4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * The signatures of the functions to get and set options on SSL and
27 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
28 Some source code changes may be required.
32 * Add "abspath" and "includedir" pragma's to config files, to prevent,
33 or modify relative pathname inclusion.
37 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
38 validated. The module is implemented as an OpenSSL provider, the so-called
39 FIPS provider. A list of all changes related to the FIPS provider would go
40 beyond the scope of this CHANGES file, please consult the README-FIPS and
41 README-PROVIDERS files, as well as the migration guide.
43 The FIPS provider is disabled by default and needs to be enabled explicitly
44 at configuration time using the `enable-fips` option. If it is enabled,
45 the FIPS provider gets built and installed in addition to the default and
46 the legacy provider. No separate installation procedure is necessary.
47 There is however a dedicated `install_fips` make target, which serves the
48 special purpose of installing only the FIPS provider into an existing
51 *OpenSSL team members and many third party contributors*
53 * For the key types DH and DHX the allowed settable parameters are now different.
54 Previously (in 1.1.1) these conflicting parameters were allowed, but will now
55 result in errors. See EVP_PKEY-DH(7) for further details. This affects the
56 behaviour of openssl-genpkey(1) for DH parameter generation.
60 * The openssl commands that read keys, certificates, and CRLs now
61 automatically detect the PEM or DER format of the input files so it is not
62 necessary to explicitly specify the input format anymore. However if the
63 input format option is used the specified format will be required.
65 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
67 * Added enhanced PKCS#12 APIs which accept a library context `OSSL_LIB_CTX`
68 and (where relevant) a property query. Other APIs which handle PKCS#7 and
69 PKCS#8 objects have also been enhanced where required. This includes:
71 PKCS12_add_key_ex(), PKCS12_add_safe_ex(), PKCS12_add_safes_ex(),
72 PKCS12_create_ex(), PKCS12_decrypt_skey_ex(), PKCS12_init_ex(),
73 PKCS12_item_decrypt_d2i_ex(), PKCS12_item_i2d_encrypt_ex(),
74 PKCS12_key_gen_asc_ex(), PKCS12_key_gen_uni_ex(), PKCS12_key_gen_utf8_ex(),
75 PKCS12_pack_p7encdata_ex(), PKCS12_pbe_crypt_ex(), PKCS12_PBE_keyivgen_ex(),
76 PKCS12_SAFEBAG_create_pkcs8_encrypt_ex(), PKCS5_pbe2_set_iv_ex(),
77 PKCS5_pbe_set0_algor_ex(), PKCS5_pbe_set_ex(), PKCS5_pbkdf2_set_ex(),
78 PKCS5_v2_PBE_keyivgen_ex(), PKCS5_v2_scrypt_keyivgen_ex(),
79 PKCS8_decrypt_ex(), PKCS8_encrypt_ex(), PKCS8_set0_pbe_ex().
81 As part of this change the EVP_PBE_xxx APIs can also accept a library
82 context and property query and will call an extended version of the key/IV
83 derivation function which supports these parameters. This includes
84 EVP_PBE_CipherInit_ex(), EVP_PBE_find_ex() and EVP_PBE_scrypt_ex().
88 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
92 * Added support for Kernel TLS (KTLS). In order to use KTLS, support for it
93 must be compiled in using the "enable-ktls" compile time option. It must
94 also be enabled at run time using the SSL_OP_ENABLE_KTLS option.
96 *Boris Pismenny, John Baldwin and Andrew Gallatin*
98 * Support for RFC 5746 secure renegotiation is now required by default for
99 SSL or TLS connections to succeed. Applications that require the ability
100 to connect to legacy peers will need to explicitly set
101 SSL_OP_LEGACY_SERVER_CONNECT. Accordingly, SSL_OP_LEGACY_SERVER_CONNECT
102 is no longer set as part of SSL_OP_ALL.
106 * The signature of the `copy` functional parameter of the
107 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
108 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
109 the signature of the `pub_decode` functional parameter of the
110 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
111 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
115 * The error return values from some control calls (ctrl) have changed.
116 One significant change is that controls which used to return -2 for
117 invalid inputs, now return -1 indicating a generic error condition instead.
121 * A public key check is now performed during EVP_PKEY_derive_set_peer().
122 Previously DH was internally doing this during EVP_PKEY_derive().
123 To disable this check use EVP_PKEY_derive_set_peer_ex(dh, peer, 0). This
124 may mean that an error can occur in EVP_PKEY_derive_set_peer() rather than
125 during EVP_PKEY_derive().
129 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
130 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
131 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
132 are deprecated. They are not invoked by the OpenSSL library anymore and
133 are replaced by direct checks of the key operation against the key type
134 when the operation is initialized.
138 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
139 more key types including RSA, DSA, ED25519, X25519, ED448 and X448.
140 Previously (in 1.1.1) they would return -2. For key types that do not have
141 parameters then EVP_PKEY_param_check() will always return 1.
143 * The output from the command line applications may have minor
144 changes. These are primarily changes in capitalisation and white
145 space. However, in some cases, there are additional differences.
146 For example, the DH parameters output from `dhparam` now lists 'P',
147 'Q', 'G' and 'pcounter' instead of 'prime', 'generator', 'subgroup
148 order' and 'counter' respectively.
152 * The output from numerous "printing" functions such as X509_signature_print(),
153 X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been
154 amended such that there may be cosmetic differences between the output
155 observed in 1.1.1 and 3.0. This also applies to the "-text" output from the
156 x509 and crl applications.
160 * Windows thread synchronization uses read/write primitives (SRWLock) when
161 supported by the OS, otherwise CriticalSection continues to be used.
165 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
166 work on read only BIO source/sinks that do not support these functions.
167 This allows piping or redirection of a file BIO using stdin to be buffered
168 into memory. This is used internally in OSSL_DECODER_from_bio().
172 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
173 this function would return one of the values OSSL_STORE_INFO_NAME,
174 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
175 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
176 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
177 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
178 using this function should be amended to handle the changed return value.
182 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
183 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
184 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
185 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
186 contains more than one certificate identifier: This means that all
187 certificates referenced there MUST be part of the validation chain.
191 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
192 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
193 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
194 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
195 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
196 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
197 using the EVP APIs to access these ciphers should instead use more modern
198 ciphers. If that is not possible then these applications should ensure that
199 the legacy provider has been loaded. This can be achieved either
200 programmatically or via configuration. See the provider(7) man page for
205 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
206 RIPEMD-160 have been moved to the legacy provider. Applications using the
207 EVP APIs to access these digests should instead use more modern digests. If
208 that is not possible then these applications should ensure that the legacy
209 provider has been loaded. This can be achieved either programmatically or via
210 configuration. See the provider(7) man page for further details.
214 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
219 * The deprecated functions EVP_PKEY_get0_RSA(),
220 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
221 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
222 well as the similarly named "get1" functions behave slightly differently in
223 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
224 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
225 provider. Calling these functions will only return a handle on the internal
226 key where the EVP_PKEY was constructed using this key in the first place, for
227 example using a function or macro such as EVP_PKEY_assign_RSA(),
228 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
229 then these functions now return a cached copy of the key. Changes to
230 the internal provider key that take place after the first time the cached key
231 is accessed will not be reflected back in the cached copy. Similarly any
232 changes made to the cached copy by application code will not be reflected
233 back in the internal provider key.
235 For the above reasons the keys returned from these functions should typically
236 be treated as read-only. To emphasise this the value returned from
237 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
238 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
239 break some existing code. Applications broken by this change should be
240 modified. The preferred solution is to refactor the code to avoid the use of
241 these deprecated functions. Failing this the code should be modified to use a
242 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
243 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
244 non-const pointer to enable them to be "freed". However they should also be
245 treated as read-only.
249 * A number of functions handling low-level keys or engines were deprecated
250 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
251 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
252 EVP_PKEY_get0_siphash(). Applications using engines should instead use
253 providers. Applications getting or setting low-level keys in an EVP_PKEY
254 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
255 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
259 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
260 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
261 and require returning octet ptr parameters from providers that
262 would like to support them which complicates provider implementations.
