).
11957 *Ben Laurie (CHATS) and Richard Levitte*
11959 * Avoid using fixed-size buffers for one-line DNs.
11961 *Ben Laurie (CHATS)*
11963 * Add BUF_MEM_grow_clean() to avoid information leakage when
11964 resizing buffers containing secrets, and use where appropriate.
11966 *Ben Laurie (CHATS)*
11968 * Avoid using fixed size buffers for configuration file location.
11970 *Ben Laurie (CHATS)*
11972 * Avoid filename truncation for various CA files.
11974 *Ben Laurie (CHATS)*
11976 * Use sizeof in preference to magic numbers.
11978 *Ben Laurie (CHATS)*
11980 * Avoid filename truncation in cert requests.
11982 *Ben Laurie (CHATS)*
11984 * Add assertions to check for (supposedly impossible) buffer
11987 *Ben Laurie (CHATS)*
11989 * Don't cache truncated DNS entries in the local cache (this could
11990 potentially lead to a spoofing attack).
11992 *Ben Laurie (CHATS)*
11994 * Fix various buffers to be large enough for hex/decimal
11995 representations in a platform independent manner.
11997 *Ben Laurie (CHATS)*
11999 * Add CRYPTO_realloc_clean() to avoid information leakage when
12000 resizing buffers containing secrets, and use where appropriate.
12002 *Ben Laurie (CHATS)*
12004 * Add BIO_indent() to avoid much slightly worrying code to do
12007 *Ben Laurie (CHATS)*
12009 * Convert sprintf()/BIO_puts() to BIO_printf().
12011 *Ben Laurie (CHATS)*
12013 * buffer_gets() could terminate with the buffer only half
12016 *Ben Laurie (CHATS)*
12018 * Add assertions to prevent user-supplied crypto functions from
12019 overflowing internal buffers by having large block sizes, etc.
12021 *Ben Laurie (CHATS)*
12023 * New OPENSSL_assert() macro (similar to assert(), but enabled
12026 *Ben Laurie (CHATS)*
12028 * Eliminate unused copy of key in RC4.
12030 *Ben Laurie (CHATS)*
12032 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12034 *Ben Laurie (CHATS)*
12036 * Fix off-by-one error in EGD path.
12038 *Ben Laurie (CHATS)*
12040 * If RANDFILE path is too long, ignore instead of truncating.
12042 *Ben Laurie (CHATS)*
12044 * Eliminate unused and incorrectly sized X.509 structure
12047 *Ben Laurie (CHATS)*
12049 * Eliminate unused and dangerous function knumber().
12051 *Ben Laurie (CHATS)*
12053 * Eliminate unused and dangerous structure, KSSL_ERR.
12055 *Ben Laurie (CHATS)*
12057 * Protect against overlong session ID context length in an encoded
12058 session object. Since these are local, this does not appear to be
12061 *Ben Laurie (CHATS)*
12063 * Change from security patch (see 0.9.6e below) that did not affect
12064 the 0.9.6 release series:
12066 Remote buffer overflow in SSL3 protocol - an attacker could
12067 supply an oversized master key in Kerberos-enabled versions.
12070 *Ben Laurie (CHATS)*
12072 * Change the SSL kerb5 codes to match RFC 2712.
12076 * Make -nameopt work fully for req and add -reqopt switch.
12078 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12080 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12082 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12084 * Make sure tests can be performed even if the corresponding algorithms
12085 have been removed entirely. This was also the last step to make
12086 OpenSSL compilable with DJGPP under all reasonable conditions.
12088 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12090 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12091 to allow version independent disabling of normally unselected ciphers,
12092 which may be activated as a side-effect of selecting a single cipher.
12094 (E.g., cipher list string "RSA" enables ciphersuites that are left
12095 out of "ALL" because they do not provide symmetric encryption.
12096 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12098 *Lutz Jaenicke, Bodo Moeller*
12100 * Add appropriate support for separate platform-dependent build
12101 directories. The recommended way to make a platform-dependent
12102 build directory is the following (tested on Linux), maybe with
12105 # Place yourself outside of the OpenSSL source tree. In
12106 # this example, the environment variable OPENSSL_SOURCE
12107 # is assumed to contain the absolute OpenSSL source directory.
12108 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12109 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12110 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12111 mkdir -p `dirname $F`
12112 ln -s $OPENSSL_SOURCE/$F $F
12115 To be absolutely sure not to disturb the source tree, a "make clean"
12116 is a good thing. If it isn't successful, don't worry about it,
12117 it probably means the source directory is very clean.
12121 * Make sure any ENGINE control commands make local copies of string
12122 pointers passed to them whenever necessary. Otherwise it is possible
12123 the caller may have overwritten (or deallocated) the original string
12124 data when a later ENGINE operation tries to use the stored values.
12126 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12128 * Improve diagnostics in file reading and command-line digests.
12130 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12132 * Add AES modes CFB and OFB to the object database. Correct an
12133 error in AES-CFB decryption.
12137 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12138 allows existing EVP_CIPHER_CTX structures to be reused after
12139 calling `EVP_*Final()`. This behaviour is used by encryption
12140 BIOs and some applications. This has the side effect that
12141 applications must explicitly clean up cipher contexts with
12142 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12146 * Check the values of dna and dnb in bn_mul_recursive before calling
12147 bn_mul_comba (a non zero value means the a or b arrays do not contain
12148 n2 elements) and fallback to bn_mul_normal if either is not zero.
12152 * Fix escaping of non-ASCII characters when using the -subj option
12153 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12157 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12158 form for "surname", serialNumber has no short form.
12159 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12160 therefore remove "mail" short name for "internet 7".
12161 The OID for unique identifiers in X509 certificates is
12162 x500UniqueIdentifier, not uniqueIdentifier.
12163 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12167 * Add an "init" command to the ENGINE config module and auto initialize
12168 ENGINEs. Without any "init" command the ENGINE will be initialized
12169 after all ctrl commands have been executed on it. If init=1 the
12170 ENGINE is initialized at that point (ctrls before that point are run
12171 on the uninitialized ENGINE and after on the initialized one). If
12172 init=0 then the ENGINE will not be initialized at all.
12176 * Fix the 'app_verify_callback' interface so that the user-defined
12177 argument is actually passed to the callback: In the
12178 SSL_CTX_set_cert_verify_callback() prototype, the callback
12179 declaration has been changed from
12182 int (*cb)(X509_STORE_CTX *,void *);
12183 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12184 i=s->ctx->app_verify_callback(&ctx)
12185 has been changed into
12186 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12188 To update applications using SSL_CTX_set_cert_verify_callback(),
12189 a dummy argument can be added to their callback functions.
12191 *D. K. Smetters <smetters@parc.xerox.com>*
12193 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12195 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12197 * Add and OPENSSL_LOAD_CONF define which will cause
12198 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12199 This allows older applications to transparently support certain
12200 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12201 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12202 load the config file and OPENSSL_add_all_algorithms_conf() which will
12203 always load it have also been added.
12207 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12208 Adjust NIDs and EVP layer.
12210 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12212 * Config modules support in openssl utility.
12214 Most commands now load modules from the config file,
12215 though in a few (such as version) this isn't done
12216 because it couldn't be used for anything.
12218 In the case of ca and req the config file used is
12219 the same as the utility itself: that is the -config
12220 command line option can be used to specify an
12225 * Move default behaviour from OPENSSL_config(). If appname is NULL
12226 use "openssl_conf" if filename is NULL use default openssl config file.
12230 * Add an argument to OPENSSL_config() to allow the use of an alternative
12231 config section name. Add a new flag to tolerate a missing config file
12232 and move code to CONF_modules_load_file().
12236 * Support for crypto accelerator cards from Accelerated Encryption
12237 Processing, www.aep.ie. (Use engine 'aep')
12238 The support was copied from 0.9.6c [engine] and adapted/corrected
12239 to work with the new engine framework.
12241 *AEP Inc. and Richard Levitte*
12243 * Support for SureWare crypto accelerator cards from Baltimore
12244 Technologies. (Use engine 'sureware')
12245 The support was copied from 0.9.6c [engine] and adapted
12246 to work with the new engine framework.
12250 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12251 make the newer ENGINE framework commands for the CHIL engine work.
12253 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12255 * Make it possible to produce shared libraries on ReliantUNIX.
12257 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12259 * Add the configuration target debug-linux-ppro.
12260 Make 'openssl rsa' use the general key loading routines
12261 implemented in `apps.c`, and make those routines able to
12262 handle the key format FORMAT_NETSCAPE and the variant
12265 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12267 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12269 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12271 * Add -keyform to rsautl, and document -engine.
12273 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12275 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12276 BIO_R_NO_SUCH_FILE error code rather than the generic
12277 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12281 * Add new functions
12282 ERR_peek_last_error
12283 ERR_peek_last_error_line
12284 ERR_peek_last_error_line_data.
12285 These are similar to
12287 ERR_peek_error_line
12288 ERR_peek_error_line_data,
12289 but report on the latest error recorded rather than the first one
12290 still in the error queue.
12292 *Ben Laurie, Bodo Moeller*
12294 * default_algorithms option in ENGINE config module. This allows things
12296 default_algorithms = ALL
12297 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12301 * Preliminary ENGINE config module.
12305 * New experimental application configuration code.
12309 * Change the AES code to follow the same name structure as all other
12310 symmetric ciphers, and behave the same way. Move everything to
12311 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12313 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12315 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12317 *Ben Laurie and Theo de Raadt*
12319 * Add option to output public keys in req command.
12321 *Massimiliano Pala madwolf@openca.org*
12323 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12324 (up to about 10% better than before for P-192 and P-224).
12328 * New functions/macros
12330 SSL_CTX_set_msg_callback(ctx, cb)
12331 SSL_CTX_set_msg_callback_arg(ctx, arg)
12332 SSL_set_msg_callback(ssl, cb)
12333 SSL_set_msg_callback_arg(ssl, arg)
12335 to request calling a callback function
12337 void cb(int write_p, int version, int content_type,
12338 const void *buf, size_t len, SSL *ssl, void *arg)
12340 whenever a protocol message has been completely received
12341 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12342 protocol version according to which the SSL library interprets
12343 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12344 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12345 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12346 specification (change_cipher_spec(20), alert(21), handshake(22)).
12347 'buf' and 'len' point to the actual message, 'ssl' to the
12348 SSL object, and 'arg' is the application-defined value set by
12349 SSL[_CTX]_set_msg_callback_arg().
12351 'openssl s_client' and 'openssl s_server' have new '-msg' options
12352 to enable a callback that displays all protocol messages.
12356 * Change the shared library support so shared libraries are built as
12357 soon as the corresponding static library is finished, and thereby get
12358 openssl and the test programs linked against the shared library.
12359 This still only happens when the keyword "shard" has been given to
12360 the configuration scripts.
12362 NOTE: shared library support is still an experimental thing, and
12363 backward binary compatibility is still not guaranteed.
12365 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12367 * Add support for Subject Information Access extension.
12369 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12371 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12372 additional bytes when new memory had to be allocated, not just
12373 when reusing an existing buffer.
12377 * New command line and configuration option 'utf8' for the req command.
12378 This allows field values to be specified as UTF8 strings.
12382 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12383 runs for the former and machine-readable output for the latter.
12387 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12388 of the e-mail address in the DN (i.e., it will go into a certificate
12389 extension only). The new configuration file option 'email_in_dn = no'
12390 has the same effect.
12392 *Massimiliano Pala madwolf@openca.org*
12394 * Change all functions with names starting with `des_` to be starting
12395 with `DES_` instead. Add wrappers that are compatible with libdes,
12396 but are named `_ossl_old_des_*`. Finally, add macros that map the
12397 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12398 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12399 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12402 Since we provide two compatibility mappings, the user needs to
12403 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12404 compatibility is desired. The default (i.e., when that macro
12405 isn't defined) is OpenSSL 0.9.6c compatibility.
12407 There are also macros that enable and disable the support of old
12408 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12409 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12410 are defined, the default will apply: to support the old des routines.
12412 In either case, one must include openssl/des.h to get the correct
12413 definitions. Do not try to just include openssl/des_old.h, that
12416 NOTE: This is a major break of an old API into a new one. Software
12417 authors are encouraged to switch to the `DES_` style functions. Some
12418 time in the future, des_old.h and the libdes compatibility functions
12419 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12420 default), and then completely removed.
12424 * Test for certificates which contain unsupported critical extensions.
12425 If such a certificate is found during a verify operation it is
12426 rejected by default: this behaviour can be overridden by either
12427 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12428 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12429 X509_supported_extension() has also been added which returns 1 if a
12430 particular extension is supported.
12434 * Modify the behaviour of EVP cipher functions in similar way to digests
12435 to retain compatibility with existing code.
12439 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12440 compatibility with existing code. In particular the 'ctx' parameter does
12441 not have to be to be initialized before the call to EVP_DigestInit() and
12442 it is tidied up after a call to EVP_DigestFinal(). New function
12443 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12444 EVP_MD_CTX_copy() changed to not require the destination to be
12445 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12446 requires the destination to be valid.
12448 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12449 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12453 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12454 so that complete 'Handshake' protocol structures are kept in memory
12455 instead of overwriting 'msg_type' and 'length' with 'body' data.
12459 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12461 *Massimo Santin via Richard Levitte*
12463 * Major restructuring to the underlying ENGINE code. This includes
12464 reduction of linker bloat, separation of pure "ENGINE" manipulation
12465 (initialisation, etc) from functionality dealing with implementations
12466 of specific crypto interfaces. This change also introduces integrated
12467 support for symmetric ciphers and digest implementations - so ENGINEs
12468 can now accelerate these by providing EVP_CIPHER and EVP_MD
12469 implementations of their own. This is detailed in
12470 [crypto/engine/README.md](crypto/engine/README.md)
12471 as it couldn't be adequately described here. However, there are a few
12472 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12473 were changed in the original introduction of ENGINE code have now
12474 reverted back - the hooking from this code to ENGINE is now a good
12475 deal more passive and at run-time, operations deal directly with
12476 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12477 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12478 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12479 they were not being used by the framework as there is no concept of a
12480 BIGNUM_METHOD and they could not be generalised to the new
12481 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12482 ENGINE_cpy() has been removed as it cannot be consistently defined in
12487 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12491 * Change mkdef.pl to sort symbols that get the same entry number,
12492 and make sure the automatically generated functions `ERR_load_*`
12493 become part of libeay.num as well.
12497 * New function SSL_renegotiate_pending(). This returns true once
12498 renegotiation has been requested (either SSL_renegotiate() call
12499 or HelloRequest/ClientHello received from the peer) and becomes
12500 false once a handshake has been completed.
12501 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12502 sends a HelloRequest, but does not ensure that a handshake takes
12503 place. SSL_renegotiate_pending() is useful for checking if the
12504 client has followed the request.)
12508 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12509 By default, clients may request session resumption even during
12510 renegotiation (if session ID contexts permit); with this option,
12511 session resumption is possible only in the first handshake.
12513 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12514 more bits available for options that should not be part of
12515 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12519 * Add some demos for certificate and certificate request creation.
12523 * Make maximum certificate chain size accepted from the peer application
12524 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12525 "Douglas E. Engert" <deengert@anl.gov>.
12529 * Add support for shared libraries for Unixware-7
12530 (Boyd Lynn Gerber <gerberb@zenez.com>).
12534 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12535 be done prior to destruction. Use this to unload error strings from
12536 ENGINEs that load their own error strings. NB: This adds two new API
12537 functions to "get" and "set" this destroy handler in an ENGINE.
12541 * Alter all existing ENGINE implementations (except "openssl" and
12542 "openbsd") to dynamically instantiate their own error strings. This
12543 makes them more flexible to be built both as statically-linked ENGINEs
12544 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12545 Also, add stub code to each that makes building them as self-contained
12546 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12550 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12551 implementations into applications that are completely implemented in
12552 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12553 commands that can be used to configure what shared-library to load and
12554 to control aspects of the way it is handled. Also, made an update to
12555 the [README-Engine.md](README-Engine.md) file
12556 that brings its information up-to-date and
12557 provides some information and instructions on the "dynamic" ENGINE
12558 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12562 * Make it possible to unload ranges of ERR strings with a new
12563 "ERR_unload_strings" function.
12567 * Add a copy() function to EVP_MD.
12571 * Make EVP_MD routines take a context pointer instead of just the
12572 md_data void pointer.
12576 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12577 that the digest can only process a single chunk of data
12578 (typically because it is provided by a piece of
12579 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12580 is only going to provide a single chunk of data, and hence the
12581 framework needn't accumulate the data for oneshot drivers.
12585 * As with "ERR", make it possible to replace the underlying "ex_data"
12586 functions. This change also alters the storage and management of global
12587 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12588 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12589 index counters. The API functions that use this state have been changed
12590 to take a "class_index" rather than pointers to the class's local STACK
12591 and counter, and there is now an API function to dynamically create new
12592 classes. This centralisation allows us to (a) plug a lot of the
12593 thread-safety problems that existed, and (b) makes it possible to clean
12594 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12595 such data would previously have always leaked in application code and
12596 workarounds were in place to make the memory debugging turn a blind eye
12597 to it. Application code that doesn't use this new function will still
12598 leak as before, but their memory debugging output will announce it now
12599 rather than letting it slide.
12601 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12602 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12603 has a return value to indicate success or failure.
12607 * Make it possible to replace the underlying "ERR" functions such that the
12608 global state (2 LHASH tables and 2 locks) is only used by the "default"
12609 implementation. This change also adds two functions to "get" and "set"
12610 the implementation prior to it being automatically set the first time
12611 any other ERR function takes place. Ie. an application can call "get",
12612 pass the return value to a module it has just loaded, and that module
12613 can call its own "set" function using that value. This means the
12614 module's "ERR" operations will use (and modify) the error state in the
12615 application and not in its own statically linked copy of OpenSSL code.
12619 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12620 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12621 the operation, and provides a more encapsulated way for external code
12622 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12623 to use these functions rather than manually incrementing the counts.
12625 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12629 * Add EVP test program.
12633 * Add symmetric cipher support to ENGINE. Expect the API to change!
12637 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12638 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12639 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12640 These allow a CRL to be built without having to access X509_CRL fields
12641 directly. Modify 'ca' application to use new functions.
12645 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12646 bug workarounds. Rollback attack detection is a security feature.
12647 The problem will only arise on OpenSSL servers when TLSv1 is not
12648 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12649 Software authors not wanting to support TLSv1 will have special reasons
12650 for their choice and can explicitly enable this option.
12652 *Bodo Moeller, Lutz Jaenicke*
12654 * Rationalise EVP so it can be extended: don't include a union of
12655 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12656 (similar to those existing for EVP_CIPHER_CTX).
12661 EVP_MD_CTX_init(&md); /* new function call */
12662 EVP_DigestInit(&md, EVP_sha1());
12663 EVP_DigestUpdate(&md, in, len);
12664 EVP_DigestFinal(&md, out, NULL);
12665 EVP_MD_CTX_cleanup(&md); /* new function call */
12669 * Make DES key schedule conform to the usual scheme, as well as
12670 correcting its structure. This means that calls to DES functions
12671 now have to pass a pointer to a des_key_schedule instead of a
12672 plain des_key_schedule (which was actually always a pointer
12675 des_key_schedule ks;
12677 des_set_key_checked(..., &ks);
12678 des_ncbc_encrypt(..., &ks, ...);
12680 (Note that a later change renames 'des_...' into 'DES_...'.)
12684 * Initial reduction of linker bloat: the use of some functions, such as
12685 PEM causes large amounts of unused functions to be linked in due to
12686 poor organisation. For example pem_all.c contains every PEM function
12687 which has a knock on effect of linking in large amounts of (unused)
12688 ASN1 code. Grouping together similar functions and splitting unrelated
12689 functions prevents this.
12693 * Cleanup of EVP macros.
12697 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12698 correct `_ecb suffix`.
12702 * Add initial OCSP responder support to ocsp application. The
12703 revocation information is handled using the text based index
12704 use by the ca application. The responder can either handle
12705 requests generated internally, supplied in files (for example
12706 via a CGI script) or using an internal minimal server.
12710 * Add configuration choices to get zlib compression for TLS.
12714 * Changes to Kerberos SSL for RFC 2712 compliance:
12715 1. Implemented real KerberosWrapper, instead of just using
12716 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12717 2. Implemented optional authenticator field of KerberosWrapper.
12719 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12720 and authenticator structs; see crypto/krb5/.
12722 Generalized Kerberos calls to support multiple Kerberos libraries.
12723 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12724 via Richard Levitte*
12726 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12727 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12728 values for each of the key sizes rather than having just
12729 parameters (and 'speed' generating keys each time).
12733 * Speed up EVP routines.
12736 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12737 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12738 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12739 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12741 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12742 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12743 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12746 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12748 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12752 * Added the OS2-EMX target.
12754 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12756 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12757 New functions to support `NCONF `routines in extension code.
12758 New function `CONF_set_nconf()`
12759 to allow functions which take an `NCONF` to also handle the old `LHASH`
12760 structure: this means that the old `CONF` compatible routines can be
12761 retained (in particular w.rt. extensions) without having to duplicate the
12762 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12766 * Enhance the general user interface with mechanisms for inner control
12767 and with possibilities to have yes/no kind of prompts.
12771 * Change all calls to low level digest routines in the library and
12772 applications to use EVP. Add missing calls to HMAC_cleanup() and
12773 don't assume HMAC_CTX can be copied using memcpy().
12775 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12777 * Add the possibility to control engines through control names but with
12778 arbitrary arguments instead of just a string.
12779 Change the key loaders to take a UI_METHOD instead of a callback
12780 function pointer. NOTE: this breaks binary compatibility with earlier
12781 versions of OpenSSL [engine].
12782 Adapt the nCipher code for these new conditions and add a card insertion
12787 * Enhance the general user interface with mechanisms to better support
12788 dialog box interfaces, application-defined prompts, the possibility
12789 to use defaults (for example default passwords from somewhere else)
12790 and interrupts/cancellations.
12794 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12795 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12799 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12800 tidy up some unnecessarily weird code in 'sk_new()').
12802 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12804 * Change the key loading routines for ENGINEs to use the same kind
12805 callback (pem_password_cb) as all other routines that need this
12810 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12811 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12812 than this minimum value is recommended.
12816 * New random seeder for OpenVMS, using the system process statistics
12817 that are easily reachable.
12821 * Windows apparently can't transparently handle global
12822 variables defined in DLLs. Initialisations such as:
12824 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12826 won't compile. This is used by the any applications that need to
12827 declare their own ASN1 modules. This was fixed by adding the option
12828 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12829 needed for static libraries under Win32.
12833 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12834 setting of purpose and trust fields. New X509_STORE trust and
12835 purpose functions and tidy up setting in other SSL functions.
12839 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12840 structure. These are inherited by X509_STORE_CTX when it is
12841 initialised. This allows various defaults to be set in the
12842 X509_STORE structure (such as flags for CRL checking and custom
12843 purpose or trust settings) for functions which only use X509_STORE_CTX
12844 internally such as S/MIME.
12846 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12847 trust settings if they are not set in X509_STORE. This allows X509_STORE
12848 purposes and trust (in S/MIME for example) to override any set by default.
12850 Add command line options for CRL checking to smime, s_client and s_server
12855 * Initial CRL based revocation checking. If the CRL checking flag(s)
12856 are set then the CRL is looked up in the X509_STORE structure and
12857 its validity and signature checked, then if the certificate is found
12858 in the CRL the verify fails with a revoked error.
12860 Various new CRL related callbacks added to X509_STORE_CTX structure.
12862 Command line options added to 'verify' application to support this.
12864 This needs some additional work, such as being able to handle multiple
12865 CRLs with different times, extension based lookup (rather than just
12866 by subject name) and ultimately more complete V2 CRL extension
12871 * Add a general user interface API (crypto/ui/). This is designed
12872 to replace things like des_read_password and friends (backward
12873 compatibility functions using this new API are provided).
12874 The purpose is to remove prompting functions from the DES code
12875 section as well as provide for prompting through dialog boxes in
12876 a window system and the like.
12880 * Add "ex_data" support to ENGINE so implementations can add state at a
12881 per-structure level rather than having to store it globally.
12885 * Make it possible for ENGINE structures to be copied when retrieved by
12886 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12887 This causes the "original" ENGINE structure to act like a template,
12888 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12889 operational state can be localised to each ENGINE structure, despite the
12890 fact they all share the same "methods". New ENGINE structures returned in
12891 this case have no functional references and the return value is the single
12892 structural reference. This matches the single structural reference returned
12893 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12898 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12899 needs to match any other type at all we need to manually clear the
12904 * Changes to the "openssl engine" utility to include;
12905 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12906 about an ENGINE's available control commands.
12907 - executing control commands from command line arguments using the
12908 '-pre' and '-post' switches. '-post' is only used if '-t' is
12909 specified and the ENGINE is successfully initialised. The syntax for
12910 the individual commands are colon-separated, for example;
12911 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12915 * New dynamic control command support for ENGINEs. ENGINEs can now
12916 declare their own commands (numbers), names (strings), descriptions,
12917 and input types for run-time discovery by calling applications. A
12918 subset of these commands are implicitly classed as "executable"
12919 depending on their input type, and only these can be invoked through
12920 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12921 can be based on user input, config files, etc). The distinction is
12922 that "executable" commands cannot return anything other than a boolean
12923 result and can only support numeric or string input, whereas some
12924 discoverable commands may only be for direct use through
12925 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12926 pointers, or other custom uses. The "executable" commands are to
12927 support parameterisations of ENGINE behaviour that can be
12928 unambiguously defined by ENGINEs and used consistently across any
12929 OpenSSL-based application. Commands have been added to all the
12930 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12931 control over shared-library paths without source code alterations.
12935 * Changed all ENGINE implementations to dynamically allocate their
12936 ENGINEs rather than declaring them statically. Apart from this being
12937 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12938 this also allows the implementations to compile without using the
12939 internal engine_int.h header.
12943 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12944 'const' value. Any code that should be able to modify a RAND_METHOD
12945 should already have non-const pointers to it (ie. they should only
12946 modify their own ones).
12950 * Made a variety of little tweaks to the ENGINE code.
12951 - "atalla" and "ubsec" string definitions were moved from header files
12952 to C code. "nuron" string definitions were placed in variables
12953 rather than hard-coded - allowing parameterisation of these values
12954 later on via ctrl() commands.
12955 - Removed unused "#if 0"'d code.
12956 - Fixed engine list iteration code so it uses ENGINE_free() to release
12957 structural references.
12958 - Constified the RAND_METHOD element of ENGINE structures.
12959 - Constified various get/set functions as appropriate and added
12960 missing functions (including a catch-all ENGINE_cpy that duplicates
12961 all ENGINE values onto a new ENGINE except reference counts/state).
12962 - Removed NULL parameter checks in get/set functions. Setting a method
12963 or function to NULL is a way of cancelling out a previously set
12964 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12965 and doesn't justify the extra error symbols and code.
12966 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12967 flags from engine_int.h to engine.h.
12968 - Changed prototypes for ENGINE handler functions (init(), finish(),
12969 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12973 * Implement binary inversion algorithm for BN_mod_inverse in addition
12974 to the algorithm using long division. The binary algorithm can be
12975 used only if the modulus is odd. On 32-bit systems, it is faster
12976 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12977 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12978 up to 450 bits. In 64-bit environments, the binary algorithm
12979 appears to be advantageous for much longer moduli; here we use it
12980 for moduli up to 2048 bits.
12984 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12985 could not support the combine flag in choice fields.
12989 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12990 extensions from a certificate request to the certificate.
12994 * Allow multiple 'certopt' and 'nameopt' options to be separated
12995 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12996 file: this allows the display of the certificate about to be
12997 signed to be customised, to allow certain fields to be included
12998 or excluded and extension details. The old system didn't display
12999 multicharacter strings properly, omitted fields not in the policy
13000 and couldn't display additional details such as extensions.
13004 * Function EC_POINTs_mul for multiple scalar multiplication
13005 of an arbitrary number of elliptic curve points
13006 \sum scalars[i]*points[i],
13007 optionally including the generator defined for the EC_GROUP:
13008 scalar*generator + \sum scalars[i]*points[i].
13010 EC_POINT_mul is a simple wrapper function for the typical case
13011 that the point list has just one item (besides the optional
13016 * First EC_METHODs for curves over GF(p):
13018 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13019 operations and provides various method functions that can also
13020 operate with faster implementations of modular arithmetic.
13022 EC_GFp_mont_method() reuses most functions that are part of
13023 EC_GFp_simple_method, but uses Montgomery arithmetic.
13025 *Bodo Moeller; point addition and point doubling
13026 implementation directly derived from source code provided by
13027 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13029 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13030 crypto/ec/ec_lib.c):
13032 Curves are EC_GROUP objects (with an optional group generator)
13033 based on EC_METHODs that are built into the library.
13035 Points are EC_POINT objects based on EC_GROUP objects.
13037 Most of the framework would be able to handle curves over arbitrary
13038 finite fields, but as there are no obvious types for fields other
13039 than GF(p), some functions are limited to that for now.
13043 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13044 that the file contains a complete HTTP response.
13048 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13049 change the def and num file printf format specifier from "%-40sXXX"
13050 to "%-39s XXX". The latter will always guarantee a space after the
13051 field while the former will cause them to run together if the field
13052 is 40 of more characters long.
13056 * Constify the cipher and digest 'method' functions and structures
13057 and modify related functions to take constant EVP_MD and EVP_CIPHER
13062 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13063 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13067 * Modify `EVP_Digest*()` routines so they now return values. Although the
13068 internal software routines can never fail additional hardware versions
13073 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13075 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13076 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13079 ERR_R_NESTED_ASN1_ERROR
13081 ERR_R_MISSING_ASN1_EOS
13082 were 4 .. 9, conflicting with
13083 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13085 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13086 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13088 Add new error code 'ERR_R_INTERNAL_ERROR'.
13092 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13097 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13098 sets the subject name for a new request or supersedes the
13099 subject name in a given request. Formats that can be parsed are
13100 'CN=Some Name, OU=myOU, C=IT'
13102 'CN=Some Name/OU=myOU/C=IT'.
13104 Add options '-batch' and '-verbose' to 'openssl req'.
13106 *Massimiliano Pala <madwolf@hackmasters.net>*
13108 * Introduce the possibility to access global variables through
13109 functions on platform were that's the best way to handle exporting
13110 global variables in shared libraries. To enable this functionality,
13111 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13112 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13113 is normally done by Configure or something similar).
13115 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13116 in the source file (foo.c) like this:
13118 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13119 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13121 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13122 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13124 OPENSSL_DECLARE_GLOBAL(int,foo);
13125 #define foo OPENSSL_GLOBAL_REF(foo)
13126 OPENSSL_DECLARE_GLOBAL(double,bar);
13127 #define bar OPENSSL_GLOBAL_REF(bar)
13129 The #defines are very important, and therefore so is including the
13130 header file everywhere where the defined globals are used.
13132 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13133 of ASN.1 items, but that structure is a bit different.
13135 The largest change is in util/mkdef.pl which has been enhanced with
13136 better and easier to understand logic to choose which symbols should
13137 go into the Windows .def files as well as a number of fixes and code
13138 cleanup (among others, algorithm keywords are now sorted
13139 lexicographically to avoid constant rewrites).
13143 * In BN_div() keep a copy of the sign of 'num' before writing the
13144 result to 'rm' because if rm==num the value will be overwritten
13145 and produce the wrong result if 'num' is negative: this caused
13146 problems with BN_mod() and BN_nnmod().
13150 * Function OCSP_request_verify(). This checks the signature on an
13151 OCSP request and verifies the signer certificate. The signer
13152 certificate is just checked for a generic purpose and OCSP request
13157 * Add OCSP_check_validity() function to check the validity of OCSP
13158 responses. OCSP responses are prepared in real time and may only
13159 be a few seconds old. Simply checking that the current time lies
13160 between thisUpdate and nextUpdate max reject otherwise valid responses
13161 caused by either OCSP responder or client clock inaccuracy. Instead
13162 we allow thisUpdate and nextUpdate to fall within a certain period of
13163 the current time. The age of the response can also optionally be
13164 checked. Two new options -validity_period and -status_age added to
13169 * If signature or public key algorithm is unrecognized print out its
13170 OID rather that just UNKNOWN.
13174 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13175 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13176 ID to be generated from the issuer certificate alone which can then be
13177 passed to OCSP_id_issuer_cmp().
13181 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13182 ASN1 modules to export functions returning ASN1_ITEM pointers
13183 instead of the ASN1_ITEM structures themselves. This adds several
13184 new macros which allow the underlying ASN1 function/structure to
13185 be accessed transparently. As a result code should not use ASN1_ITEM
13186 references directly (such as &X509_it) but instead use the relevant
13187 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13188 use of the new ASN1 code on platforms where exporting structures
13189 is problematical (for example in shared libraries) but exporting
13190 functions returning pointers to structures is not.
13194 * Add support for overriding the generation of SSL/TLS session IDs.
13195 These callbacks can be registered either in an SSL_CTX or per SSL.
13196 The purpose of this is to allow applications to control, if they wish,
13197 the arbitrary values chosen for use as session IDs, particularly as it
13198 can be useful for session caching in multiple-server environments. A
13199 command-line switch for testing this (and any client code that wishes
13200 to use such a feature) has been added to "s_server".
13202 *Geoff Thorpe, Lutz Jaenicke*
13204 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13205 of the form `#if defined(...) || defined(...) || ...` and
13206 `#if !defined(...) && !defined(...) && ...`. This also avoids
13207 the growing number of special cases it was previously handling.
13211 * Make all configuration macros available for application by making
13212 sure they are available in opensslconf.h, by giving them names starting
13213 with `OPENSSL_` to avoid conflicts with other packages and by making
13214 sure e_os2.h will cover all platform-specific cases together with
13216 Additionally, it is now possible to define configuration/platform-
13217 specific names (called "system identities"). In the C code, these
13218 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13219 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13220 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13225 * New option -set_serial to 'req' and 'x509' this allows the serial
13226 number to use to be specified on the command line. Previously self
13227 signed certificates were hard coded with serial number 0 and the
13228 CA options of 'x509' had to use a serial number in a file which was
13233 * New options to 'ca' utility to support V2 CRL entry extensions.
13234 Currently CRL reason, invalidity date and hold instruction are
13235 supported. Add new CRL extensions to V3 code and some new objects.
13239 * New function EVP_CIPHER_CTX_set_padding() this is used to
13240 disable standard block padding (aka PKCS#5 padding) in the EVP
13241 API, which was previously mandatory. This means that the data is
13242 not padded in any way and so the total length much be a multiple
13243 of the block size, otherwise an error occurs.
13247 * Initial (incomplete) OCSP SSL support.
13251 * New function OCSP_parse_url(). This splits up a URL into its host,
13252 port and path components: primarily to parse OCSP URLs. New -url
13253 option to ocsp utility.
13257 * New nonce behavior. The return value of OCSP_check_nonce() now
13258 reflects the various checks performed. Applications can decide
13259 whether to tolerate certain situations such as an absent nonce
13260 in a response when one was present in a request: the ocsp application
13261 just prints out a warning. New function OCSP_add1_basic_nonce()
13262 this is to allow responders to include a nonce in a response even if
13263 the request is nonce-less.
13267 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13268 skipped when using openssl x509 multiple times on a single input file,
13269 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13273 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13274 set string type: to handle setting ASN1_TIME structures. Fix ca
13275 utility to correctly initialize revocation date of CRLs.
13279 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13280 the clients preferred ciphersuites and rather use its own preferences.
13281 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13282 Internet Explorer by ensuring unchanged hash method during stepup.
13283 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13287 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13288 to aes and add a new 'exist' option to print out symbols that don't
13293 * Additional options to ocsp utility to allow flags to be set and
13294 additional certificates supplied.
13298 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13299 OCSP client a number of certificate to only verify the response
13304 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13305 handle the new API. Currently only ECB, CBC modes supported. Add new
13308 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13309 Encryption Standard (AES) Ciphersuites for Transport Layer
13310 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13311 not enabled by default and were not part of the "ALL" ciphersuite
13312 alias because they were not yet official; they could be
13313 explicitly requested by specifying the "AESdraft" ciphersuite
13314 group alias. In the final release of OpenSSL 0.9.7, the group
13315 alias is called "AES" and is part of "ALL".)
13317 *Ben Laurie, Steve Henson, Bodo Moeller*
13319 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13320 request to response.
13324 * Functions for OCSP responders. OCSP_request_onereq_count(),
13325 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13326 extract information from a certificate request. OCSP_response_create()
13327 creates a response and optionally adds a basic response structure.
13328 OCSP_basic_add1_status() adds a complete single response to a basic
13329 response and returns the OCSP_SINGLERESP structure just added (to allow
13330 extensions to be included for example). OCSP_basic_add1_cert() adds a
13331 certificate to a basic response and OCSP_basic_sign() signs a basic
13332 response with various flags. New helper functions ASN1_TIME_check()
13333 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13334 (converts ASN1_TIME to GeneralizedTime).
13338 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13339 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13340 structure from a certificate. X509_pubkey_digest() digests the public_key
13341 contents: this is used in various key identifiers.
13345 * Make sk_sort() tolerate a NULL argument.
13347 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13349 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13350 passed by the function are trusted implicitly. If any of them signed the
13351 response then it is assumed to be valid and is not verified.
13355 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13356 to data. This was previously part of the PKCS7 ASN1 code. This
13357 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13358 *Steve Henson, reported by Kenneth R. Robinette
13359 <support@securenetterm.com>*
13361 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13362 routines: without these tracing memory leaks is very painful.
13363 Fix leaks in PKCS12 and PKCS7 routines.
13367 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13368 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13369 effectively meant GeneralizedTime would never be used. Now it
13370 is initialised to -1 but X509_time_adj() now has to check the value
13371 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13372 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13373 *Steve Henson, reported by Kenneth R. Robinette
13374 <support@securenetterm.com>*
13376 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13377 result in a zero length in the ASN1_INTEGER structure which was
13378 not consistent with the structure when d2i_ASN1_INTEGER() was used
13379 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13380 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13381 where it did not print out a minus for negative ASN1_INTEGER.
13385 * Add summary printout to ocsp utility. The various functions which
13386 convert status values to strings have been renamed to:
13387 OCSP_response_status_str(), OCSP_cert_status_str() and
13388 OCSP_crl_reason_str() and are no longer static. New options
13389 to verify nonce values and to disable verification. OCSP response
13390 printout format cleaned up.
13394 * Add additional OCSP certificate checks. These are those specified
13395 in RFC2560. This consists of two separate checks: the CA of the
13396 certificate being checked must either be the OCSP signer certificate
13397 or the issuer of the OCSP signer certificate. In the latter case the
13398 OCSP signer certificate must contain the OCSP signing extended key
13399 usage. This check is performed by attempting to match the OCSP
13400 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13401 in the OCSP_CERTID structures of the response.
13405 * Initial OCSP certificate verification added to OCSP_basic_verify()
13406 and related routines. This uses the standard OpenSSL certificate
13407 verify routines to perform initial checks (just CA validity) and
13408 to obtain the certificate chain. Then additional checks will be
13409 performed on the chain. Currently the root CA is checked to see
13410 if it is explicitly trusted for OCSP signing. This is used to set
13411 a root CA as a global signing root: that is any certificate that
13412 chains to that CA is an acceptable OCSP signing certificate.
13416 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13417 extensions from a separate configuration file.
13418 As when reading extensions from the main configuration file,
13419 the '-extensions ...' option may be used for specifying the
13422 *Massimiliano Pala <madwolf@comune.modena.it>*
13424 * New OCSP utility. Allows OCSP requests to be generated or
13425 read. The request can be sent to a responder and the output
13426 parsed, outputted or printed in text form. Not complete yet:
13427 still needs to check the OCSP response validity.
13431 * New subcommands for 'openssl ca':
13432 `openssl ca -status <serial>` prints the status of the cert with
13433 the given serial number (according to the index file).
13434 `openssl ca -updatedb` updates the expiry status of certificates
13437 *Massimiliano Pala <madwolf@comune.modena.it>*
13439 * New '-newreq-nodes' command option to CA.pl. This is like
13440 '-newreq', but calls 'openssl req' with the '-nodes' option
13441 so that the resulting key is not encrypted.
13443 *Damien Miller <djm@mindrot.org>*
13445 * New configuration for the GNU Hurd.
13447 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13449 * Initial code to implement OCSP basic response verify. This
13450 is currently incomplete. Currently just finds the signer's
13451 certificate and verifies the signature on the response.
13455 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13456 value of OPENSSLDIR. This is available via the new '-d' option
13457 to 'openssl version', and is also included in 'openssl version -a'.
13461 * Allowing defining memory allocation callbacks that will be given
13462 file name and line number information in additional arguments
13463 (a `const char*` and an int). The basic functionality remains, as
13464 well as the original possibility to just replace malloc(),
13465 realloc() and free() by functions that do not know about these
13466 additional arguments. To register and find out the current
13467 settings for extended allocation functions, the following
13468 functions are provided:
13470 CRYPTO_set_mem_ex_functions
13471 CRYPTO_set_locked_mem_ex_functions
13472 CRYPTO_get_mem_ex_functions
13473 CRYPTO_get_locked_mem_ex_functions
13475 These work the same way as CRYPTO_set_mem_functions and friends.
13476 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13477 extended allocation function is enabled.
13478 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13479 a conventional allocation function is enabled.
13481 *Richard Levitte, Bodo Moeller*
13483 * Finish off removing the remaining LHASH function pointer casts.
13484 There should no longer be any prototype-casting required when using
13485 the LHASH abstraction, and any casts that remain are "bugs". See
13486 the callback types and macros at the head of lhash.h for details
13487 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13491 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13492 If /dev/[u]random devices are not available or do not return enough
13493 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13495 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13496 /etc/entropy will be queried once each in this sequence, querying stops
13497 when enough entropy was collected without querying more sockets.
13501 * Change the Unix RAND_poll() variant to be able to poll several
13502 random devices, as specified by DEVRANDOM, until a sufficient amount
13503 of data has been collected. We spend at most 10 ms on each file
13504 (select timeout) and read in non-blocking mode. DEVRANDOM now
13505 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13506 (previously it was just the string "/dev/urandom"), so on typical
13507 platforms the 10 ms delay will never occur.
13508 Also separate out the Unix variant to its own file, rand_unix.c.
13509 For VMS, there's a currently-empty rand_vms.c.
13513 * Move OCSP client related routines to ocsp_cl.c. These
13514 provide utility functions which an application needing
13515 to issue a request to an OCSP responder and analyse the
13516 response will typically need: as opposed to those which an
13517 OCSP responder itself would need which will be added later.
13519 OCSP_request_sign() signs an OCSP request with an API similar
13520 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13521 response. OCSP_response_get1_basic() extracts basic response
13522 from response. OCSP_resp_find_status(): finds and extracts status
13523 information from an OCSP_CERTID structure (which will be created
13524 when the request structure is built). These are built from lower
13525 level functions which work on OCSP_SINGLERESP structures but
13526 won't normally be used unless the application wishes to examine
13527 extensions in the OCSP response for example.
13529 Replace nonce routines with a pair of functions.
13530 OCSP_request_add1_nonce() adds a nonce value and optionally
13531 generates a random value. OCSP_check_nonce() checks the
13532 validity of the nonce in an OCSP response.
13536 * Change function OCSP_request_add() to OCSP_request_add0_id().
13537 This doesn't copy the supplied OCSP_CERTID and avoids the
13538 need to free up the newly created id. Change return type
13539 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13540 This can then be used to add extensions to the request.
13541 Deleted OCSP_request_new(), since most of its functionality
13542 is now in OCSP_REQUEST_new() (and the case insensitive name
13543 clash) apart from the ability to set the request name which
13544 will be added elsewhere.
13548 * Update OCSP API. Remove obsolete extensions argument from
13549 various functions. Extensions are now handled using the new
13550 OCSP extension code. New simple OCSP HTTP function which
13551 can be used to send requests and parse the response.
13555 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13556 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13557 uses the special reorder version of SET OF to sort the attributes
13558 and reorder them to match the encoded order. This resolves a long
13559 standing problem: a verify on a PKCS7 structure just after signing
13560 it used to fail because the attribute order did not match the
13561 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13562 it uses the received order. This is necessary to tolerate some broken
13563 software that does not order SET OF. This is handled by encoding
13564 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13565 to produce the required SET OF.
13569 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13570 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13571 files to get correct declarations of the ASN.1 item variables.
13575 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13576 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13577 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13578 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13579 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13580 ASN1_ITEM and no wrapper functions.
13584 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13585 replace the old function pointer based I/O routines. Change most of
13586 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13590 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13591 lines, recognize more "algorithms" that can be deselected, and make
13592 it complain about algorithm deselection that isn't recognised.
13596 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13597 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13598 to use new functions. Add NO_ASN1_OLD which can be set to remove
13599 some old style ASN1 functions: this can be used to determine if old
13600 code will still work when these eventually go away.
13604 * New extension functions for OCSP structures, these follow the
13605 same conventions as certificates and CRLs.
13609 * New function X509V3_add1_i2d(). This automatically encodes and
13610 adds an extension. Its behaviour can be customised with various
13611 flags to append, replace or delete. Various wrappers added for
13612 certificates and CRLs.