266 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
267 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
268 RAND_set_rand_method(). Provider based random number generators should
269 be used instead via EVP_RAND(3).
273 * The SRP APIs have been deprecated. The old APIs do not work via providers,
274 and there is no EVP interface to them. Unfortunately there is no replacement
275 for these APIs at this time.
279 * Add a compile time option to prevent the caching of provider fetched
280 algorithms. This is enabled by including the no-cached-fetch option
281 at configuration time.
285 * The default algorithms for pkcs12 creation with the PKCS12_create() function
286 were changed to more modern PBKDF2 and AES based algorithms. The default
287 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
288 with the password-based encryption iteration count. The default digest
289 algorithm for the MAC computation was changed to SHA-256. The pkcs12
290 application now supports -legacy option that restores the previous
291 default algorithms to support interoperability with legacy systems.
293 *Tomáš Mráz and Sahana Prasad*
295 * The openssl speed command does not use low-level API calls anymore. This
296 implies some of the performance numbers might not be fully comparable
297 with the previous releases due to higher overhead. This applies
298 particularly to measuring performance on smaller data chunks.
302 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
305 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
307 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
308 Typically if OpenSSL has no EC or DH algorithms then it cannot support
309 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
310 through providers. Therefore third party providers may supply group
311 implementations even where there are no built-in ones. Attempting to create
312 TLS connections in such a build without also disabling TLSv1.3 at run time or
313 using third party provider groups may result in handshake failures. TLSv1.3
314 can be disabled at compile time using the "no-tls1_3" Configure option.
318 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
319 providers may supply their own group implementations (using either the "key
320 exchange" or the "key encapsulation" methods) which will automatically be
321 detected and used by libssl.
323 *Matt Caswell, Nicola Tuveri*
325 * The undocumented function X509_certificate_type() has been deprecated;
326 applications can use X509_get0_pubkey() and X509_get0_signature() to
327 get the same information.
331 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
332 functions. They are identical to BN_rand() and BN_rand_range()
337 * Removed RSA padding mode for SSLv23 (which was only used for
338 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
339 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
344 * Deprecated the obsolete X9.31 RSA key generation related functions
345 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
346 BN_X931_generate_prime_ex().
350 * The default key generation method for the regular 2-prime RSA keys was
351 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
352 Conditions Based on Auxiliary Probable Primes). This method is slower
353 than the original method.
357 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
358 They are replaced with the BN_check_prime() function that avoids possible
359 misuse and always uses at least 64 rounds of the Miller-Rabin
360 primality test. At least 64 rounds of the Miller-Rabin test are now also
361 used for all prime generation, including RSA key generation.
362 This increases key generation time, especially for larger keys.
366 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
367 as they are not useful with non-deprecated functions.
371 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
372 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
373 OCSP_REQ_CTX_i2d() and its special form OCSP_REQ_CTX_set1_req(),
375 OCSP_REQ_CTX_nbio_d2i() and its special form OCSP_sendreq_nbio(),
376 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
377 were used to collect all necessary data to form a HTTP request, and to
378 perform the HTTP transfer with that request. With OpenSSL 3.0, the
379 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
380 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
381 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
382 OSSL_HTTP_REQ_CTX_set1_req(),
383 OSSL_HTTP_REQ_CTX_nbio(), OSSL_HTTP_REQ_CTX_nbio_d2i(),
384 OSSL_HTTP_REQ_CTX_get0_mem_bio(), and
385 OSSL_HTTP_REQ_CTX_set_max_response_length().
387 *Rich Salz, Richard Levitte, and David von Oheimb*
389 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`,
390 which are superseded by `X509_load_http()` and `X509_CRL_load_http()`.
394 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
398 * Validation of SM2 keys has been separated from the validation of regular EC
399 keys, allowing to improve the SM2 validation process to reject loaded private
400 keys that are not conforming to the SM2 ISO standard.
401 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
406 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
407 switches: a validation failure triggers an early exit, returning a failure
408 exit status to the parent process.
412 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
413 to ignore unknown ciphers.
417 * The `-cipher-commands` and `-digest-commands` options
418 of the command line utility `list` have been deprecated.
419 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
423 * Added convenience functions for generating asymmetric key pairs:
424 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
425 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
429 * All of the low-level EC_KEY functions have been deprecated including:
431 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
432 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
433 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
434 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
435 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
436 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
437 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
438 EC_KEY_METHOD_get_verify,
439 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
440 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
441 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
442 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
443 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
444 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
445 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
446 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
447 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
448 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
449 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
450 Applications that need to implement an EC_KEY_METHOD need to consider
451 implementation of the functionality in a special provider.
452 For replacement of the functions manipulating the EC_KEY objects
453 see the L<EVP_PKEY-EC(7)> manual page.
454 A simple way of generating EC keys is L<EVP_EC_gen(3)>.
456 Additionally functions that read and write EC_KEY objects such as
457 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
458 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
459 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
460 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
461 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
462 have also been deprecated. Applications should instead use the
463 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
465 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
466 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
467 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
468 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
469 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
471 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
473 * Deprecated all the libcrypto and libssl error string loading
474 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
475 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
476 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
477 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
478 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
479 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
480 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
481 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
482 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
483 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
485 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
486 now loads error strings automatically.
490 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
491 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
492 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
493 are to be used by servers requiring ephemeral DH keys. Instead applications
494 should consider using the built-in DH parameters that are available by
495 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
496 necessary then applications can use the alternative functions
497 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
498 replacement for the "callback" functions. The callback was originally useful
499 in order to have different parameters for export and non-export ciphersuites.
500 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
501 functions should be replaced by one of the other methods described above.
505 * The `-crypt` option to the `passwd` command line tool has been removed.
509 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
514 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
516 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
517 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
518 The inverse ciphers use AES decryption for wrapping, and
519 AES encryption for unwrapping.
523 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
524 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
525 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
526 OpenSSL 3.0 these are replaced by the more generic functions
527 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
528 The old versions have been converted to deprecated macros that just call the
533 * The security callback, which can be customised by application code, supports
534 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
535 in the "other" parameter. In most places this is what is passed. All these
536 places occur server side. However there was one client side call of this
537 security operation and it passed a DH object instead. This is incorrect
538 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
539 of the other locations. Therefore this client side call has been changed to
540 pass an EVP_PKEY instead.
544 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
545 interface. Their functionality remains unchanged.
549 * Added new option for 'openssl list', '-providers', which will display the
550 list of loaded providers, their names, version and status. It optionally
551 displays their gettable parameters.
555 * Removed EVP_PKEY_set_alias_type(). This function was previously
556 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
557 type is internally recognised so the workaround is no longer needed.
559 This is a breaking change from previous OpenSSL versions.
563 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
564 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
568 * Changed all "STACK" functions to be macros instead of inline functions. Macro
569 parameters are still checked for type safety at compile time via helper
574 * Remove the RAND_DRBG API
576 The RAND_DRBG API did not fit well into the new provider concept as
577 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
578 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
579 and some of its API calls are rather low-level. This holds in particular
580 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
582 Adding a compatibility layer to continue supporting the RAND_DRBG API as
583 a legacy API for a regular deprecation period turned out to come at the
584 price of complicating the new provider API unnecessarily. Since the
585 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
588 *Paul Dale and Matthias St. Pierre*
590 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
591 as well as actual hostnames.
595 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
596 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
597 conversely, silently ignore DTLS protocol version bounds when configuring
598 TLS-based contexts. The commands can be repeated to set bounds of both
599 types. The same applies with the corresponding "min_protocol" and
600 "max_protocol" command-line switches, in case some application uses both TLS
603 SSL_CTX instances that are created for a fixed protocol version (e.g.
604 `TLSv1_server_method()`) also silently ignore version bounds. Previously
605 attempts to apply bounds to these protocol versions would result in an
606 error. Now only the "version-flexible" SSL_CTX instances are subject to
607 limits in configuration files in command-line options.