13616 * Fix to avoid calling the underlying ASN1 print routine when
13617 an extension cannot be parsed. Correct a typo in the
13618 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13622 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13623 entries for variables.
13627 * Add functionality to `apps/openssl.c` for detecting locking
13628 problems: As the program is single-threaded, all we have
13629 to do is register a locking callback using an array for
13630 storing which locks are currently held by the program.
13634 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13635 SSL_get_ex_data_X509_STORE_idx(), which is used in
13636 ssl_verify_cert_chain() and thus can be called at any time
13637 during TLS/SSL handshakes so that thread-safety is essential.
13638 Unfortunately, the ex_data design is not at all suited
13639 for multi-threaded use, so it probably should be abolished.
13643 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13645 *Broadcom, tweaked and integrated by Geoff Thorpe*
13647 * Move common extension printing code to new function
13648 X509V3_print_extensions(). Reorganise OCSP print routines and
13649 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13653 * New function X509_signature_print() to remove duplication in some
13658 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13659 set (this was treated exactly the same as SET OF previously). This
13660 is used to reorder the STACK representing the structure to match the
13661 encoding. This will be used to get round a problem where a PKCS7
13662 structure which was signed could not be verified because the STACK
13663 order did not reflect the encoded order.
13667 * Reimplement the OCSP ASN1 module using the new code.
13671 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13672 for its ASN1 operations. The old style function pointers still exist
13673 for now but they will eventually go away.
13677 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13678 completely replaces the old ASN1 functionality with a table driven
13679 encoder and decoder which interprets an ASN1_ITEM structure describing
13680 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13681 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13682 has also been converted to the new form.
13686 * Change BN_mod_exp_recp so that negative moduli are tolerated
13687 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13688 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13689 for negative moduli.
13693 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13694 of not touching the result's sign bit.
13698 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13703 * Changed the LHASH code to use prototypes for callbacks, and created
13704 macros to declare and implement thin (optionally static) functions
13705 that provide type-safety and avoid function pointer casting for the
13706 type-specific callbacks.
13710 * Added Kerberos Cipher Suites to be used with TLS, as written in
13712 *Veers Staats <staatsvr@asc.hpc.mil>,
13713 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13715 * Reformat the FAQ so the different questions and answers can be divided
13716 in sections depending on the subject.
13720 * Have the zlib compression code load ZLIB.DLL dynamically under
13725 * New function BN_mod_sqrt for computing square roots modulo a prime
13726 (using the probabilistic Tonelli-Shanks algorithm unless
13727 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13728 be handled deterministically).
13730 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13732 * Make BN_mod_inverse faster by explicitly handling small quotients
13733 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13734 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13738 * New function BN_kronecker.
13742 * Fix BN_gcd so that it works on negative inputs; the result is
13743 positive unless both parameters are zero.
13744 Previously something reasonably close to an infinite loop was
13745 possible because numbers could be growing instead of shrinking
13746 in the implementation of Euclid's algorithm.
13750 * Fix BN_is_word() and BN_is_one() macros to take into account the
13751 sign of the number in question.
13753 Fix BN_is_word(a,w) to work correctly for w == 0.
13755 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13756 because its test if the absolute value of 'a' equals 'w'.
13757 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13758 it exists mostly for use in the implementations of BN_is_zero(),
13759 BN_is_one(), and BN_is_word().
13763 * New function BN_swap.
13767 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13768 the exponentiation functions are more likely to produce reasonable
13769 results on negative inputs.
13773 * Change BN_mod_mul so that the result is always non-negative.
13774 Previously, it could be negative if one of the factors was negative;
13775 I don't think anyone really wanted that behaviour.
13779 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13780 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13781 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13782 and add new functions:
13791 BN_mod_lshift1_quick
13793 BN_mod_lshift_quick
13795 These functions always generate non-negative results.
13797 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13798 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13800 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13801 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13802 be reduced modulo `m`.
13804 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13807 The following entry accidentally appeared in the CHANGES file
13808 distributed with OpenSSL 0.9.7. The modifications described in
13809 it do *not* apply to OpenSSL 0.9.7.
13811 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13812 was actually never needed) and in BN_mul(). The removal in BN_mul()
13813 required a small change in bn_mul_part_recursive() and the addition
13814 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13815 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13816 bn_sub_words() and bn_add_words() except they take arrays with
13822 * In 'openssl passwd', verify passwords read from the terminal
13823 unless the '-salt' option is used (which usually means that
13824 verification would just waste user's time since the resulting
13825 hash is going to be compared with some given password hash)
13826 or the new '-noverify' option is used.
13828 This is an incompatible change, but it does not affect
13829 non-interactive use of 'openssl passwd' (passwords on the command
13830 line, '-stdin' option, '-in ...' option) and thus should not
13831 cause any problems.
13835 * Remove all references to RSAref, since there's no more need for it.
13839 * Make DSO load along a path given through an environment variable
13840 (SHLIB_PATH) with shl_load().
13844 * Constify the ENGINE code as a result of BIGNUM constification.
13845 Also constify the RSA code and most things related to it. In a
13846 few places, most notable in the depth of the ASN.1 code, ugly
13847 casts back to non-const were required (to be solved at a later
13852 * Make it so the openssl application has all engines loaded by default.
13856 * Constify the BIGNUM routines a little more.
13860 * Add the following functions:
13862 ENGINE_load_cswift()
13864 ENGINE_load_atalla()
13865 ENGINE_load_nuron()
13866 ENGINE_load_builtin_engines()
13868 That way, an application can itself choose if external engines that
13869 are built-in in OpenSSL shall ever be used or not. The benefit is
13870 that applications won't have to be linked with libdl or other dso
13871 libraries unless it's really needed.
13873 Changed 'openssl engine' to load all engines on demand.
13874 Changed the engine header files to avoid the duplication of some
13875 declarations (they differed!).
13879 * 'openssl engine' can now list capabilities.
13883 * Better error reporting in 'openssl engine'.
13887 * Never call load_dh_param(NULL) in s_server.
13891 * Add engine application. It can currently list engines by name and
13892 identity, and test if they are actually available.
13896 * Improve RPM specification file by forcing symbolic linking and making
13897 sure the installed documentation is also owned by root.root.
13899 *Damien Miller <djm@mindrot.org>*
13901 * Give the OpenSSL applications more possibilities to make use of
13902 keys (public as well as private) handled by engines.
13906 * Add OCSP code that comes from CertCo.
13910 * Add VMS support for the Rijndael code.
13914 * Added untested support for Nuron crypto accelerator.
13918 * Add support for external cryptographic devices. This code was
13919 previously distributed separately as the "engine" branch.
13921 *Geoff Thorpe, Richard Levitte*
13923 * Rework the filename-translation in the DSO code. It is now possible to
13924 have far greater control over how a "name" is turned into a filename
13925 depending on the operating environment and any oddities about the
13926 different shared library filenames on each system.
13930 * Support threads on FreeBSD-elf in Configure.
13934 * Fix for SHA1 assembly problem with MASM: it produces
13935 warnings about corrupt line number information when assembling
13936 with debugging information. This is caused by the overlapping
13939 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13942 NCONF_get_number() has no error checking at all. As a replacement,
13943 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13944 promoted strongly. The old NCONF_get_number is kept around for
13945 binary backward compatibility.
13946 Make it possible for methods to load from something other than a BIO,
13947 by providing a function pointer that is given a name instead of a BIO.
13948 For example, this could be used to load configuration data from an
13953 * Fix for non blocking accept BIOs. Added new I/O special reason
13954 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13955 with non blocking I/O was not possible because no retry code was
13956 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13961 * Added the beginnings of Rijndael support.
13965 * Fix for bug in DirectoryString mask setting. Add support for
13966 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13967 to allow certificate printing to more controllable, additional
13968 'certopt' option to 'x509' to allow new printing options to be
13973 * Clean old EAY MD5 hack from e_os.h.
13977 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13979 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13980 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13982 *Joe Orton, Steve Henson*
13984 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13986 * Fix additional bug revealed by the NISCC test suite:
13988 Stop bug triggering large recursion when presented with
13989 certain ASN.1 tags ([CVE-2003-0851])
13993 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13995 * Fix various bugs revealed by running the NISCC test suite:
13997 Stop out of bounds reads in the ASN1 code when presented with
13998 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14000 If verify callback ignores invalid public key errors don't try to check
14001 certificate signature with the NULL public key.
14005 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14006 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14011 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14012 extra data after the compression methods not only for TLS 1.0
14013 but also for SSL 3.0 (as required by the specification).
14015 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14017 * Change X509_certificate_type() to mark the key as exported/exportable
14018 when it's 512 *bits* long, not 512 bytes.
14022 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14024 * Countermeasure against the Klima-Pokorny-Rosa extension of
14025 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14026 a protocol version number mismatch like a decryption error
14027 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14031 * Turn on RSA blinding by default in the default implementation
14032 to avoid a timing attack. Applications that don't want it can call
14033 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14034 They would be ill-advised to do so in most cases.
14036 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14038 * Change RSA blinding code so that it works when the PRNG is not
14039 seeded (in this case, the secret RSA exponent is abused as
14040 an unpredictable seed -- if it is not unpredictable, there
14041 is no point in blinding anyway). Make RSA blinding thread-safe
14042 by remembering the creator's thread ID in rsa->blinding and
14043 having all other threads use local one-time blinding factors
14044 (this requires more computation than sharing rsa->blinding, but
14045 avoids excessive locking; and if an RSA object is not shared
14046 between threads, blinding will still be very fast).
14050 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14052 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14053 via timing by performing a MAC computation even if incorrect
14054 block cipher padding has been found. This is a countermeasure
14055 against active attacks where the attacker has to distinguish
14056 between bad padding and a MAC verification error. ([CVE-2003-0078])
14058 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14059 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14060 Martin Vuagnoux (EPFL, Ilion)*
14062 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14064 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14065 memory from its contents. This is done with a counter that will
14066 place alternating values in each byte. This can be used to solve
14067 two issues: 1) the removal of calls to memset() by highly optimizing
14068 compilers, and 2) cleansing with other values than 0, since those can
14069 be read through on certain media, for example a swap space on disk.
14073 * Bugfix: client side session caching did not work with external caching,
14074 because the session->cipher setting was not restored when reloading
14075 from the external cache. This problem was masked, when
14076 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14077 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14081 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14082 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14084 *Zeev Lieber <zeev-l@yahoo.com>*
14086 * Undo an undocumented change introduced in 0.9.6e which caused
14087 repeated calls to OpenSSL_add_all_ciphers() and
14088 OpenSSL_add_all_digests() to be ignored, even after calling
14093 * Change the default configuration reader to deal with last line not
14094 being properly terminated.
14098 * Change X509_NAME_cmp() so it applies the special rules on handling
14099 DN values that are of type PrintableString, as well as RDNs of type
14100 emailAddress where the value has the type ia5String.
14102 *stefank@valicert.com via Richard Levitte*
14104 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14105 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14106 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14107 the bitwise-OR of the two for use by the majority of applications
14108 wanting this behaviour, and update the docs. The documented
14109 behaviour and actual behaviour were inconsistent and had been
14110 changing anyway, so this is more a bug-fix than a behavioural
14113 *Geoff Thorpe, diagnosed by Nadav Har'El*
14115 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14116 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14120 * Fix initialization code race conditions in
14121 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14122 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14123 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14124 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14125 ssl2_get_cipher_by_char(),
14126 ssl3_get_cipher_by_char().
14128 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14130 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14131 the cached sessions are flushed, as the remove_cb() might use ex_data
14132 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14133 (see [openssl.org #212]).
14135 *Geoff Thorpe, Lutz Jaenicke*
14137 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14138 length, instead of the encoding length to d2i_ASN1_OBJECT.
14142 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14144 * [In 0.9.6g-engine release:]
14145 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14147 *Lynn Gazis <lgazis@rainbow.com>*
14149 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14151 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14152 and get fix the header length calculation.
14153 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14154 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14156 * Use proper error handling instead of 'assertions' in buffer
14157 overflow checks added in 0.9.6e. This prevents DoS (the
14158 assertions could call abort()).
14160 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14162 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14164 * Add various sanity checks to asn1_get_length() to reject
14165 the ASN1 length bytes if they exceed sizeof(long), will appear
14166 negative or the content length exceeds the length of the
14169 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14171 * Fix cipher selection routines: ciphers without encryption had no flags
14172 for the cipher strength set and where therefore not handled correctly
14173 by the selection routines (PR #130).
14177 * Fix EVP_dsa_sha macro.
14182 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14183 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14184 that was added in OpenSSL 0.9.6d.
14186 As the countermeasure turned out to be incompatible with some
14187 broken SSL implementations, the new option is part of SSL_OP_ALL.
14188 SSL_OP_ALL is usually employed when compatibility with weird SSL
14189 implementations is desired (e.g. '-bugs' option to 's_client' and
14190 's_server'), so the new option is automatically set in many
14195 * Changes in security patch:
14197 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14198 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14199 Air Force Materiel Command, USAF, under agreement number
14202 * Add various sanity checks to asn1_get_length() to reject
14203 the ASN1 length bytes if they exceed sizeof(long), will appear
14204 negative or the content length exceeds the length of the
14205 supplied buffer. ([CVE-2002-0659])
14207 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14209 * Assertions for various potential buffer overflows, not known to
14210 happen in practice.
14212 *Ben Laurie (CHATS)*
14214 * Various temporary buffers to hold ASCII versions of integers were
14215 too small for 64 bit platforms. ([CVE-2002-0655])
14216 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14218 * Remote buffer overflow in SSL3 protocol - an attacker could
14219 supply an oversized session ID to a client. ([CVE-2002-0656])
14221 *Ben Laurie (CHATS)*
14223 * Remote buffer overflow in SSL2 protocol - an attacker could
14224 supply an oversized client master key. ([CVE-2002-0656])
14226 *Ben Laurie (CHATS)*
14228 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14230 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14231 encoded as NULL) with id-dsa-with-sha1.
14233 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14235 * Check various `X509_...()` return values in `apps/req.c`.
14237 *Nils Larsch <nla@trustcenter.de>*
14239 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14240 an end-of-file condition would erroneously be flagged, when the CRLF
14241 was just at the end of a processed block. The bug was discovered when
14242 processing data through a buffering memory BIO handing the data to a
14243 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14244 <ptsekov@syntrex.com> and Nedelcho Stanev.
14248 * Implement a countermeasure against a vulnerability recently found
14249 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14250 before application data chunks to avoid the use of known IVs
14251 with data potentially chosen by the attacker.
14255 * Fix length checks in ssl3_get_client_hello().
14259 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14260 to prevent ssl3_read_internal() from incorrectly assuming that
14261 ssl3_read_bytes() found application data while handshake
14262 processing was enabled when in fact s->s3->in_read_app_data was
14263 merely automatically cleared during the initial handshake.
14265 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14267 * Fix object definitions for Private and Enterprise: they were not
14268 recognized in their shortname (=lowercase) representation. Extend
14269 obj_dat.pl to issue an error when using undefined keywords instead
14270 of silently ignoring the problem (Svenning Sorensen
14271 <sss@sss.dnsalias.net>).
14275 * Fix DH_generate_parameters() so that it works for 'non-standard'
14276 generators, i.e. generators other than 2 and 5. (Previously, the
14277 code did not properly initialise the 'add' and 'rem' values to
14278 BN_generate_prime().)
14280 In the new general case, we do not insist that 'generator' is
14281 actually a primitive root: This requirement is rather pointless;
14282 a generator of the order-q subgroup is just as good, if not
14287 * Map new X509 verification errors to alerts. Discovered and submitted by
14288 Tom Wu <tom@arcot.com>.
14292 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14293 returning non-zero before the data has been completely received
14294 when using non-blocking I/O.
14296 *Bodo Moeller; problem pointed out by John Hughes*
14298 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14300 *Ben Laurie, Lutz Jaenicke*
14302 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14303 Yoram Zahavi <YoramZ@gilian.com>).
14307 * Add information about CygWin 1.3 and on, and preserve proper
14308 configuration for the versions before that.
14310 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14312 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14313 check whether we deal with a copy of a session and do not delete from
14314 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14315 <izhar@checkpoint.com>.
14319 * Do not store session data into the internal session cache, if it
14320 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14321 flag is set). Proposed by Aslam <aslam@funk.com>.
14325 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14330 * [In 0.9.6d-engine release:]
14331 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14333 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14335 * Add the configuration target linux-s390x.
14337 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14339 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14340 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14341 variable as an indication that a ClientHello message has been
14342 received. As the flag value will be lost between multiple
14343 invocations of ssl3_accept when using non-blocking I/O, the
14344 function may not be aware that a handshake has actually taken
14345 place, thus preventing a new session from being added to the
14348 To avoid this problem, we now set s->new_session to 2 instead of
14349 using a local variable.
14351 *Lutz Jaenicke, Bodo Moeller*
14353 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14354 if the SSL_R_LENGTH_MISMATCH error is detected.
14356 *Geoff Thorpe, Bodo Moeller*
14358 * New 'shared_ldflag' column in Configure platform table.
14362 * Fix EVP_CIPHER_mode macro.
14364 *"Dan S. Camper" <dan@bti.net>*
14366 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14367 type, we must throw them away by setting rr->length to 0.
14369 *D P Chang <dpc@qualys.com>*
14371 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14373 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14374 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14375 worked incorrectly for those cases where range = `10..._2` and
14376 `3*range` is two bits longer than range.)
14380 * Only add signing time to PKCS7 structures if it is not already
14385 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14386 OBJ_ld_ce should be OBJ_id_ce.
14387 Also some ip-pda OIDs in crypto/objects/objects.txt were
14388 incorrect (cf. RFC 3039).
14390 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14392 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14393 returns early because it has nothing to do.
14395 *Andy Schneider <andy.schneider@bjss.co.uk>*
14397 * [In 0.9.6c-engine release:]
14398 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14400 *Andy Schneider <andy.schneider@bjss.co.uk>*
14402 * [In 0.9.6c-engine release:]
14403 Add support for Cryptographic Appliance's keyserver technology.
14404 (Use engine 'keyclient')
14406 *Cryptographic Appliances and Geoff Thorpe*
14408 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14409 is called via tools/c89.sh because arguments have to be
14410 rearranged (all '-L' options must appear before the first object
14413 *Richard Shapiro <rshapiro@abinitio.com>*
14415 * [In 0.9.6c-engine release:]
14416 Add support for Broadcom crypto accelerator cards, backported
14419 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14421 * [In 0.9.6c-engine release:]
14422 Add support for SureWare crypto accelerator cards from
14423 Baltimore Technologies. (Use engine 'sureware')
14425 *Baltimore Technologies and Mark Cox*
14427 * [In 0.9.6c-engine release:]
14428 Add support for crypto accelerator cards from Accelerated
14429 Encryption Processing, www.aep.ie. (Use engine 'aep')
14431 *AEP Inc. and Mark Cox*
14433 * Add a configuration entry for gcc on UnixWare.
14435 *Gary Benson <gbenson@redhat.com>*
14437 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14438 messages are stored in a single piece (fixed-length part and
14439 variable-length part combined) and fix various bugs found on the way.
14443 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14444 instead. BIO_gethostbyname() does not know what timeouts are
14445 appropriate, so entries would stay in cache even when they have
14447 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14449 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14450 faced with a pathologically small ClientHello fragment that does
14451 not contain client_version: Instead of aborting with an error,
14452 simply choose the highest available protocol version (i.e.,
14453 TLS 1.0 unless it is disabled). In practice, ClientHello
14454 messages are never sent like this, but this change gives us
14455 strictly correct behaviour at least for TLS.
14459 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14460 never resets s->method to s->ctx->method when called from within
14461 one of the SSL handshake functions.
14463 *Bodo Moeller; problem pointed out by Niko Baric*
14465 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14466 (sent using the client's version number) if client_version is
14467 smaller than the protocol version in use. Also change
14468 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14469 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14470 the client will at least see that alert.
14474 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14479 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14480 client receives HelloRequest while in a handshake.
14482 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14484 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14485 should end in 'break', not 'goto end' which circumvents various
14486 cleanups done in state SSL_ST_OK. But session related stuff
14487 must be disabled for SSL_ST_OK in the case that we just sent a
14490 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14491 before just sending a HelloRequest.
14493 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14495 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14496 reveal whether illegal block cipher padding was found or a MAC
14497 verification error occurred. (Neither SSLerr() codes nor alerts
14498 are directly visible to potential attackers, but the information
14499 may leak via logfiles.)
14501 Similar changes are not required for the SSL 2.0 implementation
14502 because the number of padding bytes is sent in clear for SSL 2.0,
14503 and the extra bytes are just ignored. However ssl/s2_pkt.c
14504 failed to verify that the purported number of padding bytes is in
14509 * Add OpenUNIX-8 support including shared libraries
14510 (Boyd Lynn Gerber <gerberb@zenez.com>).
14514 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14515 'wristwatch attack' using huge encoding parameters (cf.
14516 James H. Manger's CRYPTO 2001 paper). Note that the
14517 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14518 encoding parameters and hence was not vulnerable.
14522 * BN_sqr() bug fix.
14524 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14526 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14527 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14528 followed by modular reduction.
14530 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14532 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14533 equivalent based on BN_pseudo_rand() instead of BN_rand().
14537 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14538 This function was broken, as the check for a new client hello message
14539 to handle SGC did not allow these large messages.
14540 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14544 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14548 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14549 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14553 * Rework the configuration and shared library support for Tru64 Unix.
14554 The configuration part makes use of modern compiler features and
14555 still retains old compiler behavior for those that run older versions
14556 of the OS. The shared library support part includes a variant that
14557 uses the RPATH feature, and is available through the special
14558 configuration target "alpha-cc-rpath", which will never be selected
14561 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14563 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14564 with the same message size as in ssl3_get_certificate_request().
14565 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14566 messages might inadvertently be reject as too long.
14568 *Petr Lampa <lampa@fee.vutbr.cz>*
14570 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14574 * Modified SSL library such that the verify_callback that has been set
14575 specifically for an SSL object with SSL_set_verify() is actually being
14576 used. Before the change, a verify_callback set with this function was
14577 ignored and the verify_callback() set in the SSL_CTX at the time of
14578 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14579 to allow the necessary settings.
14583 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14584 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14585 done automatically (in contradiction to the requirements of the C
14586 standard). This made problems when used from OpenSSH.
14590 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14591 dh->length and always used
14593 BN_rand_range(priv_key, dh->p).
14595 BN_rand_range() is not necessary for Diffie-Hellman, and this
14596 specific range makes Diffie-Hellman unnecessarily inefficient if
14597 dh->length (recommended exponent length) is much smaller than the
14598 length of dh->p. We could use BN_rand_range() if the order of
14599 the subgroup was stored in the DH structure, but we only have
14604 BN_rand(priv_key, l, ...)
14606 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14613 RSA_eay_public_encrypt
14614 RSA_eay_private_decrypt
14615 RSA_eay_private_encrypt (signing)
14616 RSA_eay_public_decrypt (signature verification)
14618 (default implementations for RSA_public_encrypt,
14619 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14620 always reject numbers >= n.
14624 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14625 to synchronize access to 'locking_thread'. This is necessary on
14626 systems where access to 'locking_thread' (an 'unsigned long'
14627 variable) is not atomic.
14631 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14632 *before* setting the 'crypto_lock_rand' flag. The previous code had
14633 a race condition if 0 is a valid thread ID.
14635 *Travis Vitek <vitek@roguewave.com>*
14637 * Add support for shared libraries under Irix.
14639 *Albert Chin-A-Young <china@thewrittenword.com>*
14641 * Add configuration option to build on Linux on both big-endian and
14642 little-endian MIPS.
14644 *Ralf Baechle <ralf@uni-koblenz.de>*
14646 * Add the possibility to create shared libraries on HP-UX.
14650 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14652 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14653 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14654 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14655 PRNG state recovery was possible based on the output of
14656 one PRNG request appropriately sized to gain knowledge on
14657 'md' followed by enough consecutive 1-byte PRNG requests
14658 to traverse all of 'state'.
14660 1. When updating 'md_local' (the current thread's copy of 'md')
14661 during PRNG output generation, hash all of the previous
14662 'md_local' value, not just the half used for PRNG output.
14664 2. Make the number of bytes from 'state' included into the hash
14665 independent from the number of PRNG bytes requested.
14667 The first measure alone would be sufficient to avoid
14668 Markku-Juhani's attack. (Actually it had never occurred
14669 to me that the half of 'md_local' used for chaining was the
14670 half from which PRNG output bytes were taken -- I had always
14671 assumed that the secret half would be used.) The second
14672 measure makes sure that additional data from 'state' is never
14673 mixed into 'md_local' in small portions; this heuristically
14674 further strengthens the PRNG.
14678 * Fix crypto/bn/asm/mips3.s.
14682 * When only the key is given to "enc", the IV is undefined. Print out
14683 an error message in this case.
14687 * Handle special case when X509_NAME is empty in X509 printing routines.
14691 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14692 positive and less than q.
14696 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14697 used: it isn't thread safe and the add_lock_callback should handle
14700 *Paul Rose <Paul.Rose@bridge.com>*
14702 * Verify that incoming data obeys the block size in
14703 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14709 *Ulf Möller, Bodo Möller*
14711 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14712 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14713 when fixing the server behaviour for backwards-compatible 'client
14714 hello' messages. (Note that the attack is impractical against
14715 SSL 3.0 and TLS 1.0 anyway because length and version checking
14716 means that the probability of guessing a valid ciphertext is
14717 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14720 Before 0.9.5, the countermeasure (hide the error by generating a
14721 random 'decryption result') did not work properly because
14722 ERR_clear_error() was missing, meaning that SSL_get_error() would
14723 detect the supposedly ignored error.
14725 Both problems are now fixed.
14729 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14730 (previously it was 1024).
14734 * Fix for compatibility mode trust settings: ignore trust settings
14735 unless some valid trust or reject settings are present.
14739 * Fix for blowfish EVP: its a variable length cipher.
14743 * Fix various bugs related to DSA S/MIME verification. Handle missing
14744 parameters in DSA public key structures and return an error in the
14745 DSA routines if parameters are absent.
14749 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14750 in the current directory if neither $RANDFILE nor $HOME was set.
14751 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14752 caused some confusion to Windows users who haven't defined $HOME.
14753 Thus RAND_file_name() is changed again: e_os.h can define a
14754 DEFAULT_HOME, which will be used if $HOME is not set.
14755 For Windows, we use "C:"; on other platforms, we still require
14756 environment variables.
14758 * Move 'if (!initialized) RAND_poll()' into regions protected by
14759 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14760 having multiple threads call RAND_poll() concurrently.
14764 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14765 combination of a flag and a thread ID variable.
14766 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14767 flag), *other* threads can enter ssleay_add_bytes without obeying
14768 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14769 that they do not hold after the first thread unsets add_do_not_lock).
14773 * Change bctest again: '-x' expressions are not available in all
14774 versions of 'test'.
14778 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14780 * Fix a couple of memory leaks in PKCS7_dataDecode()
14782 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14784 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14785 the default extension for executables, if any. Also, make the perl
14786 scripts that use symlink() to test if it really exists and use "cp"
14787 if it doesn't. All this made OpenSSL compilable and installable in
14792 * Fix for asn1_GetSequence() for indefinite length constructed data.
14793 If SEQUENCE is length is indefinite just set c->slen to the total
14794 amount of data available.
14796 *Steve Henson, reported by shige@FreeBSD.org*
14798 *This change does not apply to 0.9.7.*
14800 * Change bctest to avoid here-documents inside command substitution
14801 (workaround for FreeBSD /bin/sh bug).
14802 For compatibility with Ultrix, avoid shell functions (introduced
14803 in the bctest version that searches along $PATH).
14807 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14808 with des_encrypt() defined on some operating systems, like Solaris
14813 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14814 On the Importance of Eliminating Errors in Cryptographic
14815 Computations, J. Cryptology 14 (2001) 2, 101-119,
14816 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14820 * MIPS assembler BIGNUM division bug fix.
14824 * Disabled incorrect Alpha assembler code.
14828 * Fix PKCS#7 decode routines so they correctly update the length
14829 after reading an EOC for the EXPLICIT tag.
14833 *This change does not apply to 0.9.7.*
14835 * Fix bug in PKCS#12 key generation routines. This was triggered
14836 if a 3DES key was generated with a 0 initial byte. Include
14837 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14838 (but broken) behaviour.
14842 * Enhance bctest to search for a working bc along $PATH and print
14845 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14847 * Fix memory leaks in err.c: free err_data string if necessary;
14848 don't write to the wrong index in ERR_set_error_data.
14852 * Implement ssl23_peek (analogous to ssl23_read), which previously
14857 * Replace rdtsc with `_emit` statements for VC++ version 5.
14859 *Jeremy Cooper <jeremy@baymoo.org>*
14861 * Make it possible to reuse SSLv2 sessions.
14865 * In copy_email() check for >= 0 as a return value for
14866 X509_NAME_get_index_by_NID() since 0 is a valid index.
14868 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14870 * Avoid coredump with unsupported or invalid public keys by checking if
14871 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14872 PKCS7_verify() fails with non detached data.
14876 * Don't use getenv in library functions when run as setuid/setgid.
14877 New function OPENSSL_issetugid().
14881 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14882 due to incorrect handling of multi-threading:
14884 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14886 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14888 3. Count how many times MemCheck_off() has been called so that
14889 nested use can be treated correctly. This also avoids
14890 inband-signalling in the previous code (which relied on the
14891 assumption that thread ID 0 is impossible).
14895 * Add "-rand" option also to s_client and s_server.
14899 * Fix CPU detection on Irix 6.x.
14900 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14901 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14903 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14908 *This change does not apply to 0.9.7.*
14910 * Use the cached encoding of an X509_NAME structure rather than
14911 copying it. This is apparently the reason for the libsafe "errors"
14912 but the code is actually correct.
14916 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14917 Bleichenbacher's DSA attack.
14918 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14919 to be set and top=0 forces the highest bit to be set; top=-1 is new
14920 and leaves the highest bit random.
14922 *Ulf Moeller, Bodo Moeller*
14924 * In the `NCONF_...`-based implementations for `CONF_...` queries
14925 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14926 a temporary CONF structure with the data component set to NULL
14927 (which gives segmentation faults in lh_retrieve).
14928 Instead, use NULL for the CONF pointer in CONF_get_string and
14929 CONF_get_number (which may use environment variables) and directly
14930 return NULL from CONF_get_section.
14934 * Fix potential buffer overrun for EBCDIC.
14938 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14939 keyUsage if basicConstraints absent for a CA.
14943 * Make SMIME_write_PKCS7() write mail header values with a format that
14944 is more generally accepted (no spaces before the semicolon), since
14945 some programs can't parse those values properly otherwise. Also make
14946 sure BIO's that break lines after each write do not create invalid
14951 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14952 macros previously used would not encode an empty SEQUENCE OF
14953 and break the signature.
14957 *This change does not apply to 0.9.7.*
14959 * Zero the premaster secret after deriving the master secret in
14964 * Add some EVP_add_digest_alias registrations (as found in
14965 OpenSSL_add_all_digests()) to SSL_library_init()
14966 aka OpenSSL_add_ssl_algorithms(). This provides improved
14967 compatibility with peers using X.509 certificates
14968 with unconventional AlgorithmIdentifier OIDs.
14972 * Fix for Irix with NO_ASM.
14974 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14976 * ./config script fixes.
14978 *Ulf Moeller, Richard Levitte*
14980 * Fix 'openssl passwd -1'.
14984 * Change PKCS12_key_gen_asc() so it can cope with non null
14985 terminated strings whose length is passed in the passlen
14986 parameter, for example from PEM callbacks. This was done
14987 by adding an extra length parameter to asc2uni().
14989 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14991 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14992 call failed, free the DSA structure.
14996 * Fix to uni2asc() to cope with zero length Unicode strings.
14997 These are present in some PKCS#12 files.
15001 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15002 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15003 when writing a 32767 byte record.
15005 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15007 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15008 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15010 (RSA objects have a reference count access to which is protected
15011 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15012 so they are meant to be shared between threads.)
15013 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15014 "Reddie, Steven" <Steven.Reddie@ca.com>*
15016 * Fix a deadlock in CRYPTO_mem_leaks().
15020 * Use better test patterns in bntest.
15024 * rand_win.c fix for Borland C.
15028 * BN_rshift bugfix for n == 0.
15032 * Add a 'bctest' script that checks for some known 'bc' bugs
15033 so that 'make test' does not abort just because 'bc' is broken.
15037 * Store verify_result within SSL_SESSION also for client side to
15038 avoid potential security hole. (Re-used sessions on the client side
15039 always resulted in verify_result==X509_V_OK, not using the original
15040 result of the server certificate verification.)
15044 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15045 SSL3_RT_APPLICATION_DATA, return 0.
15046 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15051 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15052 releases, have been re-implemented by renaming the previous
15053 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15054 and ssl3_read_internal, respectively, and adding 'peek' parameters
15055 to them. The new ssl[23]_{read,peek} functions are calls to
15056 ssl[23]_read_internal with the 'peek' flag set appropriately.
15057 A 'peek' parameter has also been added to ssl3_read_bytes, which
15058 does the actual work for ssl3_read_internal.
15062 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15063 the method-specific "init()" handler. Also clean up ex_data after
15064 calling the method-specific "finish()" handler. Previously, this was
15065 happening the other way round.
15069 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15070 The previous value, 12, was not always sufficient for BN_mod_exp().
15074 * Make sure that shared libraries get the internal name engine with
15075 the full version number and not just 0. This should mark the
15076 shared libraries as not backward compatible. Of course, this should
15077 be changed again when we can guarantee backward binary compatibility.
15081 * Fix typo in get_cert_by_subject() in by_dir.c
15083 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15085 * Rework the system to generate shared libraries:
15087 - Make note of the expected extension for the shared libraries and
15088 if there is a need for symbolic links from for example libcrypto.so.0
15089 to libcrypto.so.0.9.7. There is extended info in Configure for
15092 - Make as few rebuilds of the shared libraries as possible.
15094 - Still avoid linking the OpenSSL programs with the shared libraries.
15096 - When installing, install the shared libraries separately from the
15101 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15103 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15104 and not in SSL_clear because the latter is also used by the
15105 accept/connect functions; previously, the settings made by
15106 SSL_set_read_ahead would be lost during the handshake.
15108 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15110 * Correct util/mkdef.pl to be selective about disabled algorithms.
15111 Previously, it would create entries for disabled algorithms no
15116 * Added several new manual pages for SSL_* function.
15120 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15122 * In ssl23_get_client_hello, generate an error message when faced
15123 with an initial SSL 3.0/TLS record that is too small to contain the
15124 first two bytes of the ClientHello message, i.e. client_version.
15125 (Note that this is a pathologic case that probably has never happened
15126 in real life.) The previous approach was to use the version number
15127 from the record header as a substitute; but our protocol choice
15128 should not depend on that one because it is not authenticated
15129 by the Finished messages.
15133 * More robust randomness gathering functions for Windows.
15135 *Jeffrey Altman <jaltman@columbia.edu>*
15137 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15138 not set then we don't setup the error code for issuer check errors
15139 to avoid possibly overwriting other errors which the callback does
15140 handle. If an application does set the flag then we assume it knows
15141 what it is doing and can handle the new informational codes
15146 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15147 a general "ANY" type, as such it should be able to decode anything
15148 including tagged types. However it didn't check the class so it would
15149 wrongly interpret tagged types in the same way as their universal
15150 counterpart and unknown types were just rejected. Changed so that the
15151 tagged and unknown types are handled in the same way as a SEQUENCE:
15152 that is the encoding is stored intact. There is also a new type
15153 "V_ASN1_OTHER" which is used when the class is not universal, in this
15154 case we have no idea what the actual type is so we just lump them all
15159 * On VMS, stdout may very well lead to a file that is written to
15160 in a record-oriented fashion. That means that every write() will
15161 write a separate record, which will be read separately by the
15162 programs trying to read from it. This can be very confusing.
15164 The solution is to put a BIO filter in the way that will buffer
15165 text until a linefeed is reached, and then write everything a
15166 line at a time, so every record written will be an actual line,
15167 not chunks of lines and not (usually doesn't happen, but I've
15168 seen it once) several lines in one record. BIO_f_linebuffer() is
15171 Currently, it's a VMS-only method, because that's where it has
15172 been tested well enough.
15176 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15177 it can return incorrect results.
15178 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15179 but it was in 0.9.6-beta[12].)
15183 * Disable the check for content being present when verifying detached
15184 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15185 include zero length content when signing messages.
15189 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15190 BIO_ctrl (for BIO pairs).
15194 * Add DSO method for VMS.
15198 * Bug fix: Montgomery multiplication could produce results with the
15203 * Add RPM specification openssl.spec and modify it to build three
15204 packages. The default package contains applications, application
15205 documentation and run-time libraries. The devel package contains
15206 include files, static libraries and function documentation. The
15207 doc package contains the contents of the doc directory. The original
15208 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15212 * Add a large number of documentation files for many SSL routines.
15214 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15216 * Add a configuration entry for Sony News 4.
15218 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15220 * Don't set the two most significant bits to one when generating a
15221 random number < q in the DSA library.
15225 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15226 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15227 the underlying transport is blocking) if a handshake took place.
15228 (The default behaviour is needed by applications such as s_client
15229 and s_server that use select() to determine when to use SSL_read;
15230 but for applications that know in advance when to expect data, it
15231 just makes things more complicated.)
15235 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15240 * Add a few more EBCDIC conditionals that make `req` and `x509`
15241 work better on such systems.
15243 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15245 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15246 Update PKCS12_parse() so it copies the friendlyName and the
15247 keyid to the certificates aux info.
15251 * Fix bug in PKCS7_verify() which caused an infinite loop
15252 if there was more than one signature.
15254 *Sven Uszpelkat <su@celocom.de>*
15256 * Major change in util/mkdef.pl to include extra information
15257 about each symbol, as well as presenting variables as well
15258 as functions. This change means that there's n more need
15259 to rebuild the .num files when some algorithms are excluded.
15263 * Allow the verify time to be set by an application,
15264 rather than always using the current time.
15268 * Phase 2 verify code reorganisation. The certificate
15269 verify code now looks up an issuer certificate by a
15270 number of criteria: subject name, authority key id
15271 and key usage. It also verifies self signed certificates
15272 by the same criteria. The main comparison function is
15273 X509_check_issued() which performs these checks.
15275 Lot of changes were necessary in order to support this
15276 without completely rewriting the lookup code.
15278 Authority and subject key identifier are now cached.
15280 The LHASH 'certs' is X509_STORE has now been replaced
15281 by a STACK_OF(X509_OBJECT). This is mainly because an
15282 LHASH can't store or retrieve multiple objects with
15283 the same hash value.
15285 As a result various functions (which were all internal
15286 use only) have changed to handle the new X509_STORE
15287 structure. This will break anything that messed round
15288 with X509_STORE internally.
15290 The functions X509_STORE_add_cert() now checks for an
15291 exact match, rather than just subject name.
15293 The X509_STORE API doesn't directly support the retrieval
15294 of multiple certificates matching a given criteria, however
15295 this can be worked round by performing a lookup first
15296 (which will fill the cache with candidate certificates)
15297 and then examining the cache for matches. This is probably
15298 the best we can do without throwing out X509_LOOKUP
15299 entirely (maybe later...).
15301 The X509_VERIFY_CTX structure has been enhanced considerably.
15303 All certificate lookup operations now go via a get_issuer()
15304 callback. Although this currently uses an X509_STORE it
15305 can be replaced by custom lookups. This is a simple way
15306 to bypass the X509_STORE hackery necessary to make this
15307 work and makes it possible to use more efficient techniques
15308 in future. A very simple version which uses a simple
15309 STACK for its trusted certificate store is also provided
15310 using X509_STORE_CTX_trusted_stack().
15312 The verify_cb() and verify() callbacks now have equivalents
15313 in the X509_STORE_CTX structure.
15315 X509_STORE_CTX also has a 'flags' field which can be used
15316 to customise the verify behaviour.
15320 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15321 excludes S/MIME capabilities.
15325 * When a certificate request is read in keep a copy of the
15326 original encoding of the signed data and use it when outputting
15327 again. Signatures then use the original encoding rather than
15328 a decoded, encoded version which may cause problems if the
15329 request is improperly encoded.
15333 * For consistency with other BIO_puts implementations, call
15334 buffer_write(b, ...) directly in buffer_puts instead of calling
15337 In BIO_puts, increment b->num_write as in BIO_write.
15339 *Peter.Sylvester@EdelWeb.fr*
15341 * Fix BN_mul_word for the case where the word is 0. (We have to use
15342 BN_zero, we may not return a BIGNUM with an array consisting of
15343 words set to zero.)
15347 * Avoid calling abort() from within the library when problems are
15348 detected, except if preprocessor symbols have been defined
15349 (such as REF_CHECK, BN_DEBUG etc.).
15353 * New openssl application 'rsautl'. This utility can be
15354 used for low level RSA operations. DER public key
15355 BIO/fp routines also added.
15359 * New Configure entry and patches for compiling on QNX 4.
15361 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15363 * A demo state-machine implementation was sponsored by
15364 Nuron (<http://www.nuron.com/>) and is now available in
15365 demos/state_machine.
15369 * New options added to the 'dgst' utility for signature
15370 generation and verification.
15374 * Unrecognized PKCS#7 content types are now handled via a
15375 catch all ASN1_TYPE structure. This allows unsupported
15376 types to be stored as a "blob" and an application can
15377 encode and decode it manually.
15381 * Fix various signed/unsigned issues to make a_strex.c
15382 compile under VC++.
15384 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15386 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15387 length if passed a buffer. ASN1_INTEGER_to_BN failed
15388 if passed a NULL BN and its argument was negative.
15390 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15392 * Modification to PKCS#7 encoding routines to output definite
15393 length encoding. Since currently the whole structures are in
15394 memory there's not real point in using indefinite length
15395 constructed encoding. However if OpenSSL is compiled with
15396 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15400 * Added BIO_vprintf() and BIO_vsnprintf().
15404 * Added more prefixes to parse for in the strings written
15405 through a logging bio, to cover all the levels that are available
15406 through syslog. The prefixes are now:
15408 PANIC, EMERG, EMR => LOG_EMERG
15409 ALERT, ALR => LOG_ALERT
15410 CRIT, CRI => LOG_CRIT
15411 ERROR, ERR => LOG_ERR
15412 WARNING, WARN, WAR => LOG_WARNING
15413 NOTICE, NOTE, NOT => LOG_NOTICE
15414 INFO, INF => LOG_INFO
15415 DEBUG, DBG => LOG_DEBUG
15417 and as before, if none of those prefixes are present at the
15418 beginning of the string, LOG_ERR is chosen.
15420 On Win32, the `LOG_*` levels are mapped according to this:
15422 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15423 LOG_WARNING => EVENTLOG_WARNING_TYPE
15424 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15428 * Made it possible to reconfigure with just the configuration
15429 argument "reconf" or "reconfigure". The command line arguments
15430 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15431 and are retrieved from there when reconfiguring.
15437 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15439 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15443 * The obj_dat.pl script was messing up the sorting of object
15444 names. The reason was that it compared the quoted version
15445 of strings as a result "OCSP" > "OCSP Signing" because
15446 " > SPACE. Changed script to store unquoted versions of
15447 names and add quotes on output. It was also omitting some
15448 names from the lookup table if they were given a default
15449 value (that is if SN is missing it is given the same
15450 value as LN and vice versa), these are now added on the
15451 grounds that if an object has a name we should be able to
15452 look it up. Finally added warning output when duplicate
15453 short or long names are found.
15457 * Changes needed for Tandem NSK.
15459 *Scott Uroff <scott@xypro.com>*
15461 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15462 RSA_padding_check_SSLv23(), special padding was never detected
15463 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15464 version rollback attacks was not effective.
15466 In s23_clnt.c, don't use special rollback-attack detection padding
15467 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15468 client; similarly, in s23_srvr.c, don't do the rollback check if
15469 SSL 2.0 is the only protocol enabled in the server.
15473 * Make it possible to get hexdumps of unprintable data with 'openssl
15474 asn1parse'. By implication, the functions ASN1_parse_dump() and
15475 BIO_dump_indent() are added.
15479 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15480 these print out strings and name structures based on various
15481 flags including RFC2253 support and proper handling of
15482 multibyte characters. Added options to the 'x509' utility
15483 to allow the various flags to be set.
15487 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15488 Also change the functions X509_cmp_current_time() and
15489 X509_gmtime_adj() work with an ASN1_TIME structure,
15490 this will enable certificates using GeneralizedTime in validity
15491 dates to be checked.
15495 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15496 negative public key encodings) on by default,
15497 NO_NEG_PUBKEY_BUG can be set to disable it.
15501 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15502 content octets. An i2c_ASN1_OBJECT is unnecessary because
15503 the encoding can be trivially obtained from the structure.
15507 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15508 not read locks (`CRYPTO_r_[un]lock`).
15512 * A first attempt at creating official support for shared
15513 libraries through configuration. I've kept it so the
15514 default is static libraries only, and the OpenSSL programs
15515 are always statically linked for now, but there are
15516 preparations for dynamic linking in place.