611 * Deprecated the `ENGINE` API. Engines should be replaced with providers
616 * Reworked the recorded ERR codes to make better space for system errors.
617 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
618 given code is a system error (true) or an OpenSSL error (false).
622 * Reworked the test perl framework to better allow parallel testing.
624 *Nicola Tuveri and David von Oheimb*
626 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
627 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
631 * 'Configure' has been changed to figure out the configuration target if
632 none is given on the command line. Consequently, the 'config' script is
633 now only a mere wrapper. All documentation is changed to only mention
636 *Rich Salz and Richard Levitte*
638 * Added a library context `OSSL_LIB_CTX` that applications as well as
639 other libraries can use to form a separate context within which
640 libcrypto operations are performed.
642 There are two ways this can be used:
644 - Directly, by passing a library context to functions that take
645 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
647 - Indirectly, by creating a new library context and then assigning
648 it as the new default, with `OSSL_LIB_CTX_set0_default`.
650 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
651 apart from the functions directly related to `OSSL_LIB_CTX`, accept
652 NULL to indicate that the default library context should be used.
654 Library code that changes the default library context using
655 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
656 second call before returning to the caller.
658 _(Note: the library context was initially called `OPENSSL_CTX` and
659 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
663 * Handshake now fails if Extended Master Secret extension is dropped
668 * Dropped interactive mode from the `openssl` program. From now on,
669 running it without arguments is equivalent to `openssl help`.
673 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()` since their
674 return values were confusing: Unlike other `_cmp()` functions
675 they do not return 0 when their arguments are equal.
676 The new replacement functions `EVP_PKEY_eq()` and `EVP_PKEY_parameters_eq()`
679 *David von Oheimb and Shane Lontis*
681 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
682 `EC_GROUP_get_field_type()`.
686 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
687 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
688 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
689 Applications should rely on the library automatically assigning a suitable
690 EC_METHOD internally upon EC_GROUP construction.
694 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
695 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
696 assigned internally without application intervention.
697 Users of EC_GROUP_new() should switch to a different suitable constructor.
701 * Add CAdES-BES signature verification support, mostly derived
702 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
704 *Filipe Raimundo da Silva*
706 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
710 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
711 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
712 is to support encryption and decryption of a digital envelope that is both
713 authenticated and encrypted using AES GCM mode.
717 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
718 functions are not widely used and now OpenSSL automatically perform this
719 conversion when needed.
723 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
724 EC_KEY_precompute_mult(). These functions are not widely used and
725 applications should instead switch to named curves which OpenSSL has
726 hardcoded lookup tables for.
730 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
731 should instead use the L<EC_POINT_mul(3)> function.
735 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
736 that are not applicable to the new provider model. Applications should
737 instead use EVP_default_properties_is_fips_enabled() and
738 EVP_default_properties_enable_fips().
742 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
743 is set, an unexpected EOF is ignored, it pretends a close notify was received
744 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
748 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
749 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
750 used and applications should instead use the
751 L<EC_POINT_set_affine_coordinates(3)> and
752 L<EC_POINT_get_affine_coordinates(3)> functions.
756 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
757 arrays to be more easily constructed via a series of utility functions.
758 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
759 the various push functions and finally convert to a passable OSSL_PARAM
760 array using OSSL_PARAM_BLD_to_param().
764 * The security strength of SHA1 and MD5 based signatures in TLS has been
765 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
766 working at the default security level of 1 and instead requires security
767 level 0. The security level can be changed either using the cipher string
768 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
769 that where the signature algorithms extension is missing from a ClientHello
770 then the handshake will fail in TLS 1.2 at security level 1. This is because,
771 although this extension is optional, failing to provide one means that
772 OpenSSL will fallback to a default set of signature algorithms. This default
773 set requires the availability of SHA1.
777 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
778 contain a provider side internal key.
782 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
783 They are old functions that we don't use, and that you could disable with
784 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
788 * Project text documents not yet having a proper file name extension
789 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
790 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
794 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
795 have been converted to Markdown with the goal to produce documents
796 which not only look pretty when viewed online in the browser, but
797 remain well readable inside a plain text editor.
799 To achieve this goal, a 'minimalistic' Markdown style has been applied
800 which avoids formatting elements that interfere too much with the
801 reading flow in the text file. For example, it
803 * avoids [ATX headings][] and uses [setext headings][] instead
804 (which works for `<h1>` and `<h2>` headings only).
805 * avoids [inline links][] and uses [reference links][] instead.
806 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
808 [ATX headings]: https://github.github.com/gfm/#atx-headings
809 [setext headings]: https://github.github.com/gfm/#setext-headings
810 [inline links]: https://github.github.com/gfm/#inline-link
811 [reference links]: https://github.github.com/gfm/#reference-link
812 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
813 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
815 *Matthias St. Pierre*
817 * The test suite is changed to preserve results of each test recipe.
818 A new directory test-runs/ with subdirectories named like the
819 test recipes are created in the build tree for this purpose.
823 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
824 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
825 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
827 *David von Oheimb, Martin Peylo*
829 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
830 It supports arbitrary request and response content types, GET redirection,
831 TLS, connections via HTTP(S) proxies, connections and exchange via
832 user-defined BIOs (allowing implicit connections), persistent connections,
833 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
834 The legacy OCSP-focused (and only partly documented) API
835 is retained for backward compatibility, while most of it is deprecated.
839 * Added `util/check-format.pl`, a tool for checking adherence to the
840 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
841 The checks performed are incomplete and yield some false positives.
842 Still the tool should be useful for detecting most typical glitches.
846 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
847 If domain name resolution yields multiple IP addresses all of them are tried
848 after `connect()` failures.
852 * All of the low-level RSA functions have been deprecated including:
854 RSA_new, RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
855 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
856 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
857 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
858 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
859 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
860 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
861 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
862 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
863 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
864 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
865 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
866 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
867 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
868 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
869 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
870 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
871 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
872 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
873 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
874 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
875 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
876 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
877 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
878 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
879 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
880 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
881 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
883 Use of these low-level functions has been informally discouraged for a long
884 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
885 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
886 L<EVP_PKEY_decrypt(3)>.
887 For replacement of the functions manipulating the RSA objects
888 see the L<EVP_PKEY-RSA(7)> manual page.
889 A simple way of generating RSA keys is L<EVP_RSA_gen(3)>.
891 All of these low-level RSA functions have been deprecated without
894 RSA_blinding_off, RSA_blinding_on, RSA_clear_flags, RSA_get_version,
895 RSAPrivateKey_dup, RSAPublicKey_dup, RSA_set_flags, RSA_setup_blinding and
898 All of these RSA flags have been deprecated without replacement:
900 RSA_FLAG_BLINDING, RSA_FLAG_CACHE_PRIVATE, RSA_FLAG_CACHE_PUBLIC,
901 RSA_FLAG_EXT_PKEY, RSA_FLAG_NO_BLINDING, RSA_FLAG_THREAD_SAFE and
902 RSA_METHOD_FLAG_NO_CHECK.
906 * X509 certificates signed using SHA1 are no longer allowed at security
908 In TLS/SSL the default security level is 1. It can be set either
909 using the cipher string with `@SECLEVEL`, or calling
910 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
911 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
913 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
914 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
915 options of the commands.
919 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
920 modified to use PKEY APIs. These commands are now in maintenance mode
921 and no new features will be added to them.
925 * The command line utility rsautl has been deprecated.
926 Instead use the pkeyutl program.
930 * The command line utilities genrsa and rsa have been modified to use PKEY
931 APIs. They now write PKCS#8 keys by default. These commands are now in
932 maintenance mode and no new features will be added to them.
936 * All of the low-level DH functions have been deprecated including:
938 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
939 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
940 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
941 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
942 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
943 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
944 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
945 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
946 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
947 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
948 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
949 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
950 DH_meth_get_generate_params and DH_meth_set_generate_params.
952 Use of these low-level functions has been informally discouraged for a long
953 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
954 and L<EVP_PKEY_derive(3)>.