15517 This has been tested on Linux and Tru64.
15521 * Randomness polling function for Win9x, as described in:
15522 Peter Gutmann, Software Generation of Practically Strong
15527 * Fix so PRNG is seeded in req if using an already existing
15532 * New options to smime application. -inform and -outform
15533 allow alternative formats for the S/MIME message including
15534 PEM and DER. The -content option allows the content to be
15535 specified separately. This should allow things like Netscape
15536 form signing output easier to verify.
15540 * Fix the ASN1 encoding of tags using the 'long form'.
15544 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15545 STRING types. These convert content octets to and from the
15546 underlying type. The actual tag and length octets are
15547 already assumed to have been read in and checked. These
15548 are needed because all other string types have virtually
15549 identical handling apart from the tag. By having versions
15550 of the ASN1 functions that just operate on content octets
15551 IMPLICIT tagging can be handled properly. It also allows
15552 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15553 and ASN1_INTEGER are identical apart from the tag.
15557 * Change the handling of OID objects as follows:
15559 - New object identifiers are inserted in objects.txt, following
15560 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15561 - objects.pl is used to process obj_mac.num and create a new
15563 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15566 This is currently kind of a hack, and the perl code in objects.pl
15567 isn't very elegant, but it works as I intended. The simplest way
15568 to check that it worked correctly is to look in obj_dat.h and
15569 check the array nid_objs and make sure the objects haven't moved
15570 around (this is important!). Additions are OK, as well as
15571 consistent name changes.
15575 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15579 * Addition of the command line parameter '-rand file' to 'openssl req'.
15580 The given file adds to whatever has already been seeded into the
15581 random pool through the RANDFILE configuration file option or
15582 environment variable, or the default random state file.
15586 * mkstack.pl now sorts each macro group into lexical order.
15587 Previously the output order depended on the order the files
15588 appeared in the directory, resulting in needless rewriting
15593 * Patches to make OpenSSL compile under Win32 again. Mostly
15594 work arounds for the VC++ problem that it treats func() as
15595 func(void). Also stripped out the parts of mkdef.pl that
15596 added extra typesafe functions: these no longer exist.
15600 * Reorganisation of the stack code. The macros are now all
15601 collected in safestack.h . Each macro is defined in terms of
15602 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15603 DEBUG_SAFESTACK is now handled in terms of function casts,
15604 this has the advantage of retaining type safety without the
15605 use of additional functions. If DEBUG_SAFESTACK is not defined
15606 then the non typesafe macros are used instead. Also modified the
15607 mkstack.pl script to handle the new form. Needs testing to see
15608 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15609 the default if no major problems. Similar behaviour for ASN1_SET_OF
15610 and PKCS12_STACK_OF.
15614 * When some versions of IIS use the 'NET' form of private key the
15615 key derivation algorithm is different. Normally MD5(password) is
15616 used as a 128 bit RC4 key. In the modified case
15617 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15618 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15619 as the old Netscape_RSA functions except they have an additional
15620 'sgckey' parameter which uses the modified algorithm. Also added
15621 an -sgckey command line option to the rsa utility. Thanks to
15622 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15623 algorithm to openssl-dev.
15627 * The evp_local.h macros were using 'c.##kname' which resulted in
15628 invalid expansion on some systems (SCO 5.0.5 for example).
15629 Corrected to 'c.kname'.
15631 *Phillip Porch <root@theporch.com>*
15633 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15634 a STACK of email addresses from a certificate or request, these look
15635 in the subject name and the subject alternative name extensions and
15636 omit any duplicate addresses.
15640 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15641 This makes DSA verification about 2 % faster.
15645 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15646 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15647 plus overhead for 1024 bit moduli).
15648 This makes exponentiations about 0.5 % faster for 1024 bit
15649 exponents (as measured by "openssl speed rsa2048").
15653 * Rename memory handling macros to avoid conflicts with other
15655 Malloc => OPENSSL_malloc
15656 Malloc_locked => OPENSSL_malloc_locked
15657 Realloc => OPENSSL_realloc
15658 Free => OPENSSL_free
15662 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15663 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15667 * CygWin32 support.
15669 *John Jarvie <jjarvie@newsguy.com>*
15671 * The type-safe stack code has been rejigged. It is now only compiled
15672 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15673 by default all type-specific stack functions are "#define"d back to
15674 standard stack functions. This results in more streamlined output
15675 but retains the type-safety checking possibilities of the original
15680 * The STACK code has been cleaned up, and certain type declarations
15681 that didn't make a lot of sense have been brought in line. This has
15682 also involved a cleanup of sorts in safestack.h to more correctly
15683 map type-safe stack functions onto their plain stack counterparts.
15684 This work has also resulted in a variety of "const"ifications of
15685 lots of the code, especially `_cmp` operations which should normally
15686 be prototyped with "const" parameters anyway.
15690 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15691 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15692 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15693 where all of 'md' is used each time the PRNG is used, but 'state'
15694 is used only indexed by a cyclic counter. As entropy may not be
15695 well distributed from the beginning, 'md' is important as a
15696 chaining variable. However, the output function chains only half
15697 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15698 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15699 in all of 'state' being rewritten, with the new values depending
15700 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15704 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15705 the handshake is continued after ssl_verify_cert_chain();
15706 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15707 can lead to 'unexplainable' connection aborts later.
15709 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15711 * Major EVP API cipher revision.
15712 Add hooks for extra EVP features. This allows various cipher
15713 parameters to be set in the EVP interface. Support added for variable
15714 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15715 setting of RC2 and RC5 parameters.
15717 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15720 Remove lots of duplicated code from the EVP library. For example *every*
15721 cipher init() function handles the 'iv' in the same way according to the
15722 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15723 for CFB and OFB modes they zero ctx->num.
15725 New functionality allows removal of S/MIME code RC2 hack.
15727 Most of the routines have the same form and so can be declared in terms
15730 By shifting this to the top level EVP_CipherInit() it can be removed from
15731 all individual ciphers. If the cipher wants to handle IVs or keys
15732 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15735 Change lots of functions like EVP_EncryptUpdate() to now return a
15736 value: although software versions of the algorithms cannot fail
15737 any installed hardware versions can.
15741 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15742 this option is set, tolerate broken clients that send the negotiated
15743 protocol version number instead of the requested protocol version
15748 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15749 i.e. non-zero for export ciphersuites, zero otherwise.
15750 Previous versions had this flag inverted, inconsistent with
15751 rsa_tmp_cb (..._TMP_RSA_CB).
15753 *Bodo Moeller; problem reported by Amit Chopra*
15755 * Add missing DSA library text string. Work around for some IIS
15756 key files with invalid SEQUENCE encoding.
15760 * Add a document (doc/standards.txt) that list all kinds of standards
15761 and so on that are implemented in OpenSSL.
15765 * Enhance c_rehash script. Old version would mishandle certificates
15766 with the same subject name hash and wouldn't handle CRLs at all.
15767 Added -fingerprint option to crl utility, to support new c_rehash
15772 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15776 * Fix for SSL server purpose checking. Server checking was
15777 rejecting certificates which had extended key usage present
15778 but no ssl client purpose.
15780 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15782 * Make PKCS#12 code work with no password. The PKCS#12 spec
15783 is a little unclear about how a blank password is handled.
15784 Since the password in encoded as a BMPString with terminating
15785 double NULL a zero length password would end up as just the
15786 double NULL. However no password at all is different and is
15787 handled differently in the PKCS#12 key generation code. NS
15788 treats a blank password as zero length. MSIE treats it as no
15789 password on export: but it will try both on import. We now do
15790 the same: PKCS12_parse() tries zero length and no password if
15791 the password is set to "" or NULL (NULL is now a valid password:
15792 it wasn't before) as does the pkcs12 application.
15796 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15797 perror when PEM_read_bio_X509_REQ fails, the error message must
15798 be obtained from the error queue.
15802 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15803 it in ERR_remove_state if appropriate, and change ERR_get_state
15804 accordingly to avoid race conditions (this is necessary because
15805 thread_hash is no longer constant once set).
15809 * Bugfix for linux-elf makefile.one.
15813 * RSA_get_default_method() will now cause a default
15814 RSA_METHOD to be chosen if one doesn't exist already.
15815 Previously this was only set during a call to RSA_new()
15816 or RSA_new_method(NULL) meaning it was possible for
15817 RSA_get_default_method() to return NULL.
15821 * Added native name translation to the existing DSO code
15822 that will convert (if the flag to do so is set) filenames
15823 that are sufficiently small and have no path information
15824 into a canonical native form. Eg. "blah" converted to
15825 "libblah.so" or "blah.dll" etc.
15829 * New function ERR_error_string_n(e, buf, len) which is like
15830 ERR_error_string(e, buf), but writes at most 'len' bytes
15831 including the 0 terminator. For ERR_error_string_n, 'buf'
15834 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15836 * CONF library reworked to become more general. A new CONF
15837 configuration file reader "class" is implemented as well as a
15838 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15839 old `CONF_*` functions are still there, but are reimplemented to
15840 work in terms of the new functions. Also, a set of functions
15841 to handle the internal storage of the configuration data is
15842 provided to make it easier to write new configuration file
15843 reader "classes" (I can definitely see something reading a
15844 configuration file in XML format, for example), called `_CONF_*`,
15845 or "the configuration storage API"...
15847 The new configuration file reading functions are:
15849 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15850 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15852 NCONF_default, NCONF_WIN32
15854 NCONF_dump_fp, NCONF_dump_bio
15856 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15857 NCONF_new creates a new CONF object. This works in the same way
15858 as other interfaces in OpenSSL, like the BIO interface.
15859 `NCONF_dump_*` dump the internal storage of the configuration file,
15860 which is useful for debugging. All other functions take the same
15861 arguments as the old `CONF_*` functions with the exception of the
15862 first that must be a `CONF *` instead of a `LHASH *`.
15864 To make it easier to use the new classes with the old `CONF_*` functions,
15865 the function CONF_set_default_method is provided.
15869 * Add '-tls1' option to 'openssl ciphers', which was already
15870 mentioned in the documentation but had not been implemented.
15871 (This option is not yet really useful because even the additional
15872 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15876 * Initial DSO code added into libcrypto for letting OpenSSL (and
15877 OpenSSL-based applications) load shared libraries and bind to
15878 them in a portable way.
15880 *Geoff Thorpe, with contributions from Richard Levitte*
15882 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15884 * Make sure _lrotl and _lrotr are only used with MSVC.
15886 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15887 (the default implementation of RAND_status).
15889 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15890 to '-clrext' (= clear extensions), as intended and documented.
15891 *Bodo Moeller; inconsistency pointed out by Michael Attili
15892 <attili@amaxo.com>*
15894 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15895 was larger than the MD block size.
15897 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15899 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15900 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15901 using the passed key: if the passed key was a private key the result
15902 of X509_print(), for example, would be to print out all the private key
15907 * des_quad_cksum() byte order bug fix.
15908 *Ulf Möller, using the problem description in krb4-0.9.7, where
15909 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15911 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15914 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15916 * For easily testing in shell scripts whether some command
15917 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15918 returns with exit code 0 iff no command of the given name is available.
15919 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15920 the output goes to stdout and nothing is printed to stderr.
15921 Additional arguments are always ignored.
15923 Since for each cipher there is a command of the same name,
15924 the 'no-cipher' compilation switches can be tested this way.
15926 ('openssl no-XXX' is not able to detect pseudo-commands such
15927 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15931 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15935 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15936 is set; it will be thrown away anyway because each handshake creates
15938 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15939 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15940 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15941 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15945 * New s_client option -ign_eof: EOF at stdin is ignored, and
15946 'Q' and 'R' lose their special meanings (quit/renegotiate).
15947 This is part of what -quiet does; unlike -quiet, -ign_eof
15948 does not suppress any output.
15952 * Add compatibility options to the purpose and trust code. The
15953 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15954 accepts a certificate or CA, this was the previous behaviour,
15955 with all the associated security issues.
15957 X509_TRUST_COMPAT is the old trust behaviour: only and
15958 automatically trust self signed roots in certificate store. A
15959 new trust setting X509_TRUST_DEFAULT is used to specify that
15960 a purpose has no associated trust setting and it should instead
15961 use the value in the default purpose.
15965 * Fix the PKCS#8 DSA private key code so it decodes keys again
15966 and fix a memory leak.
15970 * In util/mkerr.pl (which implements 'make errors'), preserve
15971 reason strings from the previous version of the .c file, as
15972 the default to have only downcase letters (and digits) in
15973 automatically generated reasons codes is not always appropriate.
15977 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15978 using strerror. Previously, ERR_reason_error_string() returned
15979 library names as reason strings for SYSerr; but SYSerr is a special
15980 case where small numbers are errno values, not library numbers.
15984 * Add '-dsaparam' option to 'openssl dhparam' application. This
15985 converts DSA parameters into DH parameters. (When creating parameters,
15986 DSA_generate_parameters is used.)
15990 * Include 'length' (recommended exponent length) in C code generated
15991 by 'openssl dhparam -C'.
15995 * The second argument to set_label in perlasm was already being used
15996 so couldn't be used as a "file scope" flag. Moved to third argument
16001 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16002 instead of RAND_bytes for encryption IVs and salts.
16006 * Include RAND_status() into RAND_METHOD instead of implementing
16007 it only for md_rand.c Otherwise replacing the PRNG by calling
16008 RAND_set_rand_method would be impossible.
16012 * Don't let DSA_generate_key() enter an infinite loop if the random
16013 number generation fails.
16017 * New 'rand' application for creating pseudo-random output.
16021 * Added configuration support for Linux/IA64
16023 *Rolf Haberrecker <rolf@suse.de>*
16025 * Assembler module support for Mingw32.
16029 * Shared library support for HPUX (in shlib/).
16031 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16033 * Shared library support for Solaris gcc.
16035 *Lutz Behnke <behnke@trustcenter.de>*
16037 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16039 * PKCS7_encrypt() was adding text MIME headers twice because they
16040 were added manually and by SMIME_crlf_copy().
16044 * In bntest.c don't call BN_rand with zero bits argument.
16046 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16048 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16049 case was implemented. This caused BN_div_recp() to fail occasionally.
16053 * Add an optional second argument to the set_label() in the perl
16054 assembly language builder. If this argument exists and is set
16055 to 1 it signals that the assembler should use a symbol whose
16056 scope is the entire file, not just the current function. This
16057 is needed with MASM which uses the format label:: for this scope.
16059 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16061 * Change the ASN1 types so they are typedefs by default. Before
16062 almost all types were #define'd to ASN1_STRING which was causing
16063 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16068 * Change names of new functions to the new get1/get0 naming
16069 convention: After 'get1', the caller owns a reference count
16070 and has to call `..._free`; 'get0' returns a pointer to some
16071 data structure without incrementing reference counters.
16072 (Some of the existing 'get' functions increment a reference
16073 counter, some don't.)
16074 Similarly, 'set1' and 'add1' functions increase reference
16075 counters or duplicate objects.
16079 * Allow for the possibility of temp RSA key generation failure:
16080 the code used to assume it always worked and crashed on failure.
16084 * Fix potential buffer overrun problem in BIO_printf().
16085 *Ulf Möller, using public domain code by Patrick Powell; problem
16086 pointed out by David Sacerdote <das33@cornell.edu>*
16088 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16089 RAND_egd() and RAND_status(). In the command line application,
16090 the EGD socket can be specified like a seed file using RANDFILE
16095 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16096 Some CAs (e.g. Verisign) distribute certificates in this form.
16100 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16101 list to exclude them. This means that no special compilation option
16102 is needed to use anonymous DH: it just needs to be included in the
16107 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16108 EVP_MD_type. The old functionality is available in a new macro called
16109 EVP_MD_md(). Change code that uses it and update docs.
16113 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16114 where the `void *` argument is replaced by a function pointer argument.
16115 Previously `void *` was abused to point to functions, which works on
16116 many platforms, but is not correct. As these functions are usually
16117 called by macros defined in OpenSSL header files, most source code
16118 should work without changes.
16122 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16123 sections with information on -D... compiler switches used for
16124 compiling the library so that applications can see them. To enable
16125 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16126 must be defined. E.g.,
16127 #define OPENSSL_ALGORITHM_DEFINES
16128 #include <openssl/opensslconf.h>
16129 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16131 *Richard Levitte, Ulf and Bodo Möller*
16133 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16138 * Change the 'other' type in certificate aux info to a STACK_OF
16139 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16140 the required ASN1 format: arbitrary types determined by an OID.
16144 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16145 argument to 'req'. This is not because the function is newer or
16146 better than others it just uses the work 'NEW' in the certificate
16147 request header lines. Some software needs this.
16151 * Reorganise password command line arguments: now passwords can be
16152 obtained from various sources. Delete the PEM_cb function and make
16153 it the default behaviour: i.e. if the callback is NULL and the
16154 usrdata argument is not NULL interpret it as a null terminated pass
16155 phrase. If usrdata and the callback are NULL then the pass phrase
16156 is prompted for as usual.
16160 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16161 the support is automatically enabled. The resulting binaries will
16162 autodetect the card and use it if present.
16164 *Ben Laurie and Compaq Inc.*
16166 * Work around for Netscape hang bug. This sends certificate request
16167 and server done in one record. Since this is perfectly legal in the
16168 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16169 the bugs/SSLv3 entry for more info.
16173 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16177 * Add -rand argument to smime and pkcs12 applications and read/write
16182 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16186 * Add command line password options to the remaining applications.
16190 * Bug fix for BN_div_recp() for numerators with an even number of
16195 * More tests in bntest.c, and changed test_bn output.
16199 * ./config recognizes MacOS X now.
16203 * Bug fix for BN_div() when the first words of num and divisor are
16204 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16208 * Add support for various broken PKCS#8 formats, and command line
16209 options to produce them.
16213 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16214 get temporary BIGNUMs from a BN_CTX.
16218 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16223 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16224 include a #define from the old name to the new. The original intent
16225 was that statically linked binaries could for example just call
16226 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16227 link with digests. This never worked because SSLeay_add_all_digests()
16228 and SSLeay_add_all_ciphers() were in the same source file so calling
16229 one would link with the other. They are now in separate source files.
16233 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16237 * Use a less unusual form of the Miller-Rabin primality test (it used
16238 a binary algorithm for exponentiation integrated into the Miller-Rabin
16239 loop, our standard modexp algorithms are faster).
16243 * Support for the EBCDIC character set completed.
16245 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16247 * Source code cleanups: use const where appropriate, eliminate casts,
16248 use `void *` instead of `char *` in lhash.
16252 * Bugfix: ssl3_send_server_key_exchange was not restartable
16253 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16254 this the server could overwrite ephemeral keys that the client
16259 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16260 using 50 iterations of the Rabin-Miller test.
16262 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16263 iterations of the Rabin-Miller test as required by the appendix
16264 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16265 As BN_is_prime_fasttest includes trial division, DSA parameter
16266 generation becomes much faster.
16268 This implies a change for the callback functions in DSA_is_prime
16269 and DSA_generate_parameters: The callback function is called once
16270 for each positive witness in the Rabin-Miller test, not just
16271 occasionally in the inner loop; and the parameters to the
16272 callback function now provide an iteration count for the outer
16273 loop rather than for the current invocation of the inner loop.
16274 DSA_generate_parameters additionally can call the callback
16275 function with an 'iteration count' of -1, meaning that a
16276 candidate has passed the trial division test (when q is generated
16277 from an application-provided seed, trial division is skipped).
16281 * New function BN_is_prime_fasttest that optionally does trial
16282 division before starting the Rabin-Miller test and has
16283 an additional BN_CTX * argument (whereas BN_is_prime always
16284 has to allocate at least one BN_CTX).
16285 'callback(1, -1, cb_arg)' is called when a number has passed the
16286 trial division stage.
16290 * Fix for bug in CRL encoding. The validity dates weren't being handled
16295 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16299 * New function BN_pseudo_rand().
16303 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16304 bignum version of BN_from_montgomery() with the working code from
16305 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16310 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16311 made it impossible to use the same SSL_SESSION data structure in
16312 SSL2 clients in multiple threads.
16316 * The return value of RAND_load_file() no longer counts bytes obtained
16317 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16318 to seed the PRNG (previously an explicit byte count was required).
16320 *Ulf Möller, Bodo Möller*
16322 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16323 used `char *` instead of `void *` and had casts all over the place.
16327 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16331 * Retain source code compatibility for BN_prime_checks macro:
16332 BN_is_prime(..., BN_prime_checks, ...) now uses
16333 BN_prime_checks_for_size to determine the appropriate number of
16334 Rabin-Miller iterations.
16338 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16339 DH_CHECK_P_NOT_SAFE_PRIME.
16340 (Check if this is true? OpenPGP calls them "strong".)
16344 * Merge the functionality of "dh" and "gendh" programs into a new program
16345 "dhparam". The old programs are retained for now but will handle DH keys
16346 (instead of parameters) in future.
16350 * Make the ciphers, s_server and s_client programs check the return values
16351 when a new cipher list is set.
16355 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16356 ciphers. Before when the 56bit ciphers were enabled the sorting was
16359 The syntax for the cipher sorting has been extended to support sorting by
16360 cipher-strength (using the strength_bits hard coded in the tables).
16361 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16363 Fix a bug in the cipher-command parser: when supplying a cipher command
16364 string with an "undefined" symbol (neither command nor alphanumeric
16365 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16366 an error is flagged.
16368 Due to the strength-sorting extension, the code of the
16369 ssl_create_cipher_list() function was completely rearranged. I hope that
16370 the readability was also increased :-)
16372 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16374 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16375 for the first serial number and places 2 in the serial number file. This
16376 avoids problems when the root CA is created with serial number zero and
16377 the first user certificate has the same issuer name and serial number
16382 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16383 the new code. Add documentation for this stuff.
16387 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16388 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16389 structures and behave in an analogous way to the X509v3 functions:
16390 they shouldn't be called directly but wrapper functions should be used
16393 So we also now have some wrapper functions that call the X509at functions
16394 when passed certificate requests. (TO DO: similar things can be done with
16395 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16396 things. Some of these need some d2i or i2d and print functionality
16397 because they handle more complex structures.)
16401 * Add missing #ifndefs that caused missing symbols when building libssl
16402 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16403 NO_RSA in `ssl/s2*.c`.
16405 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16407 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16408 has a return value which indicates the quality of the random data
16409 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16410 error queue. New function RAND_pseudo_bytes() generates output that is
16411 guaranteed to be unique but not unpredictable. RAND_add is like
16412 RAND_seed, but takes an extra argument for an entropy estimate
16413 (RAND_seed always assumes full entropy).
16417 * Do more iterations of Rabin-Miller probable prime test (specifically,
16418 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16419 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16420 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16421 false-positive rate of at most 2^-80 for random input.