956 These low-level DH functions have been deprecated without replacement:
958 DH_clear_flags, DH_get_1024_160, DH_get_2048_224, DH_get_2048_256,
959 DH_set_flags and DH_test_flags.
961 The DH_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
962 The DH_FLAG_TYPE_DH and DH_FLAG_TYPE_DHX have been deprecated. Use
963 EVP_PKEY_is_a() to determine the type of a key. There is no replacement for
966 Additionally functions that read and write DH objects such as d2i_DHparams,
967 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
968 functions have also been deprecated. Applications should instead use the
969 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
971 Finally functions that assign or obtain DH objects from an EVP_PKEY such as
972 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
973 `EVP_PKEY_set1_DH()` are also deprecated.
974 Applications should instead either read or write an
975 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
976 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
978 *Paul Dale and Matt Caswell*
980 * All of the low-level DSA functions have been deprecated including:
982 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
983 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
984 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
985 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
986 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
987 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
988 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
989 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
990 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
991 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
992 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
993 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
994 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
995 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
996 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
997 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
998 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
1000 Use of these low-level functions has been informally discouraged for a long
1001 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
1002 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
1004 These low-level DSA functions have been deprecated without replacement:
1006 DSA_clear_flags, DSA_dup_DH, DSAparams_dup, DSA_set_flags and
1009 The DSA_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
1011 Finally functions that assign or obtain DSA objects from an EVP_PKEY such as
1012 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
1013 `EVP_PKEY_set1_DSA()` are also deprecated.
1014 Applications should instead either read or write an
1015 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
1016 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
1020 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1021 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. This is a breaking
1022 change from previous OpenSSL versions.
1024 Unlike in previous OpenSSL versions, this means that applications must not
1025 call `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
1026 The `EVP_PKEY_set_alias_type` function has now been removed.
1028 Parameter and key generation is also reworked to make it possible
1029 to generate EVP_PKEY_SM2 parameters and keys. Applications must now generate
1030 SM2 keys directly and must not create an EVP_PKEY_EC key first.
1034 * Deprecated low-level ECDH and ECDSA functions. These include:
1036 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
1037 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
1040 Use of these low-level functions has been informally discouraged for a long
1041 time. Instead applications should use the EVP_PKEY_derive(3),
1042 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
1046 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
1047 and EVP_PKEY_decrypt() instead.
1048 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
1049 and EVP_PKEY_encrypt() instead.
1053 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
1054 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
1055 a new formulation to include all the things it can be used for,
1056 as well as words of caution.
1060 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1061 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
1065 * All low level HMAC functions except for HMAC have been deprecated including:
1067 HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
1068 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
1069 and HMAC_CTX_get_md.
1071 Use of these low-level functions has been informally discouraged for a long
1072 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1073 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1074 and L<EVP_MAC_final(3)> or the single-shot MAC function L<EVP_Q_mac(3)>.
1076 *Paul Dale and David von Oheimb*
1078 * Over two thousand fixes were made to the documentation, including:
1079 - Common options (such as -rand/-writerand, TLS version control, etc)
1080 were refactored and point to newly-enhanced descriptions in openssl.pod.
1081 - Added style conformance for all options (with help from Richard Levitte),
1082 documented all reported missing options, added a CI build to check
1083 that all options are documented and that no unimplemented options
1085 - Documented some internals, such as all use of environment variables.
1086 - Addressed all internal broken L<> references.
1090 * All of the low-level CMAC functions have been deprecated including:
1092 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
1093 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
1095 Use of these low-level functions has been informally discouraged for a long
1096 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
1097 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
1098 and L<EVP_MAC_final(3)>.
1102 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
1103 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
1106 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
1107 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
1108 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
1109 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
1110 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
1111 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform,
1112 SHA256_Init, SHA256_Update, SHA256_Final, SHA256_Transform,
1113 SHA384_Init, SHA384_Update, SHA384_Final,
1114 SHA512_Init, SHA512_Update, SHA512_Final, SHA512_Transform,
1115 WHIRLPOOL, WHIRLPOOL_Init,
1116 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
1118 Use of these low-level functions has been informally discouraged
1119 for a long time. Applications should use the L<EVP_DigestInit_ex(3)>,
1120 L<EVP_DigestUpdate(3)>, and L<EVP_DigestFinal_ex(3)> functions instead.
1121 Alternatively, the quick one-shot function L<EVP_Q_digest(3)> can be used.
1122 SHA1, SHA224, SHA256, SHA384 and SHA512 have changed from functions to macros
1123 like this: (EVP_Q_digest(NULL, "SHA256", NULL, d, n, md, NULL) ? md : NULL).
1125 *Paul Dale and David von Oheimb*
1127 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1128 set of functions. The documentation mentioned negative values for some
1129 errors, but this was never the case, so the mention of negative values
1132 Code that followed the documentation and thereby check with something
1133 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1137 * All of the low-level cipher functions have been deprecated including:
1139 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
1140 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
1141 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
1142 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
1143 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
1144 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
1145 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
1146 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
1147 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
1148 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
1149 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
1150 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
1151 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
1152 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
1153 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
1154 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
1155 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
1156 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
1157 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
1158 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
1159 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
1160 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
1161 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
1162 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
1163 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
1164 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
1165 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
1166 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
1167 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
1169 Use of these low-level functions has been informally discouraged for
1170 a long time. Applications should use the high level EVP APIs, e.g.
1171 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
1172 equivalently named decrypt functions instead.
1174 *Matt Caswell and Paul Dale*
1176 * Removed include/openssl/opensslconf.h.in and replaced it with
1177 include/openssl/configuration.h.in, which differs in not including
1178 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1179 was added to include both.
1181 This allows internal hacks where one might need to modify the set
1182 of configured macros, for example this if deprecated symbols are
1183 still supposed to be available internally:
1185 #include <openssl/configuration.h>
1187 #undef OPENSSL_NO_DEPRECATED
1188 #define OPENSSL_SUPPRESS_DEPRECATED
1190 #include <openssl/macros.h>
1192 This should not be used by applications that use the exported
1193 symbols, as that will lead to linking errors.
1197 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1198 used in exponentiation with 512-bit moduli. No EC algorithms are
1199 affected. Analysis suggests that attacks against 2-prime RSA1024,
1200 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1201 difficult to perform and are not believed likely. Attacks against DH512
1202 are considered just feasible. However, for an attack the target would
1203 have to re-use the DH512 private key, which is not recommended anyway.
1204 Also applications directly using the low-level API BN_mod_exp may be
1205 affected if they use BN_FLG_CONSTTIME.
1210 * Most memory-debug features have been deprecated, and the functionality
1211 replaced with no-ops.
1215 * Added documentation for the STACK API.
1219 * Introduced a new method type and API, OSSL_ENCODER, to
1220 represent generic encoders. An implementation is expected to
1221 be able to encode an object associated with a given name (such
1222 as an algorithm name for an asymmetric key) into forms given by
1223 implementation properties.
1225 Encoders are primarily used from inside libcrypto, through
1226 calls to functions like EVP_PKEY_print_private(),
1227 PEM_write_bio_PrivateKey() and similar.
1229 Encoders are specified in such a way that they can be made to
1230 directly handle the provider side portion of an object, if this
1231 provider side part comes from the same provider as the encoder
1232 itself, but can also be made to handle objects in parametrized
1233 form (as an OSSL_PARAM array of data). This allows a provider to
1234 offer generic encoders as a service for any other provider.
1238 * Added a .pragma directive to the syntax of configuration files, to
1239 allow varying behavior in a supported and predictable manner.
1240 Currently added pragma:
1244 This allows dollar signs to be a keyword character unless it's
1245 followed by a opening brace or parenthesis. This is useful for
1246 platforms where dollar signs are commonly used in names, such as
1247 volume names and system directory names on VMS.
1251 * Added functionality to create an EVP_PKEY from user data. This
1252 is effectively the same as creating a RSA, DH or DSA object and
1253 then assigning them to an EVP_PKEY, but directly using algorithm
1254 agnostic EVP functions. A benefit is that this should be future
1255 proof for public key algorithms to come.
1259 * Change the interpretation of the '--api' configuration option to
1260 mean that this is a desired API compatibility level with no
1261 further meaning. The previous interpretation, that this would
1262 also mean to remove all deprecated symbols up to and including
1263 the given version, no requires that 'no-deprecated' is also used
1264 in the configuration.
1266 When building applications, the desired API compatibility level
1267 can be set with the OPENSSL_API_COMPAT macro like before. For
1268 API compatibility version below 3.0, the old style numerical
1269 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1270 For version 3.0 and on, the value is expected to be the decimal
1271 value calculated from the major and minor version like this:
1273 MAJOR * 10000 + MINOR * 100
1277 -DOPENSSL_API_COMPAT=30000 For 3.0
1278 -DOPENSSL_API_COMPAT=30200 For 3.2
1280 To hide declarations that are deprecated up to and including the
1281 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1282 given when building the application as well.
1286 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1287 access to certificate and CRL stores via URIs and OSSL_STORE
1290 This adds the following functions:
1292 - X509_LOOKUP_store()
1293 - X509_STORE_load_file()
1294 - X509_STORE_load_path()
1295 - X509_STORE_load_store()
1296 - SSL_add_store_cert_subjects_to_stack()
1297 - SSL_CTX_set_default_verify_store()
1298 - SSL_CTX_load_verify_file()
1299 - SSL_CTX_load_verify_dir()
1300 - SSL_CTX_load_verify_store()
1304 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1305 The presence of this system service is determined at run-time.
1309 * Added functionality to create an EVP_PKEY context based on data
1310 for methods from providers. This takes an algorithm name and a
1311 property query string and simply stores them, with the intent
1312 that any operation that uses this context will use those strings
1313 to fetch the needed methods implicitly, thereby making the port
1314 of application written for pre-3.0 OpenSSL easier.
1318 * The undocumented function NCONF_WIN32() has been deprecated; for
1319 conversion details see the HISTORY section of doc/man5/config.pod
1323 * Introduced the new functions EVP_DigestSignInit_ex() and
1324 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1325 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1326 pages for further details.
1330 * Over two thousand fixes were made to the documentation, including:
1331 adding missing command flags, better style conformance, documentation
1334 *Rich Salz, Richard Levitte*
1336 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1337 X25519, X448, Ed25519 and Ed448.
1341 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1346 * Deprecated the public definition of `ERR_STATE` as well as the function
1347 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1352 * Added ERR functionality to give callers access to the stored function
1353 names that have replaced the older function code based functions.
1355 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1356 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1357 ERR_peek_error_all() and ERR_peek_last_error_all().
1359 These functions have become deprecated: ERR_get_error_line(),
1360 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1361 ERR_peek_last_error_line_data() and ERR_func_error_string().
1363 Users are recommended to use ERR_get_error_all(), or to pick information
1364 with ERR_peek functions and finish off with getting the error code by using
1369 * Extended testing to be verbose for failing tests only. The make variables
1370 VERBOSE_FAILURE or VF can be used to enable this:
1372 $ make VF=1 test # Unix
1373 $ mms /macro=(VF=1) test ! OpenVMS
1374 $ nmake VF=1 test # Windows
1378 * Added the `-copy_extensions` option to the `x509` command for use with
1379 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1380 all extensions in the request are copied to the certificate or vice versa.
1382 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1384 * Added the `-copy_extensions` option to the `req` command for use with
1385 `-x509`. When given with the `copy` or `copyall` argument,
1386 all extensions in the certification request are copied to the certificate.
1390 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1391 they generate are by default RFC 5280 compliant in the following sense:
1392 There is a subjectKeyIdentifier extension with a hash value of the public key
1393 and for not self-signed certs there is an authorityKeyIdentifier extension
1394 with a keyIdentifier field or issuer information identifying the signing key.
1395 This is done unless some configuration overrides the new default behavior,
1396 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1400 * Added several checks to `X509_verify_cert()` according to requirements in
1401 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1402 (which may be done by using the CLI option `-x509_strict`):
1403 * The basicConstraints of CA certificates must be marked critical.
1404 * CA certificates must explicitly include the keyUsage extension.
1405 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1406 * The issuer name of any certificate must not be empty.
1407 * The subject name of CA certs, certs with keyUsage crlSign,
1408 and certs without subjectAlternativeName must not be empty.
1409 * If a subjectAlternativeName extension is given it must not be empty.
1410 * The signatureAlgorithm field and the cert signature must be consistent.
1411 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1412 must not be marked critical.
1413 * The authorityKeyIdentifier must be given for X.509v3 certs
1414 unless they are self-signed.
1415 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1419 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1420 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1424 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1425 used even when parsing explicit parameters, when loading a encoded key
1426 or calling `EC_GROUP_new_from_ecpkparameters()`/
1427 `EC_GROUP_new_from_ecparameters()`.
1428 This prevents bypass of security hardening and performance gains,
1429 especially for curves with specialized EC_METHODs.
1430 By default, if a key encoded with explicit parameters is loaded and later
1431 encoded, the output is still encoded with explicit parameters, even if
1432 internally a "named" EC_GROUP is used for computation.
1436 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1437 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1438 NULL. After this change, only the cofactor parameter can be NULL. It also
1439 does some minimal sanity checks on the passed order.
1444 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1445 An attack is simple, if the first CMS_recipientInfo is valid but the
1446 second CMS_recipientInfo is chosen ciphertext. If the second
1447 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1448 encryption key will be replaced by garbage, and the message cannot be
1449 decoded, but if the RSA decryption fails, the correct encryption key is
1450 used and the recipient will not notice the attack.
1451 As a work around for this potential attack the length of the decrypted
1452 key must be equal to the cipher default key length, in case the
1453 certifiate is not given and all recipientInfo are tried out.
1454 The old behaviour can be re-enabled in the CMS code by setting the
1455 CMS_DEBUG_DECRYPT flag.
1459 * Early start up entropy quality from the DEVRANDOM seed source has been
1460 improved for older Linux systems. The RAND subsystem will wait for
1461 /dev/random to be producing output before seeding from /dev/urandom.
1462 The seeded state is stored for future library initialisations using
1463 a system global shared memory segment. The shared memory identifier
1464 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1465 the desired value. The default identifier is 114.
1469 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1470 when primes for RSA keys are computed.
1471 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1472 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1473 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1474 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1475 This avoids possible fingerprinting of newly generated RSA modules.
1479 * Correct the extended master secret constant on EBCDIC systems. Without this
1480 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1481 negotiate EMS will fail. Unfortunately this also means that TLS connections
1482 between EBCDIC systems with this fix, and EBCDIC systems without this
1483 fix will fail if they negotiate EMS.
1487 * Changed the library initialisation so that the config file is now loaded
1488 by default. This was already the case for libssl. It now occurs for both
1489 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1490 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1494 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1495 where the former acts as a replacement for `ERR_put_error()`, and the
1496 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1497 `ERR_raise_data()` adds more flexibility by taking a format string and
1498 an arbitrary number of arguments following it, to be processed with
1503 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1504 to check if a named provider is loaded and available. When called, it
1505 will also activate all fallback providers if such are still present.
1509 * Enforce a minimum DH modulus size of 512 bits.
1513 * Changed DH parameters to generate the order q subgroup instead of 2q.
1514 Previously generated DH parameters are still accepted by DH_check
1515 but DH_generate_key works around that by clearing bit 0 of the
1516 private key for those. This avoids leaking bit 0 of the private key.
1520 * Significantly reduce secure memory usage by the randomness pools.
1524 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1529 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1530 algorithms. An implementation of a key exchange algorithm can be obtained
1531 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1532 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1533 the older EVP_PKEY_derive_init() function. See the man pages for the new
1534 functions for further details.