16425 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16429 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16430 in the 0.9.5 release), this returns the chain
16431 from an X509_CTX structure with a dup of the stack and all
16432 the X509 reference counts upped: so the stack will exist
16433 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16436 Also make SSL_SESSION_print() print out the verify return
16441 * Add manpage for the pkcs12 command. Also change the default
16442 behaviour so MAC iteration counts are used unless the new
16443 -nomaciter option is used. This improves file security and
16444 only older versions of MSIE (4.0 for example) need it.
16448 * Honor the no-xxx Configure options when creating .DEF files.
16452 * Add PKCS#10 attributes to field table: challengePassword,
16453 unstructuredName and unstructuredAddress. These are taken from
16454 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16455 international characters are used.
16457 More changes to X509_ATTRIBUTE code: allow the setting of types
16458 based on strings. Remove the 'loc' parameter when adding
16459 attributes because these will be a SET OF encoding which is sorted
16464 * Initial changes to the 'req' utility to allow request generation
16465 automation. This will allow an application to just generate a template
16466 file containing all the field values and have req construct the
16469 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16470 used all over the place including certificate requests and PKCS#7
16471 structures. They are currently handled manually where necessary with
16472 some primitive wrappers for PKCS#7. The new functions behave in a
16473 manner analogous to the X509 extension functions: they allow
16474 attributes to be looked up by NID and added.
16476 Later something similar to the X509V3 code would be desirable to
16477 automatically handle the encoding, decoding and printing of the
16478 more complex types. The string types like challengePassword can
16479 be handled by the string table functions.
16481 Also modified the multi byte string table handling. Now there is
16482 a 'global mask' which masks out certain types. The table itself
16483 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16484 is useful when for example there is only one permissible type
16485 (as in countryName) and using the mask might result in no valid
16490 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16491 SSL_get_peer_finished to allow applications to obtain the latest
16492 Finished messages sent to the peer or expected from the peer,
16493 respectively. (SSL_get_peer_finished is usually the Finished message
16494 actually received from the peer, otherwise the protocol will be aborted.)
16496 As the Finished message are message digests of the complete handshake
16497 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16498 be used for external authentication procedures when the authentication
16499 provided by SSL/TLS is not desired or is not enough.
16503 * Enhanced support for Alpha Linux is added. Now ./config checks if
16504 the host supports BWX extension and if Compaq C is present on the
16505 $PATH. Just exploiting of the BWX extension results in 20-30%
16506 performance kick for some algorithms, e.g. DES and RC4 to mention
16507 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16512 * Add support for MS "fast SGC". This is arguably a violation of the
16513 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16514 weak crypto and after checking the certificate is SGC a second one
16515 with strong crypto. MS SGC stops the first handshake after receiving
16516 the server certificate message and sends a second client hello. Since
16517 a server will typically do all the time consuming operations before
16518 expecting any further messages from the client (server key exchange
16519 is the most expensive) there is little difference between the two.
16521 To get OpenSSL to support MS SGC we have to permit a second client
16522 hello message after we have sent server done. In addition we have to
16523 reset the MAC if we do get this second client hello.
16527 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16528 if a DER encoded private key is RSA or DSA traditional format. Changed
16529 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16530 format DER encoded private key. Newer code should use PKCS#8 format which
16531 has the key type encoded in the ASN1 structure. Added DER private key
16532 support to pkcs8 application.
16536 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16537 ciphersuites has been selected (as required by the SSL 3/TLS 1
16538 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16539 is set, we interpret this as a request to violate the specification
16540 (the worst that can happen is a handshake failure, and 'correct'
16541 behaviour would result in a handshake failure anyway).
16545 * In SSL_CTX_add_session, take into account that there might be multiple
16546 SSL_SESSION structures with the same session ID (e.g. when two threads
16547 concurrently obtain them from an external cache).
16548 The internal cache can handle only one SSL_SESSION with a given ID,
16549 so if there's a conflict, we now throw out the old one to achieve
16554 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16555 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16556 some routines that use cipher OIDs: some ciphers do not have OIDs
16557 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16562 * Simplify the trust setting structure and code. Now we just have
16563 two sequences of OIDs for trusted and rejected settings. These will
16564 typically have values the same as the extended key usage extension
16565 and any application specific purposes.
16567 The trust checking code now has a default behaviour: it will just
16568 check for an object with the same NID as the passed id. Functions can
16569 be provided to override either the default behaviour or the behaviour
16570 for a given id. SSL client, server and email already have functions
16571 in place for compatibility: they check the NID and also return "trusted"
16572 if the certificate is self signed.
16576 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16577 traditional format into an EVP_PKEY structure.
16581 * Add a password callback function PEM_cb() which either prompts for
16582 a password if usr_data is NULL or otherwise assumes it is a null
16583 terminated password. Allow passwords to be passed on command line
16584 environment or config files in a few more utilities.
16588 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16589 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16590 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16591 Update documentation.
16595 * Support for ASN1 "NULL" type. This could be handled before by using
16596 ASN1_TYPE but there wasn't any function that would try to read a NULL
16597 and produce an error if it couldn't. For compatibility we also have
16598 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16599 don't allocate anything because they don't need to.
16603 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16606 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16608 * Rebuild of the memory allocation routines used by OpenSSL code and
16609 possibly others as well. The purpose is to make an interface that
16610 provide hooks so anyone can build a separate set of allocation and
16611 deallocation routines to be used by OpenSSL, for example memory
16612 pool implementations, or something else, which was previously hard
16613 since Malloc(), Realloc() and Free() were defined as macros having
16614 the values malloc, realloc and free, respectively (except for Win32
16615 compilations). The same is provided for memory debugging code.
16616 OpenSSL already comes with functionality to find memory leaks, but
16617 this gives people a chance to debug other memory problems.
16619 With these changes, a new set of functions and macros have appeared:
16621 CRYPTO_set_mem_debug_functions() [F]
16622 CRYPTO_get_mem_debug_functions() [F]
16623 CRYPTO_dbg_set_options() [F]
16624 CRYPTO_dbg_get_options() [F]
16625 CRYPTO_malloc_debug_init() [M]
16627 The memory debug functions are NULL by default, unless the library
16628 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16629 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16630 gives the standard debugging functions that come with OpenSSL) or
16631 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16632 provided by the library user) must be used. When the standard
16633 debugging functions are used, CRYPTO_dbg_set_options can be used to
16634 request additional information:
16635 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16636 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16638 Also, things like CRYPTO_set_mem_functions will always give the
16639 expected result (the new set of functions is used for allocation
16640 and deallocation) at all times, regardless of platform and compiler
16643 To finish it up, some functions that were never use in any other
16644 way than through macros have a new API and new semantic:
16646 CRYPTO_dbg_malloc()
16647 CRYPTO_dbg_realloc()
16650 All macros of value have retained their old syntax.
16652 *Richard Levitte and Bodo Moeller*
16654 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16655 ordering of SMIMECapabilities wasn't in "strength order" and there
16656 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16661 * Some ASN1 types with illegal zero length encoding (INTEGER,
16662 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16664 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16666 * Merge in my S/MIME library for OpenSSL. This provides a simple
16667 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16668 functionality to handle multipart/signed properly) and a utility
16669 called 'smime' to call all this stuff. This is based on code I
16670 originally wrote for Celo who have kindly allowed it to be
16671 included in OpenSSL.
16675 * Add variants des_set_key_checked and des_set_key_unchecked of
16676 des_set_key (aka des_key_sched). Global variable des_check_key
16677 decides which of these is called by des_set_key; this way
16678 des_check_key behaves as it always did, but applications and
16679 the library itself, which was buggy for des_check_key == 1,
16680 have a cleaner way to pick the version they need.
16684 * New function PKCS12_newpass() which changes the password of a
16689 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16690 dynamic mix. In both cases the ids can be used as an index into the
16691 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16692 functions so they accept a list of the field values and the
16693 application doesn't need to directly manipulate the X509_TRUST
16698 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16703 * Modify the way the V3 extension code looks up extensions. This now
16704 works in a similar way to the object code: we have some "standard"
16705 extensions in a static table which is searched with OBJ_bsearch()
16706 and the application can add dynamic ones if needed. The file
16707 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16708 updated whenever a new extension is added to the core code and kept
16709 in ext_nid order. There is a simple program 'tabtest.c' which checks
16710 this. New extensions are not added too often so this file can readily
16711 be maintained manually.
16713 There are two big advantages in doing things this way. The extensions
16714 can be looked up immediately and no longer need to be "added" using
16715 X509V3_add_standard_extensions(): this function now does nothing.
16716 Side note: I get *lots* of email saying the extension code doesn't
16717 work because people forget to call this function.
16718 Also no dynamic allocation is done unless new extensions are added:
16719 so if we don't add custom extensions there is no need to call
16720 X509V3_EXT_cleanup().
16724 * Modify enc utility's salting as follows: make salting the default. Add a
16725 magic header, so unsalted files fail gracefully instead of just decrypting
16726 to garbage. This is because not salting is a big security hole, so people
16727 should be discouraged from doing it.
16731 * Fixes and enhancements to the 'x509' utility. It allowed a message
16732 digest to be passed on the command line but it only used this
16733 parameter when signing a certificate. Modified so all relevant
16734 operations are affected by the digest parameter including the
16735 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16736 DSA key was used because it didn't fix the digest.
16740 * Initial certificate chain verify code. Currently tests the untrusted
16741 certificates for consistency with the verify purpose (which is set
16742 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16744 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16745 this is because it will reject chains with invalid extensions whereas
16746 every previous version of OpenSSL and SSLeay made no checks at all.
16748 Trust code: checks the root CA for the relevant trust settings. Trust
16749 settings have an initial value consistent with the verify purpose: e.g.
16750 if the verify purpose is for SSL client use it expects the CA to be
16751 trusted for SSL client use. However the default value can be changed to
16752 permit custom trust settings: one example of this would be to only trust
16753 certificates from a specific "secure" set of CAs.
16755 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16756 which should be used for version portability: especially since the
16757 verify structure is likely to change more often now.
16759 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16760 to set them. If not set then assume SSL clients will verify SSL servers
16763 Two new options to the verify program: -untrusted allows a set of
16764 untrusted certificates to be passed in and -purpose which sets the
16765 intended purpose of the certificate. If a purpose is set then the
16766 new chain verify code is used to check extension consistency.
16770 * Support for the authority information access extension.
16774 * Modify RSA and DSA PEM read routines to transparently handle
16775 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16776 public keys in a format compatible with certificate
16777 SubjectPublicKeyInfo structures. Unfortunately there were already
16778 functions called *_PublicKey_* which used various odd formats so
16779 these are retained for compatibility: however the DSA variants were
16780 never in a public release so they have been deleted. Changed dsa/rsa
16781 utilities to handle the new format: note no releases ever handled public
16782 keys so we should be OK.
16784 The primary motivation for this change is to avoid the same fiasco
16785 that dogs private keys: there are several incompatible private key
16786 formats some of which are standard and some OpenSSL specific and
16787 require various evil hacks to allow partial transparent handling and
16788 even then it doesn't work with DER formats. Given the option anything
16789 other than PKCS#8 should be dumped: but the other formats have to
16790 stay in the name of compatibility.
16792 With public keys and the benefit of hindsight one standard format
16793 is used which works with EVP_PKEY, RSA or DSA structures: though
16794 it clearly returns an error if you try to read the wrong kind of key.
16796 Added a -pubkey option to the 'x509' utility to output the public key.
16797 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16798 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16799 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16800 that do the same as the `EVP_PKEY_assign_*()` except they up the
16801 reference count of the added key (they don't "swallow" the
16806 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16807 CRLs would fail if the file contained no certificates or no CRLs:
16808 added a new function to read in both types and return the number
16809 read: this means that if none are read it will be an error. The
16810 DER versions of the certificate and CRL reader would always fail
16811 because it isn't possible to mix certificates and CRLs in DER format
16812 without choking one or the other routine. Changed this to just read
16813 a certificate: this is the best we can do. Also modified the code
16814 in `apps/verify.c` to take notice of return codes: it was previously
16815 attempting to read in certificates from NULL pointers and ignoring
16816 any errors: this is one reason why the cert and CRL reader seemed
16817 to work. It doesn't check return codes from the default certificate
16818 routines: these may well fail if the certificates aren't installed.
16822 * Code to support otherName option in GeneralName.
16826 * First update to verify code. Change the verify utility
16827 so it warns if it is passed a self signed certificate:
16828 for consistency with the normal behaviour. X509_verify
16829 has been modified to it will now verify a self signed
16830 certificate if *exactly* the same certificate appears
16831 in the store: it was previously impossible to trust a
16832 single self signed certificate. This means that:
16833 openssl verify ss.pem
16834 now gives a warning about a self signed certificate but
16835 openssl verify -CAfile ss.pem ss.pem
16840 * For servers, store verify_result in SSL_SESSION data structure
16841 (and add it to external session representation).
16842 This is needed when client certificate verifications fails,
16843 but an application-provided verification callback (set by
16844 SSL_CTX_set_cert_verify_callback) allows accepting the session
16845 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16846 but returns 1): When the session is reused, we have to set
16847 ssl->verify_result to the appropriate error code to avoid
16850 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16852 * Fix a bug in the new PKCS#7 code: it didn't consider the
16853 case in PKCS7_dataInit() where the signed PKCS7 structure
16854 didn't contain any existing data because it was being created.
16856 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16858 * Add a salt to the key derivation routines in enc.c. This
16859 forms the first 8 bytes of the encrypted file. Also add a
16860 -S option to allow a salt to be input on the command line.
16864 * New function X509_cmp(). Oddly enough there wasn't a function
16865 to compare two certificates. We do this by working out the SHA1
16866 hash and comparing that. X509_cmp() will be needed by the trust
16871 * SSL_get1_session() is like SSL_get_session(), but increments
16872 the reference count in the SSL_SESSION returned.
16874 *Geoff Thorpe <geoff@eu.c2.net>*
16876 * Fix for 'req': it was adding a null to request attributes.
16877 Also change the X509_LOOKUP and X509_INFO code to handle
16878 certificate auxiliary information.
16882 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16887 * Add the possibility to add extra information to the memory leak
16888 detecting output, to form tracebacks, showing from where each
16889 allocation was originated: CRYPTO_push_info("constant string") adds
16890 the string plus current file name and line number to a per-thread
16891 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16892 is like calling CYRPTO_pop_info() until the stack is empty.
16893 Also updated memory leak detection code to be multi-thread-safe.
16897 * Add options -text and -noout to pkcs7 utility and delete the
16898 encryption options which never did anything. Update docs.
16902 * Add options to some of the utilities to allow the pass phrase
16903 to be included on either the command line (not recommended on
16904 OSes like Unix) or read from the environment. Update the
16905 manpages and fix a few bugs.
16909 * Add a few manpages for some of the openssl commands.
16913 * Fix the -revoke option in ca. It was freeing up memory twice,
16914 leaking and not finding already revoked certificates.
16918 * Extensive changes to support certificate auxiliary information.
16919 This involves the use of X509_CERT_AUX structure and X509_AUX
16920 functions. An X509_AUX function such as PEM_read_X509_AUX()
16921 can still read in a certificate file in the usual way but it
16922 will also read in any additional "auxiliary information". By
16923 doing things this way a fair degree of compatibility can be
16924 retained: existing certificates can have this information added
16925 using the new 'x509' options.
16927 Current auxiliary information includes an "alias" and some trust
16928 settings. The trust settings will ultimately be used in enhanced
16929 certificate chain verification routines: currently a certificate
16930 can only be trusted if it is self signed and then it is trusted
16935 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16936 The problem was that one of the replacement routines had not been working
16937 since SSLeay releases. For now the offending routine has been replaced
16938 with non-optimised assembler. Even so, this now gives around 95%
16939 performance improvement for 1024 bit RSA signs.
16943 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16944 handling. Most clients have the effective key size in bits equal to
16945 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16946 A few however don't do this and instead use the size of the decrypted key
16947 to determine the RC2 key length and the AlgorithmIdentifier to determine
16948 the effective key length. In this case the effective key length can still
16949 be 40 bits but the key length can be 168 bits for example. This is fixed
16950 by manually forcing an RC2 key into the EVP_PKEY structure because the
16951 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16952 the key length and effective key length are equal.
16956 * Add a bunch of functions that should simplify the creation of
16957 X509_NAME structures. Now you should be able to do:
16958 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16959 and have it automatically work out the correct field type and fill in
16960 the structures. The more adventurous can try:
16961 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16962 and it will (hopefully) work out the correct multibyte encoding.
16966 * Change the 'req' utility to use the new field handling and multibyte
16967 copy routines. Before the DN field creation was handled in an ad hoc
16968 way in req, ca, and x509 which was rather broken and didn't support
16969 BMPStrings or UTF8Strings. Since some software doesn't implement
16970 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16971 using the dirstring_type option. See the new comment in the default
16972 openssl.cnf for more info.
16976 * Make crypto/rand/md_rand.c more robust:
16977 - Assure unique random numbers after fork().
16978 - Make sure that concurrent threads access the global counter and
16979 md serializably so that we never lose entropy in them
16980 or use exactly the same state in multiple threads.
16981 Access to the large state is not always serializable because
16982 the additional locking could be a performance killer, and
16983 md should be large enough anyway.
16987 * New file `apps/app_rand.c` with commonly needed functionality
16988 for handling the random seed file.
16990 Use the random seed file in some applications that previously did not:
16992 dsaparam -genkey (which also ignored its '-rand' option),
16995 x509 (when signing).
16996 Except on systems with /dev/urandom, it is crucial to have a random
16997 seed file at least for key creation, DSA signing, and for DH exchanges;
16998 for RSA signatures we could do without one.
17000 gendh and gendsa (unlike genrsa) used to read only the first byte
17001 of each file listed in the '-rand' option. The function as previously
17002 found in genrsa is now in app_rand.c and is used by all programs
17003 that support '-rand'.
17007 * In RAND_write_file, use mode 0600 for creating files;
17008 don't just chmod when it may be too late.
17012 * Report an error from X509_STORE_load_locations
17013 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17017 * New function ASN1_mbstring_copy() this copies a string in either
17018 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17019 into an ASN1_STRING type. A mask of permissible types is passed
17020 and it chooses the "minimal" type to use or an error if not type
17025 * Add function equivalents to the various macros in asn1.h. The old
17026 macros are retained with an `M_` prefix. Code inside the library can
17027 use the `M_` macros. External code (including the openssl utility)
17028 should *NOT* in order to be "shared library friendly".
17032 * Add various functions that can check a certificate's extensions
17033 to see if it usable for various purposes such as SSL client,
17034 server or S/MIME and CAs of these types. This is currently
17035 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17036 verification. Also added a -purpose flag to x509 utility to
17037 print out all the purposes.
17041 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17046 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17047 for, obtain and decode and extension and obtain its critical flag.
17048 This allows all the necessary extension code to be handled in a
17049 single function call.
17053 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17054 platforms. See crypto/rc4/rc4_enc.c for further details.
17058 * New -noout option to asn1parse. This causes no output to be produced
17059 its main use is when combined with -strparse and -out to extract data
17060 from a file (which may not be in ASN.1 format).
17064 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17065 when producing the local key id.
17067 *Richard Levitte <levitte@stacken.kth.se>*
17069 * New option -dhparam in s_server. This allows a DH parameter file to be
17070 stated explicitly. If it is not stated then it tries the first server
17071 certificate file. The previous behaviour hard coded the filename
17076 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17077 a public key to be input or output. For example:
17078 openssl rsa -in key.pem -pubout -out pubkey.pem
17079 Also added necessary DSA public key functions to handle this.
17083 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17084 in the message. This was handled by allowing
17085 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17087 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17089 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17090 to the end of the strings whereas this didn't. This would cause problems
17091 if strings read with d2i_ASN1_bytes() were later modified.
17093 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17095 * Fix for base64 decode bug. When a base64 bio reads only one line of
17096 data and it contains EOF it will end up returning an error. This is
17097 caused by input 46 bytes long. The cause is due to the way base64
17098 BIOs find the start of base64 encoded data. They do this by trying a
17099 trial decode on each line until they find one that works. When they
17100 do a flag is set and it starts again knowing it can pass all the
17101 data directly through the decoder. Unfortunately it doesn't reset
17102 the context it uses. This means that if EOF is reached an attempt
17103 is made to pass two EOFs through the context and this causes the
17104 resulting error. This can also cause other problems as well. As is
17105 usual with these problems it takes *ages* to find and the fix is
17106 trivial: move one line.
17108 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17110 * Ugly workaround to get s_client and s_server working under Windows. The
17111 old code wouldn't work because it needed to select() on sockets and the
17112 tty (for keypresses and to see if data could be written). Win32 only
17113 supports select() on sockets so we select() with a 1s timeout on the
17114 sockets and then see if any characters are waiting to be read, if none
17115 are present then we retry, we also assume we can always write data to
17116 the tty. This isn't nice because the code then blocks until we've
17117 received a complete line of data and it is effectively polling the
17118 keyboard at 1s intervals: however it's quite a bit better than not
17119 working at all :-) A dedicated Windows application might handle this
17120 with an event loop for example.
17124 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17125 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17126 will be called when RSA_sign() and RSA_verify() are used. This is useful
17127 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17128 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17129 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17130 This necessitated the support of an extra signature type NID_md5_sha1
17131 for SSL signatures and modifications to the SSL library to use it instead
17132 of calling RSA_public_decrypt() and RSA_private_encrypt().
17136 * Add new -verify -CAfile and -CApath options to the crl program, these
17137 will lookup a CRL issuers certificate and verify the signature in a
17138 similar way to the verify program. Tidy up the crl program so it
17139 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17140 less strict. It will now permit CRL extensions even if it is not
17141 a V2 CRL: this will allow it to tolerate some broken CRLs.
17145 * Initialize all non-automatic variables each time one of the openssl
17146 sub-programs is started (this is necessary as they may be started
17147 multiple times from the "OpenSSL>" prompt).
17149 *Lennart Bang, Bodo Moeller*
17151 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17152 removing all other RSA functionality (this is what NO_RSA does). This
17153 is so (for example) those in the US can disable those operations covered
17154 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17159 * Non-copying interface to BIO pairs.
17160 (still largely untested)
17164 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17165 ASCII string. This was handled independently in various places before.
17169 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17170 UTF8 strings a character at a time.