1538 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1542 * Removed the function names from error messages and deprecated the
1545 * Removed NextStep support and the macro OPENSSL_UNISTD
1549 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1550 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1551 Also removed "export var as function" capability; we do not export
1552 variables, only functions.
1556 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1557 an error and 1 indicating success. In previous versions of OpenSSL this
1558 was a void type. If a key was set longer than the maximum possible this
1563 * Support SM2 signing and verification schemes with X509 certificate.
1567 * Use SHA256 as the default digest for TS query in the `ts` app.
1571 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1572 This checks that the salt length is at least 128 bits, the derived key
1573 length is at least 112 bits, and that the iteration count is at least 1000.
1574 For backwards compatibility these checks are disabled by default in the
1575 default provider, but are enabled by default in the fips provider.
1576 To enable or disable these checks use the control
1577 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1581 * Default cipher lists/suites are now available via a function, the
1582 #defines are deprecated.
1586 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1587 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1588 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1592 * Join the directories crypto/x509 and crypto/x509v3
1596 * Added command 'openssl kdf' that uses the EVP_KDF API.
1600 * Added command 'openssl mac' that uses the EVP_MAC API.
1604 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1605 as default directories. Also added the command 'openssl info'
1606 for scripting purposes.
1610 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1611 deprecated. These undocumented functions were never integrated into the EVP
1612 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1613 Bi-directional IGE mode. These modes were never formally standardised and
1614 usage of these functions is believed to be very small. In particular
1615 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1616 is ever used. The security implications are believed to be minimal, but
1617 this issue was never fixed for backwards compatibility reasons. New code
1618 should not use these modes.
1622 * Add prediction resistance to the DRBG reseeding process.
1626 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1627 mandated by IEEE Std 1619-2018.
1631 * Added newline escaping functionality to a filename when using openssl dgst.
1632 This output format is to replicate the output format found in the `*sum`
1633 checksum programs. This aims to preserve backward compatibility.
1635 *Matt Eaton, Richard Levitte, and Paul Dale*
1637 * Removed the heartbeat message in DTLS feature, as it has very
1638 little usage and doesn't seem to fulfill a valuable purpose.
1639 The configuration option is now deprecated.
1643 * Changed the output of 'openssl {digestname} < file' to display the
1644 digest name in its output.
1648 * Added a new generic trace API which provides support for enabling
1649 instrumentation through trace output. This feature is mainly intended
1650 as an aid for developers and is disabled by default. To utilize it,
1651 OpenSSL needs to be configured with the `enable-trace` option.
1653 If the tracing API is enabled, the application can activate trace output
1654 by registering BIOs as trace channels for a number of tracing and debugging
1657 The `openssl` program has been expanded to enable any of the types
1658 available via environment variables defined by the user, and serves as
1659 one possible example on how to use this functionality.
1661 *Richard Levitte & Matthias St. Pierre*
1663 * Added build tests for C++. These are generated files that only do one
1664 thing, to include one public OpenSSL head file each. This tests that
1665 the public header files can be usefully included in a C++ application.
1667 This test isn't enabled by default. It can be enabled with the option
1668 'enable-buildtest-c++'.
1672 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1676 * Add KMAC to EVP_MAC.
1680 * Added property based algorithm implementation selection framework to
1685 * Added SCA hardening for modular field inversion in EC_GROUP through
1686 a new dedicated field_inv() pointer in EC_METHOD.
1687 This also addresses a leakage affecting conversions from projective
1688 to affine coordinates.
1690 *Billy Bob Brumley, Nicola Tuveri*
1692 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1693 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1694 those algorithms that were already supported through the EVP_PKEY API
1695 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1696 and scrypt are now wrappers that call EVP_KDF.
1700 * Build devcrypto engine as a dynamic engine.
1702 *Eneas U de Queiroz*
1704 * Add keyed BLAKE2 to EVP_MAC.
1708 * Fix a bug in the computation of the endpoint-pair shared secret used
1709 by DTLS over SCTP. This breaks interoperability with older versions
1710 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1711 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1712 interoperability with such broken implementations. However, enabling
1713 this switch breaks interoperability with correct implementations.
1715 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1716 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1720 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1724 * Changed the license to the Apache License v2.0.
1728 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1730 - Major releases (indicated by incrementing the MAJOR release number)
1731 may introduce incompatible API/ABI changes.
1732 - Minor releases (indicated by incrementing the MINOR release number)
1733 may introduce new features but retain API/ABI compatibility.
1734 - Patch releases (indicated by incrementing the PATCH number)
1735 are intended for bug fixes and other improvements of existing
1736 features only (like improving performance or adding documentation)
1737 and retain API/ABI compatibility.
1741 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1745 * Remove the 'dist' target and add a tarball building script. The
1746 'dist' target has fallen out of use, and it shouldn't be
1747 necessary to configure just to create a source distribution.
1751 * Recreate the OS390-Unix config target. It no longer relies on a
1752 special script like it did for OpenSSL pre-1.1.0.
1756 * Instead of having the source directories listed in Configure, add
1757 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1762 * Add GMAC to EVP_MAC.
1766 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1770 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1771 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1772 to facilitate the continued use of MACs through raw private keys in
1773 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1777 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1778 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1782 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1783 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1784 are retained for backwards compatibility.
1788 * AES-XTS mode now enforces that its two keys are different to mitigate
1789 the attacked described in "Efficient Instantiations of Tweakable
1790 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1791 Details of this attack can be obtained from:
1792 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1796 * Rename the object files, i.e. give them other names than in previous
1797 versions. Their names now include the name of the final product, as
1798 well as its type mnemonic (bin, lib, shlib).
1802 * Added new option for 'openssl list', '-objects', which will display the
1803 list of built in objects, i.e. OIDs with names.
1807 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1808 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1813 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1814 improves application performance by removing data copies and providing
1815 applications with zero-copy system calls such as sendfile and splice.
1819 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1820 option is set, openssl cleanses (zeroize) plaintext bytes from
1821 internal buffers after delivering them to the application. Note,
1822 the application is still responsible for cleansing other copies
1823 (e.g.: data received by SSL_read(3)).
1827 * `PKCS12_parse` now maintains the order of the parsed certificates
1828 when outputting them via `*ca` (rather than reversing it).
1832 * Deprecated pthread fork support methods. These were unused so no
1833 replacement is required.
1835 - OPENSSL_fork_prepare()
1836 - OPENSSL_fork_parent()
1837 - OPENSSL_fork_child()
1841 * Added support for FFDHE key exchange in TLS 1.3.
1848 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1850 * Fixed a problem with verifying a certificate chain when using the
1851 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1852 the certificates present in a certificate chain. It is not set by default.
1854 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1855 the chain that have explicitly encoded elliptic curve parameters was added
1856 as an additional strict check.
1858 An error in the implementation of this check meant that the result of a
1859 previous check to confirm that certificates in the chain are valid CA
1860 certificates was overwritten. This effectively bypasses the check
1861 that non-CA certificates must not be able to issue other certificates.
1863 If a "purpose" has been configured then there is a subsequent opportunity
1864 for checks that the certificate is a valid CA. All of the named "purpose"
1865 values implemented in libcrypto perform this check. Therefore, where
1866 a purpose is set the certificate chain will still be rejected even when the
1867 strict flag has been used. A purpose is set by default in libssl client and
1868 server certificate verification routines, but it can be overridden or
1869 removed by an application.
1871 In order to be affected, an application must explicitly set the
1872 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1873 for the certificate verification or, in the case of TLS client or server
1874 applications, override the default purpose.
1879 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1880 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1881 renegotiation ClientHello omits the signature_algorithms extension (where it
1882 was present in the initial ClientHello), but includes a
1883 signature_algorithms_cert extension then a NULL pointer dereference will
1884 result, leading to a crash and a denial of service attack.