17174 * Use client_version from client hello to select the protocol
17175 (s23_srvr.c) and for RSA client key exchange verification
17176 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17180 * Add various utility functions to handle SPKACs, these were previously
17181 handled by poking round in the structure internals. Added new function
17182 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17183 print, verify and generate SPKACs. Based on an original idea from
17184 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17188 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17192 * Allow the config file extension section to be overwritten on the
17193 command line. Based on an original idea from Massimiliano Pala
17194 <madwolf@comune.modena.it>. The new option is called -extensions
17195 and can be applied to ca, req and x509. Also -reqexts to override
17196 the request extensions in req and -crlexts to override the crl extensions
17201 * Add new feature to the SPKAC handling in ca. Now you can include
17202 the same field multiple times by preceding it by "XXXX." for example:
17205 this is the same syntax as used in the req config file.
17209 * Allow certificate extensions to be added to certificate requests. These
17210 are specified in a 'req_extensions' option of the req section of the
17211 config file. They can be printed out with the -text option to req but
17212 are otherwise ignored at present.
17216 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17217 data read consists of only the final block it would not decrypted because
17218 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17219 A misplaced 'break' also meant the decrypted final block might not be
17220 copied until the next read.
17224 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17225 a few extra parameters to the DH structure: these will be useful if
17226 for example we want the value of 'q' or implement X9.42 DH.
17230 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17231 provides hooks that allow the default DSA functions or functions on a
17232 "per key" basis to be replaced. This allows hardware acceleration and
17233 hardware key storage to be handled without major modification to the
17234 library. Also added low level modexp hooks and CRYPTO_EX structure and
17235 associated functions.
17239 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17240 as "read only": it can't be written to and the buffer it points to will
17241 not be freed. Reading from a read only BIO is much more efficient than
17242 a normal memory BIO. This was added because there are several times when
17243 an area of memory needs to be read from a BIO. The previous method was
17244 to create a memory BIO and write the data to it, this results in two
17245 copies of the data and an O(n^2) reading algorithm. There is a new
17246 function BIO_new_mem_buf() which creates a read only memory BIO from
17247 an area of memory. Also modified the PKCS#7 routines to use read only
17252 * Bugfix: ssl23_get_client_hello did not work properly when called in
17253 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17254 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17255 but a retry condition occurred while trying to read the rest.
17259 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17260 NID_pkcs7_encrypted by default: this was wrong since this should almost
17261 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17262 the encrypted data type: this is a more sensible place to put it and it
17263 allows the PKCS#12 code to be tidied up that duplicated this
17268 * Changed obj_dat.pl script so it takes its input and output files on
17269 the command line. This should avoid shell escape redirection problems
17274 * Initial support for certificate extension requests, these are included
17275 in things like Xenroll certificate requests. Included functions to allow
17276 extensions to be obtained and added.
17280 * -crlf option to s_client and s_server for sending newlines as
17281 CRLF (as required by many protocols).
17285 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17287 * Install libRSAglue.a when OpenSSL is built with RSAref.
17289 *Ralf S. Engelschall*
17291 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17293 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17295 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17300 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17301 DH parameters/keys (q is lost during that conversion, but the resulting
17302 DH parameters contain its length).
17304 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17305 much faster than DH_generate_parameters (which creates parameters
17306 where `p = 2*q + 1`), and also the smaller q makes DH computations
17307 much more efficient (160-bit exponentiation instead of 1024-bit
17308 exponentiation); so this provides a convenient way to support DHE
17309 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17310 utter importance to use
17311 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17313 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17314 when such DH parameters are used, because otherwise small subgroup
17315 attacks may become possible!
17319 * Avoid memory leak in i2d_DHparams.
17323 * Allow the -k option to be used more than once in the enc program:
17324 this allows the same encrypted message to be read by multiple recipients.
17328 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17329 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17330 it will always use the numerical form of the OID, even if it has a short
17335 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17336 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17337 otherwise bn_mod_exp was called. In the case of hardware keys for example
17338 no private key components need be present and it might store extra data
17339 in the RSA structure, which cannot be accessed from bn_mod_exp.
17340 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17341 private key operations.
17345 * Added support for SPARC Linux.
17349 * pem_password_cb function type incompatibly changed from
17350 typedef int pem_password_cb(char *buf, int size, int rwflag);
17352 ....(char *buf, int size, int rwflag, void *userdata);
17353 so that applications can pass data to their callbacks:
17354 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17355 additional void * argument, which is just handed through whenever
17356 the password callback is called.
17358 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17360 New function SSL_CTX_set_default_passwd_cb_userdata.
17362 Compatibility note: As many C implementations push function arguments
17363 onto the stack in reverse order, the new library version is likely to
17364 interoperate with programs that have been compiled with the old
17365 pem_password_cb definition (PEM_whatever takes some data that
17366 happens to be on the stack as its last argument, and the callback
17367 just ignores this garbage); but there is no guarantee whatsoever that
17370 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17371 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17372 problems not only on Windows, but also on some Unix platforms.
17373 To avoid problematic command lines, these definitions are now in an
17374 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17375 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17379 * MIPS III/IV assembler module is reimplemented.
17383 * More DES library cleanups: remove references to srand/rand and
17384 delete an unused file.
17388 * Add support for the free Netwide assembler (NASM) under Win32,
17389 since not many people have MASM (ml) and it can be hard to obtain.
17390 This is currently experimental but it seems to work OK and pass all
17391 the tests. Check out INSTALL.W32 for info.
17395 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17396 without temporary keys kept an extra copy of the server key,
17397 and connections with temporary keys did not free everything in case
17402 * New function RSA_check_key and new openssl rsa option -check
17403 for verifying the consistency of RSA keys.
17405 *Ulf Moeller, Bodo Moeller*
17407 * Various changes to make Win32 compile work:
17408 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17409 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17410 comparison" warnings.
17411 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17415 * Add a debugging option to PKCS#5 v2 key generation function: when
17416 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17417 derived keys are printed to stderr.
17421 * Copy the flags in ASN1_STRING_dup().
17423 *Roman E. Pavlov <pre@mo.msk.ru>*
17425 * The x509 application mishandled signing requests containing DSA
17426 keys when the signing key was also DSA and the parameters didn't match.
17428 It was supposed to omit the parameters when they matched the signing key:
17429 the verifying software was then supposed to automatically use the CA's
17430 parameters if they were absent from the end user certificate.
17432 Omitting parameters is no longer recommended. The test was also
17433 the wrong way round! This was probably due to unusual behaviour in
17434 EVP_cmp_parameters() which returns 1 if the parameters match.
17435 This meant that parameters were omitted when they *didn't* match and
17436 the certificate was useless. Certificates signed with 'ca' didn't have
17439 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17441 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17442 The interface is as follows:
17443 Applications can use
17444 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17445 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17446 "off" is now the default.
17447 The library internally uses
17448 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17449 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17450 to disable memory-checking temporarily.
17452 Some inconsistent states that previously were possible (and were
17453 even the default) are now avoided.
17455 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17456 with each memory chunk allocated; this is occasionally more helpful
17457 than just having a counter.
17459 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17461 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17466 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17467 which largely parallels "options", but is for changing API behaviour,
17468 whereas "options" are about protocol behaviour.
17469 Initial "mode" flags are:
17471 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17472 a single record has been written.
17473 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17474 retries use the same buffer location.
17475 (But all of the contents must be
17480 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17483 * Fix problems with no-hmac etc.
17485 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17487 * New functions RSA_get_default_method(), RSA_set_method() and
17488 RSA_get_method(). These allows replacement of RSA_METHODs without having
17489 to mess around with the internals of an RSA structure.
17493 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17494 Also really enable memory leak checks in openssl.c and in some
17497 *Chad C. Mulligan, Bodo Moeller*
17499 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17500 up the length of negative integers. This has now been simplified to just
17501 store the length when it is first determined and use it later, rather
17502 than trying to keep track of where data is copied and updating it to
17504 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17506 * Add a new function PKCS7_signatureVerify. This allows the verification
17507 of a PKCS#7 signature but with the signing certificate passed to the
17508 function itself. This contrasts with PKCS7_dataVerify which assumes the
17509 certificate is present in the PKCS#7 structure. This isn't always the
17510 case: certificates can be omitted from a PKCS#7 structure and be
17511 distributed by "out of band" means (such as a certificate database).
17515 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17516 function prototypes in pem.h, also change util/mkdef.pl to add the
17517 necessary function names.
17521 * mk1mf.pl (used by Windows builds) did not properly read the
17522 options set by Configure in the top level Makefile, and Configure
17523 was not even able to write more than one option correctly.
17524 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17528 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17529 file to be loaded from a BIO or FILE pointer. The BIO version will
17530 for example allow memory BIOs to contain config info.
17534 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17535 Whoever hopes to achieve shared-library compatibility across versions
17536 must use this, not the compile-time macro.
17537 (Exercise 0.9.4: Which is the minimum library version required by
17539 Note: All this applies only to multi-threaded programs, others don't
17544 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17545 through a BIO pair triggered the default case, i.e.
17546 SSLerr(...,SSL_R_UNKNOWN_STATE).
17550 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17551 can use the SSL library even if none of the specific BIOs is
17556 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17557 for the encoded length.
17559 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17561 * Add initial documentation of the X509V3 functions.
17565 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17566 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17567 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17568 secure PKCS#8 private key format with a high iteration count.
17572 * Fix determination of Perl interpreter: A perl or perl5
17573 *directory* in $PATH was also accepted as the interpreter.
17575 *Ralf S. Engelschall*
17577 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17578 wrong with it but it was very old and did things like calling
17579 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17580 unusual formatting.
17584 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17585 to use the new extension code.
17589 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17590 with macros. This should make it easier to change their form, add extra
17591 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17596 * Add to configuration table a new entry that can specify an alternative
17597 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17598 according to Mark Crispin <MRC@Panda.COM>.
17602 * DES CBC did not update the IV. Weird.
17606 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17607 Changing the behaviour of the former might break existing programs --
17608 where IV updating is needed, des_ncbc_encrypt can be used.
17611 * When bntest is run from "make test" it drives bc to check its
17612 calculations, as well as internally checking them. If an internal check
17613 fails, it needs to cause bc to give a non-zero result or make test carries
17614 on without noticing the failure. Fixed.
17618 * DES library cleanups.
17622 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17623 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17624 ciphers. NOTE: although the key derivation function has been verified
17625 against some published test vectors it has not been extensively tested
17626 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17631 * Instead of "mkdir -p", which is not fully portable, use new
17632 Perl script "util/mkdir-p.pl".
17636 * Rewrite the way password based encryption (PBE) is handled. It used to
17637 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17638 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17639 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17640 the 'parameter' field of the AlgorithmIdentifier is passed to the
17641 underlying key generation function so it must do its own ASN1 parsing.
17642 This has also changed the EVP_PBE_CipherInit() function which now has a
17643 'parameter' argument instead of literal salt and iteration count values
17644 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17648 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17649 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17650 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17651 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17652 value was just used as a "magic string" and not used directly its
17653 value doesn't matter.
17657 * Introduce some semblance of const correctness to BN. Shame C doesn't
17662 * "linux-sparc64" configuration (ultrapenguin).
17664 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17665 "linux-sparc" configuration.
17667 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17669 * config now generates no-xxx options for missing ciphers.
17673 * Support the EBCDIC character set (work in progress).
17674 File ebcdic.c not yet included because it has a different license.
17676 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17678 * Support BS2000/OSD-POSIX.
17680 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17682 * Make callbacks for key generation use `void *` instead of `char *`.
17686 * Make S/MIME samples compile (not yet tested).
17690 * Additional typesafe stacks.
17694 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17698 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17700 * New configuration variant "sco5-gcc".
17702 * Updated some demos.
17704 *Sean O Riordain, Wade Scholine*
17706 * Add missing BIO_free at exit of pkcs12 application.
17710 * Fix memory leak in conf.c.
17714 * Updates for Win32 to assembler version of MD5.
17718 * Set #! path to perl in `apps/der_chop` to where we found it
17719 instead of using a fixed path.
17723 * SHA library changes for irix64-mips4-cc.
17727 * Improvements for VMS support.
17731 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17733 * Bignum library bug fix. IRIX 6 passes "make test" now!
17734 This also avoids the problems with SC4.2 and unpatched SC5.
17736 *Andy Polyakov <appro@fy.chalmers.se>*
17738 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17739 These are required because of the typesafe stack would otherwise break
17740 existing code. If old code used a structure member which used to be STACK
17741 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17742 sk_num or sk_value it would produce an error because the num, data members
17743 are not present in STACK_OF. Now it just produces a warning. sk_set
17744 replaces the old method of assigning a value to sk_value
17745 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17746 that does this will no longer work (and should use sk_set instead) but
17747 this could be regarded as a "questionable" behaviour anyway.
17751 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17752 correctly handle encrypted S/MIME data.
17756 * Change type of various DES function arguments from des_cblock
17757 (which means, in function argument declarations, pointer to char)
17758 to des_cblock * (meaning pointer to array with 8 char elements),
17759 which allows the compiler to do more typechecking; it was like
17760 that back in SSLeay, but with lots of ugly casts.
17762 Introduce new type const_des_cblock.
17766 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17767 problems: find RecipientInfo structure that matches recipient certificate
17768 and initialise the ASN1 structures properly based on passed cipher.
17772 * Belatedly make the BN tests actually check the results.
17776 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17777 to and from BNs: it was completely broken. New compilation option
17778 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17779 key elements as negative integers.
17783 * Reorganize and speed up MD5.
17785 *Andy Polyakov <appro@fy.chalmers.se>*
17789 *Richard Levitte <richard@levitte.org>*
17791 * New option -out to asn1parse to allow the parsed structure to be
17792 output to a file. This is most useful when combined with the -strparse
17793 option to examine the output of things like OCTET STRINGS.
17797 * Make SSL library a little more fool-proof by not requiring any longer
17798 that `SSL_set_{accept,connect}_state` be called before
17799 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17800 in many applications because usually everything *appeared* to work as
17801 intended anyway -- now it really works as intended).
17805 * Move openssl.cnf out of lib/.
17809 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17810 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17811 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17813 *Ralf S. Engelschall*
17815 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17816 handle PKCS#7 enveloped data properly.
17818 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17820 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17821 copying pointers. The cert_st handling is changed by this in
17822 various ways (and thus what used to be known as ctx->default_cert
17823 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17824 any longer when s->cert does not give us what we need).
17825 ssl_cert_instantiate becomes obsolete by this change.
17826 As soon as we've got the new code right (possibly it already is?),
17827 we have solved a couple of bugs of the earlier code where s->cert
17828 was used as if it could not have been shared with other SSL structures.
17830 Note that using the SSL API in certain dirty ways now will result
17831 in different behaviour than observed with earlier library versions:
17832 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17833 does not influence s as it used to.
17835 In order to clean up things more thoroughly, inside SSL_SESSION
17836 we don't use CERT any longer, but a new structure SESS_CERT
17837 that holds per-session data (if available); currently, this is
17838 the peer's certificate chain and, for clients, the server's certificate
17839 and temporary key. CERT holds only those values that can have
17840 meaningful defaults in an SSL_CTX.
17844 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17845 from the internal representation. Various PKCS#7 fixes: remove some
17846 evil casts and set the enc_dig_alg field properly based on the signing
17851 * Allow PKCS#12 password to be set from the command line or the
17852 environment. Let 'ca' get its config file name from the environment
17853 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17858 * Allow certificate policies extension to use an IA5STRING for the
17859 organization field. This is contrary to the PKIX definition but
17860 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17865 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17866 without disallowing inline assembler and the like for non-pedantic builds.
17870 * Support Borland C++ builder.
17872 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17878 * SHA-1 cleanups and performance enhancements.
17880 *Andy Polyakov <appro@fy.chalmers.se>*
17882 * Sparc v8plus assembler for the bignum library.
17884 *Andy Polyakov <appro@fy.chalmers.se>*
17886 * Accept any -xxx and +xxx compiler options in Configure.
17890 * Update HPUX configuration.
17894 * Add missing `sk_<type>_unshift()` function to safestack.h
17896 *Ralf S. Engelschall*
17898 * New function SSL_CTX_use_certificate_chain_file that sets the
17899 "extra_cert"s in addition to the certificate. (This makes sense
17900 only for "PEM" format files, as chains as a whole are not
17905 * Support verify_depth from the SSL API.
17906 x509_vfy.c had what can be considered an off-by-one-error:
17907 Its depth (which was not part of the external interface)
17908 was actually counting the number of certificates in a chain;
17909 now it really counts the depth.
17913 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17914 instead of X509err, which often resulted in confusing error
17915 messages since the error codes are not globally unique
17916 (e.g. an alleged error in ssl3_accept when a certificate
17917 didn't match the private key).
17919 * New function SSL_CTX_set_session_id_context that allows to set a default
17920 value (so that you don't need SSL_set_session_id_context for each
17921 connection using the SSL_CTX).
17925 * OAEP decoding bug fix.
17929 * Support INSTALL_PREFIX for package builders, as proposed by
17934 * New Configure options "threads" and "no-threads". For systems
17935 where the proper compiler options are known (currently Solaris
17936 and Linux), "threads" is the default.
17940 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17944 * Install various scripts to $(OPENSSLDIR)/misc, not to
17945 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17946 such as /usr/local/bin.
17950 * "make linux-shared" to build shared libraries.
17952 *Niels Poppe <niels@netbox.org>*
17954 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17958 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17959 extension adding in x509 utility.
17963 * Remove NOPROTO sections and error code comments.
17967 * Partial rewrite of the DEF file generator to now parse the ANSI
17972 * New Configure options --prefix=DIR and --openssldir=DIR.
17976 * Complete rewrite of the error code script(s). It is all now handled
17977 by one script at the top level which handles error code gathering,
17978 header rewriting and C source file generation. It should be much better
17979 than the old method: it now uses a modified version of Ulf's parser to
17980 read the ANSI prototypes in all header files (thus the old K&R definitions
17981 aren't needed for error creation any more) and do a better job of
17982 translating function codes into names. The old 'ASN1 error code embedded
17983 in a comment' is no longer necessary and it doesn't use .err files which
17984 have now been deleted. Also the error code call doesn't have to appear all
17985 on one line (which resulted in some large lines...).
17989 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17993 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17994 0 (which usually indicates a closed connection), but continue reading.
17998 * Fix some race conditions.
18002 * Add support for CRL distribution points extension. Add Certificate
18003 Policies and CRL distribution points documentation.
18007 * Move the autogenerated header file parts to crypto/opensslconf.h.
18011 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18012 8 of keying material. Merlin has also confirmed interop with this fix
18013 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18015 *Merlin Hughes <merlin@baltimore.ie>*
18017 * Fix lots of warnings.
18019 *Richard Levitte <levitte@stacken.kth.se>*
18021 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18022 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18024 *Richard Levitte <levitte@stacken.kth.se>*
18026 * Fix problems with sizeof(long) == 8.
18028 *Andy Polyakov <appro@fy.chalmers.se>*
18030 * Change functions to ANSI C.
18034 * Fix typos in error codes.
18036 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18038 * Remove defunct assembler files from Configure.
18042 * SPARC v8 assembler BIGNUM implementation.
18044 *Andy Polyakov <appro@fy.chalmers.se>*
18046 * Support for Certificate Policies extension: both print and set.
18047 Various additions to support the r2i method this uses.
18051 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18052 return a const string when you are expecting an allocated buffer.
18056 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18057 types DirectoryString and DisplayText.
18061 * Add code to allow r2i extensions to access the configuration database,
18062 add an LHASH database driver and add several ctx helper functions.
18066 * Fix an evil bug in bn_expand2() which caused various BN functions to
18067 fail when they extended the size of a BIGNUM.
18071 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18072 support typesafe stack.
18076 * Fix typo in SSL_[gs]et_options().
18078 *Nils Frostberg <nils@medcom.se>*
18080 * Delete various functions and files that belonged to the (now obsolete)
18081 old X509V3 handling code.
18085 * New Configure option "rsaref".
18089 * Don't auto-generate pem.h.
18093 * Introduce type-safe ASN.1 SETs.
18097 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18099 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18101 * Introduce type-safe STACKs. This will almost certainly break lots of code
18102 that links with OpenSSL (well at least cause lots of warnings), but fear
18103 not: the conversion is trivial, and it eliminates loads of evil casts. A
18104 few STACKed things have been converted already. Feel free to convert more.
18105 In the fullness of time, I'll do away with the STACK type altogether.
18109 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18110 specified in `<certfile>` by updating the entry in the index.txt file.
18111 This way one no longer has to edit the index.txt file manually for
18112 revoking a certificate. The -revoke option does the gory details now.
18114 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18116 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18117 `-text` option at all and this way the `-noout -text` combination was
18118 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18120 *Ralf S. Engelschall*
18122 * Make sure a corresponding plain text error message exists for the
18123 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18124 verify callback function determined that a certificate was revoked.
18126 *Ralf S. Engelschall*
18128 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18129 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18130 all available ciphers including rc5, which was forgotten until now.
18131 In order to let the testing shell script know which algorithms
18132 are available, a new (up to now undocumented) command
18133 `openssl list-cipher-commands` is used.
18137 * Bugfix: s_client occasionally would sleep in select() when
18138 it should have checked SSL_pending() first.
18142 * New functions DSA_do_sign and DSA_do_verify to provide access to
18143 the raw DSA values prior to ASN.1 encoding.
18147 * Tweaks to Configure
18149 *Niels Poppe <niels@netbox.org>*
18151 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18156 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18160 * New config option to avoid instructions that are illegal on the 80386.
18161 The default code is faster, but requires at least a 486.
18165 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18166 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18167 same as SSL2_VERSION anyway.
18171 * New "-showcerts" option for s_client.
18175 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18176 application. Various cleanups and fixes.
18180 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18181 modify error routines to work internally. Add error codes and PBE init
18182 to library startup routines.
18186 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18187 packing functions to asn1 and evp. Changed function names and error
18188 codes along the way.
18192 * PKCS12 integration: and so it begins... First of several patches to
18193 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18194 objects to objects.h
18198 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18199 and display support for Thawte strong extranet extension.
18203 * Add LinuxPPC support.
18205 *Jeff Dubrule <igor@pobox.org>*
18207 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18208 bn_div_words in alpha.s.
18210 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18212 * Make sure the RSA OAEP test is skipped under -DRSAref because
18213 OAEP isn't supported when OpenSSL is built with RSAref.
18215 *Ulf Moeller <ulf@fitug.de>*
18217 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18218 so they no longer are missing under -DNOPROTO.
18220 *Soren S. Jorvang <soren@t.dk>*
18222 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18224 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18225 doesn't work when the session is reused. Coming soon!
18229 * Fix a security hole, that allows sessions to be reused in the wrong
18230 context thus bypassing client cert protection! All software that uses
18231 client certs and session caches in multiple contexts NEEDS PATCHING to
18232 allow session reuse! A fuller solution is in the works.
18234 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18236 * Some more source tree cleanups (removed obsolete files
18237 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18238 permission on "config" script to be executable) and a fix for the INSTALL
18241 *Ulf Moeller <ulf@fitug.de>*
18243 * Remove some legacy and erroneous uses of malloc, free instead of
18246 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18248 * Make rsa_oaep_test return non-zero on error.