1886 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1887 (which is the default configuration). OpenSSL TLS clients are not impacted by
1891 *Peter Kästle and Samuel Sapalski*
1893 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1895 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1896 create a unique hash value based on the issuer and serial number data
1897 contained within an X509 certificate. However it was failing to correctly
1898 handle any errors that may occur while parsing the issuer field (which might
1899 occur if the issuer field is maliciously constructed). This may subsequently
1900 result in a NULL pointer deref and a crash leading to a potential denial of
1906 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1907 padding mode to correctly check for rollback attacks. This is considered a
1908 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1913 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1914 functions. Previously they could overflow the output length argument in some
1915 cases where the input length is close to the maximum permissable length for
1916 an integer on the platform. In such cases the return value from the function
1917 call would be 1 (indicating success), but the output length value would be
1918 negative. This could cause applications to behave incorrectly or crash.
1923 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1924 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1925 could be exploited in a side channel attack to recover the password. Since
1926 the attack is local host only this is outside of the current OpenSSL
1927 threat model and therefore no CVE is assigned.
1929 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1934 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1936 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1937 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1938 If an attacker can control both items being compared then this could lead
1939 to a possible denial of service attack. OpenSSL itself uses the
1940 GENERAL_NAME_cmp function for two purposes:
1941 1) Comparing CRL distribution point names between an available CRL and a
1942 CRL distribution point embedded in an X509 certificate
1943 2) When verifying that a timestamp response token signer matches the
1944 timestamp authority name (exposed via the API functions
1945 TS_RESP_verify_response and TS_RESP_verify_token)
1950 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1952 * Certificates with explicit curve parameters are now disallowed in
1953 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1957 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1958 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1959 conversely, silently ignore DTLS protocol version bounds when configuring
1960 TLS-based contexts. The commands can be repeated to set bounds of both
1961 types. The same applies with the corresponding "min_protocol" and
1962 "max_protocol" command-line switches, in case some application uses both TLS
1965 SSL_CTX instances that are created for a fixed protocol version (e.g.
1966 TLSv1_server_method()) also silently ignore version bounds. Previously
1967 attempts to apply bounds to these protocol versions would result in an
1968 error. Now only the "version-flexible" SSL_CTX instances are subject to
1969 limits in configuration files in command-line options.
1973 * Handshake now fails if Extended Master Secret extension is dropped
1978 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1980 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1982 * Fixed segmentation fault in SSL_check_chain()
1983 Server or client applications that call the SSL_check_chain() function
1984 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1985 dereference as a result of incorrect handling of the
1986 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1987 or unrecognised signature algorithm is received from the peer. This could
1988 be exploited by a malicious peer in a Denial of Service attack.
1993 * Added AES consttime code for no-asm configurations
1994 an optional constant time support for AES was added
1995 when building openssl for no-asm.
1996 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1997 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1998 At this time this feature is by default disabled.
1999 It will be enabled by default in 3.0.
2003 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2005 * Revert the change of EOF detection while reading in libssl to avoid
2006 regressions in applications depending on the current way of reporting
2007 the EOF. As the existing method is not fully accurate the change to
2008 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2009 branch and will be present in the 3.0 release.
2013 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2014 when primes for RSA keys are computed.
2015 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2016 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2017 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2018 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2019 This avoids possible fingerprinting of newly generated RSA modules.
2023 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2025 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2026 while reading in libssl then we would report an error back to the
2027 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2028 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2029 therefore give a hint as to what went wrong.
2033 * Check that ed25519 and ed448 are allowed by the security level. Previously
2034 signature algorithms not using an MD were not being checked that they were
2035 allowed by the security level.
2039 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2040 was not quite right. The behaviour was not consistent between resumption
2041 and normal handshakes, and also not quite consistent with historical
2042 behaviour. The behaviour in various scenarios has been clarified and
2043 it has been updated to make it match historical behaviour as closely as
2048 * *[VMS only]* The header files that the VMS compilers include automatically,
2049 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2050 that the C++ compiler doesn't understand. This is a shortcoming in the
2051 compiler, but can be worked around with `__cplusplus` guards.
2053 C++ applications that use OpenSSL libraries must be compiled using the
2054 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2055 functions. Otherwise, only functions with symbols of less than 31
2056 characters can be used, as the linker will not be able to successfully
2057 resolve symbols with longer names.
2061 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2062 The presence of this system service is determined at run-time.
2066 * Added newline escaping functionality to a filename when using openssl dgst.
2067 This output format is to replicate the output format found in the `*sum`
2068 checksum programs. This aims to preserve backward compatibility.
2070 *Matt Eaton, Richard Levitte, and Paul Dale*
2072 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2077 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2079 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2080 number generator (RNG). This was intended to include protection in the
2081 event of a fork() system call in order to ensure that the parent and child
2082 processes did not share the same RNG state. However this protection was not
2083 being used in the default case.
2085 A partial mitigation for this issue is that the output from a high
2086 precision timer is mixed into the RNG state so the likelihood of a parent
2087 and child process sharing state is significantly reduced.
2089 If an application already calls OPENSSL_init_crypto() explicitly using
2090 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2093 *Matthias St. Pierre*
2095 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2096 used even when parsing explicit parameters, when loading a encoded key
2097 or calling `EC_GROUP_new_from_ecpkparameters()`/
2098 `EC_GROUP_new_from_ecparameters()`.
2099 This prevents bypass of security hardening and performance gains,
2100 especially for curves with specialized EC_METHODs.
2101 By default, if a key encoded with explicit parameters is loaded and later
2102 encoded, the output is still encoded with explicit parameters, even if
2103 internally a "named" EC_GROUP is used for computation.
2107 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2108 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2109 NULL. After this change, only the cofactor parameter can be NULL. It also
2110 does some minimal sanity checks on the passed order.
2115 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2116 An attack is simple, if the first CMS_recipientInfo is valid but the
2117 second CMS_recipientInfo is chosen ciphertext. If the second
2118 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2119 encryption key will be replaced by garbage, and the message cannot be
2120 decoded, but if the RSA decryption fails, the correct encryption key is
2121 used and the recipient will not notice the attack.
2122 As a work around for this potential attack the length of the decrypted
2123 key must be equal to the cipher default key length, in case the
2124 certifiate is not given and all recipientInfo are tried out.
2125 The old behaviour can be re-enabled in the CMS code by setting the
2126 CMS_DEBUG_DECRYPT flag.
2131 * Early start up entropy quality from the DEVRANDOM seed source has been
2132 improved for older Linux systems. The RAND subsystem will wait for
2133 /dev/random to be producing output before seeding from /dev/urandom.
2134 The seeded state is stored for future library initialisations using
2135 a system global shared memory segment. The shared memory identifier
2136 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2137 the desired value. The default identifier is 114.
2141 * Correct the extended master secret constant on EBCDIC systems. Without this
2142 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2143 negotiate EMS will fail. Unfortunately this also means that TLS connections
2144 between EBCDIC systems with this fix, and EBCDIC systems without this
2145 fix will fail if they negotiate EMS.
2149 * Use Windows installation paths in the mingw builds
2151 Mingw isn't a POSIX environment per se, which means that Windows
2152 paths should be used for installation.
2157 * Changed DH_check to accept parameters with order q and 2q subgroups.
2158 With order 2q subgroups the bit 0 of the private key is not secret
2159 but DH_generate_key works around that by clearing bit 0 of the
2160 private key for those. This avoids leaking bit 0 of the private key.
2164 * Significantly reduce secure memory usage by the randomness pools.
2168 * Revert the DEVRANDOM_WAIT feature for Linux systems
2170 The DEVRANDOM_WAIT feature added a select() call to wait for the
2171 /dev/random device to become readable before reading from the
2172 /dev/urandom device.
2174 It turned out that this change had negative side effects on
2175 performance which were not acceptable. After some discussion it
2176 was decided to revert this feature and leave it up to the OS
2177 resp. the platform maintainer to ensure a proper initialization
2178 during early boot time.
2180 *Matthias St. Pierre*
2182 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2184 * Add build tests for C++. These are generated files that only do one
2185 thing, to include one public OpenSSL head file each. This tests that
2186 the public header files can be usefully included in a C++ application.
2188 This test isn't enabled by default. It can be enabled with the option
2189 'enable-buildtest-c++'.