18250 *Ulf Moeller <ulf@fitug.de>*
18252 * Add support for native Solaris shared libraries. Configure
18253 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18254 if someone would make that last step automatic.
18256 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18258 * ctx_size was not built with the right compiler during "make links". Fixed.
18262 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18263 except NULL ciphers". This means the default cipher list will no longer
18264 enable NULL ciphers. They need to be specifically enabled e.g. with
18265 the string "DEFAULT:eNULL".
18269 * Fix to RSA private encryption routines: if p < q then it would
18270 occasionally produce an invalid result. This will only happen with
18271 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18275 * Be less restrictive and allow also `perl util/perlpath.pl
18276 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18277 because this way one can also use an interpreter named `perl5` (which is
18278 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18279 installed as `perl`).
18281 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18283 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18285 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18287 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18288 advapi32.lib to Win32 build and change the pem test comparison
18289 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18290 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18291 and crypto/des/ede_cbcm_enc.c.
18295 * DES quad checksum was broken on big-endian architectures. Fixed.
18299 * Comment out two functions in bio.h that aren't implemented. Fix up the
18300 Win32 test batch file so it (might) work again. The Win32 test batch file
18301 is horrible: I feel ill....
18305 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18306 in e_os.h. Audit of header files to check ANSI and non ANSI
18307 sections: 10 functions were absent from non ANSI section and not exported
18308 from Windows DLLs. Fixed up libeay.num for new functions.
18312 * Make `openssl version` output lines consistent.
18314 *Ralf S. Engelschall*
18316 * Fix Win32 symbol export lists for BIO functions: Added
18317 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18318 to ms/libeay{16,32}.def.
18320 *Ralf S. Engelschall*
18322 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18323 fine under Unix and passes some trivial tests I've now added. But the
18324 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18325 added to make sure no one expects that this stuff really works in the
18326 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18327 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18330 *Ralf S. Engelschall*
18332 * Fix the generation of two part addresses in perl.
18334 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18336 * Add config entry for Linux on MIPS.
18338 *John Tobey <jtobey@channel1.com>*
18340 * Make links whenever Configure is run, unless we are on Windoze.
18344 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18345 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18350 * Add a useful kludge to allow package maintainers to specify compiler and
18351 other platforms details on the command line without having to patch the
18352 Configure script every time: One now can use
18353 `perl Configure <id>:<details>`,
18354 i.e. platform ids are allowed to have details appended
18355 to them (separated by colons). This is treated as there would be a static
18356 pre-configured entry in Configure's %table under key `<id>` with value
18357 `<details>` and `perl Configure <id>` is called. So, when you want to
18358 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18359 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18360 now, which overrides the FreeBSD-elf entry on-the-fly.
18362 *Ralf S. Engelschall*
18364 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18368 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18369 on the `perl Configure ...` command line. This way one can compile
18370 OpenSSL libraries with Position Independent Code (PIC) which is needed
18371 for linking it into DSOs.
18373 *Ralf S. Engelschall*
18375 * Remarkably, export ciphers were totally broken and no-one had noticed!
18380 * Cleaned up the LICENSE document: The official contact for any license
18381 questions now is the OpenSSL core team under openssl-core@openssl.org.
18382 And add a paragraph about the dual-license situation to make sure people
18383 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18384 to the OpenSSL toolkit.
18386 *Ralf S. Engelschall*
18388 * General source tree makefile cleanups: Made `making xxx in yyy...`
18389 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18390 Additionally cleaned up the `make links` target: Remove unnecessary
18391 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18392 to speed processing and no longer clutter the display with confusing
18393 stuff. Instead only the actually done links are displayed.
18395 *Ralf S. Engelschall*
18397 * Permit null encryption ciphersuites, used for authentication only. It used
18398 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18399 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18404 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18405 signed attributes when verifying signatures (this would break them),
18406 the detached data encoding was wrong and public keys obtained using
18407 X509_get_pubkey() weren't freed.
18411 * Add text documentation for the BUFFER functions. Also added a work around
18412 to a Win95 console bug. This was triggered by the password read stuff: the
18413 last character typed gets carried over to the next fread(). If you were
18414 generating a new cert request using 'req' for example then the last
18415 character of the passphrase would be CR which would then enter the first
18420 * Added the new 'Includes OpenSSL Cryptography Software' button as
18421 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18422 button and can be used by applications based on OpenSSL to show the
18423 relationship to the OpenSSL project.
18425 *Ralf S. Engelschall*
18427 * Remove confusing variables in function signatures in files
18428 ssl/ssl_lib.c and ssl/ssl.h.
18430 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18432 * Don't install bss_file.c under PREFIX/include/
18434 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18436 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18437 functions that return function pointers and has support for NT specific
18438 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18439 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18440 unsigned to signed types: this was killing the Win32 compile.
18444 * Add new certificate file to stack functions,
18445 SSL_add_dir_cert_subjects_to_stack() and
18446 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18447 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18448 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18449 This means that Apache-SSL and similar packages don't have to mess around
18450 to add as many CAs as they want to the preferred list.
18454 * Experiment with doxygen documentation. Currently only partially applied to
18456 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18457 openssl.doxy as the configuration file.
18461 * Get rid of remaining C++-style comments which strict C compilers hate.
18463 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18465 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18466 compiled in by default: it has problems with large keys.
18470 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18471 DH private keys and/or callback functions which directly correspond to
18472 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18473 is needed for applications which have to configure certificates on a
18474 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18476 For the RSA certificate situation is makes no difference, but
18477 for the DSA certificate situation this fixes the "no shared cipher"
18478 problem where the OpenSSL cipher selection procedure failed because the
18479 temporary keys were not overtaken from the context and the API provided
18480 no way to reconfigure them.
18481 The new functions now let applications reconfigure the stuff and they
18482 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18483 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18484 non-public-API function ssl_cert_instantiate() is used as a helper
18485 function and also to reduce code redundancy inside ssl_rsa.c.
18487 *Ralf S. Engelschall*
18489 * Move s_server -dcert and -dkey options out of the undocumented feature
18490 area because they are useful for the DSA situation and should be
18491 recognized by the users.
18493 *Ralf S. Engelschall*
18495 * Fix the cipher decision scheme for export ciphers: the export bits are
18496 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18497 SSL_EXP_MASK. So, the original variable has to be used instead of the
18498 already masked variable.
18500 *Richard Levitte <levitte@stacken.kth.se>*
18502 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18504 *Richard Levitte <levitte@stacken.kth.se>*
18506 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18507 from `int` to `unsigned int` because it is a length and initialized by
18508 EVP_DigestFinal() which expects an `unsigned int *`.
18510 *Richard Levitte <levitte@stacken.kth.se>*
18512 * Don't hard-code path to Perl interpreter on shebang line of Configure
18513 script. Instead use the usual Shell->Perl transition trick.
18515 *Ralf S. Engelschall*
18517 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18518 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18519 -noout -modulus` as it's already the case for `openssl rsa -noout
18520 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18521 currently the public key is printed (a decision which was already done by
18522 `openssl dsa -modulus` in the past) which serves a similar purpose.
18523 Additionally the NO_RSA no longer completely removes the whole -modulus
18524 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18527 *Ralf S. Engelschall*
18529 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18530 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18532 *Arne Ansper <arne@ats.cyber.ee>*
18534 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18535 to be added. Now both 'req' and 'ca' can use new objects defined in the
18540 * Add cool BIO that does syslog (or event log on NT).
18542 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18544 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18545 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18546 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18547 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18551 * Add preliminary config info for new extension code.
18555 * Make RSA_NO_PADDING really use no padding.
18557 *Ulf Moeller <ulf@fitug.de>*
18559 * Generate errors when private/public key check is done.
18563 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18564 for some CRL extensions and new objects added.
18568 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18569 key usage extension and fuller support for authority key id.
18573 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18574 padding method for RSA, which is recommended for new applications in PKCS
18575 #1 v2.0 (RFC 2437, October 1998).
18576 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18577 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18578 against Bleichbacher's attack on RSA.
18579 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18582 * Updates to the new SSL compression code
18584 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18586 * Fix so that the version number in the master secret, when passed
18587 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18588 (because the server will not accept higher), that the version number
18589 is 0x03,0x01, not 0x03,0x00
18591 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18593 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18594 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18595 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18599 * Support for RAW extensions where an arbitrary extension can be
18600 created by including its DER encoding. See `apps/openssl.cnf` for
18605 * Make sure latest Perl versions don't interpret some generated C array
18606 code as Perl array code in the crypto/err/err_genc.pl script.
18608 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18610 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18611 not many people have the assembler. Various Win32 compilation fixes and
18612 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18613 build instructions.
18617 * Modify configure script 'Configure' to automatically create crypto/date.h
18618 file under Win32 and also build pem.h from pem.org. New script
18619 util/mkfiles.pl to create the MINFO file on environments that can't do a
18620 'make files': perl util/mkfiles.pl >MINFO should work.
18624 * Major rework of DES function declarations, in the pursuit of correctness
18625 and purity. As a result, many evil casts evaporated, and some weirdness,
18626 too. You may find this causes warnings in your code. Zapping your evil
18627 casts will probably fix them. Mostly.
18631 * Fix for a typo in asn1.h. Bug fix to object creation script
18632 obj_dat.pl. It considered a zero in an object definition to mean
18633 "end of object": none of the objects in objects.h have any zeros
18634 so it wasn't spotted.
18636 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18638 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18639 Masking (CBCM). In the absence of test vectors, the best I have been able
18640 to do is check that the decrypt undoes the encrypt, so far. Send me test
18641 vectors if you have them.
18645 * Correct calculation of key length for export ciphers (too much space was
18646 allocated for null ciphers). This has not been tested!
18650 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18651 message is now correct (it understands "crypto" and "ssl" on its
18652 command line). There is also now an "update" option. This will update
18653 the util/ssleay.num and util/libeay.num files with any new functions.
18655 perl util/mkdef.pl crypto ssl update
18656 it will update them.
18660 * Overhauled the Perl interface:
18661 - ported BN stuff to OpenSSL's different BN library
18662 - made the perl/ source tree CVS-aware
18663 - renamed the package from SSLeay to OpenSSL (the files still contain
18664 their history because I've copied them in the repository)
18665 - removed obsolete files (the test scripts will be replaced
18666 by better Test::Harness variants in the future)
18668 *Ralf S. Engelschall*
18670 * First cut for a very conservative source tree cleanup:
18671 1. merge various obsolete readme texts into doc/ssleay.txt
18672 where we collect the old documents and readme texts.
18673 2. remove the first part of files where I'm already sure that we no
18674 longer need them because of three reasons: either they are just temporary
18675 files which were left by Eric or they are preserved original files where
18676 I've verified that the diff is also available in the CVS via "cvs diff
18677 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18678 the crypto/md/ stuff).
18680 *Ralf S. Engelschall*
18682 * More extension code. Incomplete support for subject and issuer alt
18683 name, issuer and authority key id. Change the i2v function parameters
18684 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18685 what that's for :-) Fix to ASN1 macro which messed up
18686 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18690 * Preliminary support for ENUMERATED type. This is largely copied from the
18695 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18697 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18699 * Make sure `make rehash` target really finds the `openssl` program.
18701 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18703 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18704 like to hear about it if this slows down other processors.
18708 * Add CygWin32 platform information to Configure script.
18710 *Alan Batie <batie@aahz.jf.intel.com>*
18712 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18714 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18716 * New program nseq to manipulate netscape certificate sequences
18720 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18725 * Fixes to BN code. Previously the default was to define BN_RECURSION
18726 but the BN code had some problems that would cause failures when
18727 doing certificate verification and some other functions.
18729 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18731 * Add ASN1 and PEM code to support netscape certificate sequences.
18735 * Add ASN1 and PEM code to support netscape certificate sequences.
18739 * Add several PKIX and private extended key usage OIDs.
18743 * Modify the 'ca' program to handle the new extension code. Modify
18744 openssl.cnf for new extension format, add comments.
18748 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18749 and add a sample to openssl.cnf so req -x509 now adds appropriate
18754 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18755 error code, add initial support to X509_print() and x509 application.
18759 * Takes a deep breath and start adding X509 V3 extension support code. Add
18760 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18761 stuff is currently isolated and isn't even compiled yet.
18765 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18766 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18767 Removed the versions check from X509 routines when loading extensions:
18768 this allows certain broken certificates that don't set the version
18769 properly to be processed.
18773 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18774 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18775 can still be regenerated with "make depend".
18779 * Spelling mistake in C version of CAST-128.
18781 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18783 * Changes to the error generation code. The perl script err-code.pl
18784 now reads in the old error codes and retains the old numbers, only
18785 adding new ones if necessary. It also only changes the .err files if new
18786 codes are added. The makefiles have been modified to only insert errors
18787 when needed (to avoid needlessly modifying header files). This is done
18788 by only inserting errors if the .err file is newer than the auto generated
18789 C file. To rebuild all the error codes from scratch (the old behaviour)
18790 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18791 or delete all the .err files.
18795 * CAST-128 was incorrectly implemented for short keys. The C version has
18796 been fixed, but is untested. The assembler versions are also fixed, but
18797 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18798 to regenerate it if needed.
18799 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18800 Hagino <itojun@kame.net>*
18802 * File was opened incorrectly in randfile.c.
18804 *Ulf Möller <ulf@fitug.de>*
18806 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18807 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18808 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18809 al: it's just almost always a UTCTime. Note this patch adds new error
18810 codes so do a "make errors" if there are problems.
18814 * Correct Linux 1 recognition in config.
18816 *Ulf Möller <ulf@fitug.de>*
18818 * Remove pointless MD5 hash when using DSA keys in ca.
18820 *Anonymous <nobody@replay.com>*
18822 * Generate an error if given an empty string as a cert directory. Also
18823 generate an error if handed NULL (previously returned 0 to indicate an
18824 error, but didn't set one).
18826 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18828 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18832 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18833 parameters. This was causing a warning which killed off the Win32 compile.
18837 * Remove C++ style comments from crypto/bn/bn_local.h.
18839 *Neil Costigan <neil.costigan@celocom.com>*
18841 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18842 based on a text string, looking up short and long names and finally
18843 "dot" format. The "dot" format stuff didn't work. Added new function
18844 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18845 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18846 OID is not part of the table.
18850 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18851 X509_LOOKUP_by_alias().
18855 * Sort openssl functions by name.
18859 * Get the `gendsa` command working and add it to the `list` command. Remove
18860 encryption from sample DSA keys (in case anyone is interested the password
18865 * Make *all* `*_free` functions accept a NULL pointer.
18867 *Frans Heymans <fheymans@isaserver.be>*
18869 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18872 *Anonymous <nobody@replay.com>*
18874 * s_server should send the CAfile as acceptable CAs, not its own cert.
18876 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18878 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18880 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18882 * Temp key "for export" tests were wrong in s3_srvr.c.
18884 *Anonymous <nobody@replay.com>*
18886 * Add prototype for temp key callback functions
18887 SSL_CTX_set_tmp_{rsa,dh}_callback().
18891 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18892 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18896 * X509_name_add_entry() freed the wrong thing after an error.
18898 *Arne Ansper <arne@ats.cyber.ee>*
18900 * rsa_eay.c would attempt to free a NULL context.
18902 *Arne Ansper <arne@ats.cyber.ee>*
18904 * BIO_s_socket() had a broken should_retry() on Windoze.
18906 *Arne Ansper <arne@ats.cyber.ee>*
18908 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18910 *Arne Ansper <arne@ats.cyber.ee>*
18912 * Make sure the already existing X509_STORE->depth variable is initialized
18913 in X509_STORE_new(), but document the fact that this variable is still
18914 unused in the certificate verification process.
18916 *Ralf S. Engelschall*
18918 * Fix the various library and `apps/` files to free up pkeys obtained from
18919 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18923 * Fix reference counting in X509_PUBKEY_get(). This makes
18924 demos/maurice/example2.c work, amongst others, probably.
18926 *Steve Henson and Ben Laurie*
18928 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18929 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18930 are no longer created. This way we have a single and consistent command
18931 line interface `openssl <command>`, similar to `cvs <command>`.
18933 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18935 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18936 BIT STRING wrapper always have zero unused bits.
18940 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18944 * Make the top-level INSTALL documentation easier to understand.
18948 * Makefiles updated to exit if an error occurs in a sub-directory
18949 make (including if user presses ^C) [Paul Sutton]
18951 * Make Montgomery context stuff explicit in RSA data structure.
18955 * Fix build order of pem and err to allow for generated pem.h.
18959 * Fix renumbering bug in X509_NAME_delete_entry().
18963 * Enhanced the err-ins.pl script so it makes the error library number
18964 global and can add a library name. This is needed for external ASN1 and
18965 other error libraries.
18969 * Fixed sk_insert which never worked properly.
18973 * Fix ASN1 macros so they can handle indefinite length constructed
18974 EXPLICIT tags. Some non standard certificates use these: they can now
18979 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18980 into a single doc/ssleay.txt bundle. This way the information is still
18981 preserved but no longer messes up this directory. Now it's new room for
18982 the new set of documentation files.
18984 *Ralf S. Engelschall*
18986 * SETs were incorrectly DER encoded. This was a major pain, because they
18987 shared code with SEQUENCEs, which aren't coded the same. This means that
18988 almost everything to do with SETs or SEQUENCEs has either changed name or
18989 number of arguments.
18991 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18993 * Fix test data to work with the above.
18997 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18998 was already fixed by Eric for 0.9.1 it seems.
19000 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19002 * Autodetect FreeBSD3.
19006 * Fix various bugs in Configure. This affects the following platforms:
19010 unixware-2.0-pentium
19015 * Eliminate generated files from CVS. Reorder tests to regenerate files
19016 before they are needed.
19020 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19024 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19026 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19027 changed SSLeay to OpenSSL in version strings.
19029 *Ralf S. Engelschall*
19031 * Some fixups to the top-level documents.
19035 * Fixed the nasty bug where rsaref.h was not found under compile-time
19036 because the symlink to include/ was missing.
19038 *Ralf S. Engelschall*
19040 * Incorporated the popular no-RSA/DSA-only patches
19041 which allow to compile a RSA-free SSLeay.
19043 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19045 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19046 when "ssleay" is still not found.
19048 *Ralf S. Engelschall*
19050 * Added more platforms to Configure: Cray T3E, HPUX 11,
19052 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19054 * Updated the README file.
19056 *Ralf S. Engelschall*
19058 * Added various .cvsignore files in the CVS repository subdirs
19059 to make a "cvs update" really silent.
19061 *Ralf S. Engelschall*
19063 * Recompiled the error-definition header files and added
19064 missing symbols to the Win32 linker tables.
19066 *Ralf S. Engelschall*
19068 * Cleaned up the top-level documents;
19069 o new files: CHANGES and LICENSE
19070 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19071 o merged COPYRIGHT into LICENSE
19072 o removed obsolete TODO file
19073 o renamed MICROSOFT to INSTALL.W32
19075 *Ralf S. Engelschall*
19077 * Removed dummy files from the 0.9.1b source tree:
19079 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19080 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19081 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19082 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19083 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19086 *Ralf S. Engelschall*
19088 * Added various platform portability fixes.
19092 * The Genesis of the OpenSSL rpject:
19093 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19094 Young and Tim J. Hudson created while they were working for C2Net until
19097 *The OpenSSL Project*
19099 ### Changes between 0.9.0b and 0.9.1b [not released]
19101 * Updated a few CA certificates under certs/
19105 * Changed some BIGNUM api stuff.
19109 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19110 DGUX x86, Linux Alpha, etc.
19114 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19115 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19120 * Add -strparse option to asn1pars program which parses nested
19123 *Dr Stephen Henson <shenson@bigfoot.com>*
19125 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19129 * DSA fix for "ca" program.
19133 * Added "-genkey" option to "dsaparam" program.
19137 * Added RIPE MD160 (rmd160) message digest.
19141 * Added -a (all) option to "ssleay version" command.
19145 * Added PLATFORM define which is the id given to Configure.
19149 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19153 * Extended the ASN.1 parser routines.
19157 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19161 * Added a BN_CTX to the BN library.
19165 * Fixed the weak key values in DES library
19169 * Changed API in EVP library for cipher aliases.
19173 * Added support for RC2/64bit cipher.
19177 * Converted the lhash library to the crypto/mem.c functions.
19181 * Added more recognized ASN.1 object ids.
19185 * Added more RSA padding checks for SSL/TLS.
19189 * Added BIO proxy/filter functionality.
19193 * Added extra_certs to SSL_CTX which can be used
19194 send extra CA certificates to the client in the CA cert chain sending
19195 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19199 * Now Fortezza is denied in the authentication phase because
19200 this is key exchange mechanism is not supported by SSLeay at all.
19204 * Additional PKCS1 checks.
19208 * Support the string "TLSv1" for all TLS v1 ciphers.
19212 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19213 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19217 * Fixed a few memory leaks.
19221 * Fixed various code and comment typos.
19225 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19226 bytes sent in the client random.
19228 *Edward Bishop <ebishop@spyglass.com>*
19232 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19233 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19234 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19235 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19236 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19237 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19238 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19239 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19240 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19241 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19242 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19243 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19244 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19245 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19246 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19247 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19248 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19249 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19250 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19251 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19252 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19253 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19254 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19255 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19256 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19257 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19258 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19259 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19260 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19261 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19262 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19263 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19264 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19265 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19266 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19267 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19268 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19269 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19270 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19271 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19272 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19273 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19274 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19275 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19276 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19277 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19278 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19279 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19280 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19281 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19282 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19283 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19284 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19285 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19286 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19287 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19288 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19289 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19290 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19291 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19292 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19293 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19294 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19295 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19296 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19297 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19298 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19299 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19300 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19301 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19302 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19303 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19304 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19305 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19306 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19307 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19308 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19309 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19310 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19311 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19312 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19313 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19314 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19315 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19316 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19317 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19318 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19319 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19320 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19321 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19322 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19323 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19324 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19325 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19326 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19327 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19328 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19329 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19330 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19331 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19332 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19333 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19334 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19335 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19336 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19337 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19338 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19339 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19340 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19341 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19342 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19343 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19344 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19345 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19346 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19347 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19348 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19349 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19350 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19351 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19352 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19353 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19354 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19355 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19356 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19357 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19358 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19359 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19360 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19361 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19362 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19363 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19364 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19365 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19366 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19367 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19368 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19369 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19370 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19371 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19372 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19373 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19374 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19375 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19376 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19377 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19378 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19379 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19380 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19381 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19382 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19383 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19384 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19385 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19386 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19387 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19388 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19389 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19390 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19391 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19392 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19393 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655