2193 * Enable SHA3 pre-hashing for ECDSA and DSA.
2197 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2198 This changes the size when using the `genpkey` command when no size is given.
2199 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2200 generation commands to use 2048 bits by default.
2204 * Reorganize the manual pages to consistently have RETURN VALUES,
2205 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2206 util/fix-doc-nits accordingly.
2208 *Paul Yang, Joshua Lock*
2210 * Add the missing accessor EVP_PKEY_get0_engine()
2214 * Have commands like `s_client` and `s_server` output the signature scheme
2215 along with other cipher suite parameters when debugging.
2219 * Make OPENSSL_config() error agnostic again.
2223 * Do the error handling in RSA decryption constant time.
2227 * Prevent over long nonces in ChaCha20-Poly1305.
2229 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2230 for every encryption operation. RFC 7539 specifies that the nonce value
2231 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2232 and front pads the nonce with 0 bytes if it is less than 12
2233 bytes. However it also incorrectly allows a nonce to be set of up to 16
2234 bytes. In this case only the last 12 bytes are significant and any
2235 additional leading bytes are ignored.
2237 It is a requirement of using this cipher that nonce values are
2238 unique. Messages encrypted using a reused nonce value are susceptible to
2239 serious confidentiality and integrity attacks. If an application changes
2240 the default nonce length to be longer than 12 bytes and then makes a
2241 change to the leading bytes of the nonce expecting the new value to be a
2242 new unique nonce then such an application could inadvertently encrypt
2243 messages with a reused nonce.
2245 Additionally the ignored bytes in a long nonce are not covered by the
2246 integrity guarantee of this cipher. Any application that relies on the
2247 integrity of these ignored leading bytes of a long nonce may be further
2248 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2249 is safe because no such use sets such a long nonce value. However user
2250 applications that use this cipher directly and set a non-default nonce
2251 length to be longer than 12 bytes may be vulnerable.
2253 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2259 * Add DEVRANDOM_WAIT feature for Linux systems
2261 On older Linux systems where the getrandom() system call is not available,
2262 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2263 Contrary to getrandom(), the /dev/urandom device will not block during
2264 early boot when the kernel CSPRNG has not been seeded yet.
2266 To mitigate this known weakness, use select() to wait for /dev/random to
2267 become readable before reading from /dev/urandom.
2269 * Ensure that SM2 only uses SM3 as digest algorithm
2273 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2275 * Change the info callback signals for the start and end of a post-handshake
2276 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2277 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2278 confused by this and assume that a TLSv1.2 renegotiation has started. This
2279 can break KeyUpdate handling. Instead we no longer signal the start and end
2280 of a post handshake message exchange (although the messages themselves are
2281 still signalled). This could break some applications that were expecting
2282 the old signals. However without this KeyUpdate is not usable for many
2287 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2289 * Timing vulnerability in DSA signature generation
2291 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2292 timing side channel attack. An attacker could use variations in the signing
2293 algorithm to recover the private key.
2295 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2300 * Timing vulnerability in ECDSA signature generation
2302 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2303 timing side channel attack. An attacker could use variations in the signing
2304 algorithm to recover the private key.
2306 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2311 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2312 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2313 of two gigabytes and the error handling improved.
2315 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2316 categorized as a normal bug, not a security issue, because the DRBG reseeds
2317 automatically and is fully functional even without additional randomness
2318 provided by the application.
2320 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2322 * Add a new ClientHello callback. Provides a callback interface that gives
2323 the application the ability to adjust the nascent SSL object at the
2324 earliest stage of ClientHello processing, immediately after extensions have
2325 been collected but before they have been processed. In particular, this
2326 callback can adjust the supported TLS versions in response to the contents
2331 * Add SM2 base algorithm support.
2335 * s390x assembly pack: add (improved) hardware-support for the following
2336 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2337 aes-cfb/cfb8, aes-ecb.
2341 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2342 parameter is no longer accepted, as it leads to a corrupt table. NULL
2343 pem_str is reserved for alias entries only.
2347 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2348 step for prime curves. The new implementation is based on formulae from
2349 differential addition-and-doubling in homogeneous projective coordinates
2350 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2351 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2352 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2353 to work in projective coordinates.
2355 *Billy Bob Brumley, Nicola Tuveri*
2357 * Change generating and checking of primes so that the error rate of not
2358 being prime depends on the intended use based on the size of the input.
2359 For larger primes this will result in more rounds of Miller-Rabin.
2360 The maximal error rate for primes with more than 1080 bits is lowered
2363 *Kurt Roeckx, Annie Yousar*
2365 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2369 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2370 moving between systems, and to avoid confusion when a Windows build is
2371 done with mingw vs with MSVC. For POSIX installs, there's still a
2372 symlink or copy named 'tsget' to avoid that confusion as well.
2376 * Revert blinding in ECDSA sign and instead make problematic addition
2377 length-invariant. Switch even to fixed-length Montgomery multiplication.
2381 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2382 step for binary curves. The new implementation is based on formulae from
2383 differential addition-and-doubling in mixed Lopez-Dahab projective
2384 coordinates, modified to independently blind the operands.
2386 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2388 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2389 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2390 EC_METHODs to implement their own specialized "ladder step", to take
2391 advantage of more favorable coordinate systems or more efficient
2392 differential addition-and-doubling algorithms.
2394 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2396 * Modified the random device based seed sources to keep the relevant
2397 file descriptors open rather than reopening them on each access.
2398 This allows such sources to operate in a chroot() jail without
2399 the associated device nodes being available. This behaviour can be
2400 controlled using RAND_keep_random_devices_open().
2404 * Numerous side-channel attack mitigations have been applied. This may have
2405 performance impacts for some algorithms for the benefit of improved
2406 security. Specific changes are noted in this change log by their respective
2411 * AIX shared library support overhaul. Switch to AIX "natural" way of
2412 handling shared libraries, which means collecting shared objects of
2413 different versions and bitnesses in one common archive. This allows to
2414 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2415 doesn't affect the way 3rd party applications are linked, only how
2416 multi-version installation is managed.
2420 * Make ec_group_do_inverse_ord() more robust and available to other
2421 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2422 mitigations are applied to the fallback BN_mod_inverse().
2423 When using this function rather than BN_mod_inverse() directly, new
2424 EC cryptosystem implementations are then safer-by-default.
2428 * Add coordinate blinding for EC_POINT and implement projective
2429 coordinate blinding for generic prime curves as a countermeasure to
2430 chosen point SCA attacks.
2432 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2434 * Add blinding to ECDSA and DSA signatures to protect against side channel
2435 attacks discovered by Keegan Ryan (NCC Group).
2439 * Enforce checking in the `pkeyutl` command to ensure that the input
2440 length does not exceed the maximum supported digest length when performing
2441 a sign, verify or verifyrecover operation.
2445 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2446 I/O in combination with something like select() or poll() will hang. This
2447 can be turned off again using SSL_CTX_clear_mode().
2448 Many applications do not properly handle non-application data records, and
2449 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2450 around the problems in those applications, but can also break some.
2451 It's recommended to read the manpages about SSL_read(), SSL_write(),
2452 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2453 SSL_CTX_set_read_ahead() again.
2457 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2458 now allow empty (zero character) pass phrases.
2462 * Apply blinding to binary field modular inversion and remove patent
2463 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2467 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2468 binary and prime elliptic curves.
2472 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2473 constant time fixed point multiplication.
2477 * Revise elliptic curve scalar multiplication with timing attack
2478 defenses: ec_wNAF_mul redirects to a constant time implementation
2479 when computing fixed point and variable point multiplication (which
2480 in OpenSSL are mostly used with secret scalars in keygen, sign,
2481 ECDH derive operations).
2482 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2485 * Updated CONTRIBUTING
2489 * Updated DRBG / RAND to request nonce and additional low entropy
2490 randomness from the system.
2492 *Matthias St. Pierre*
2494 * Updated 'openssl rehash' to use OpenSSL consistent default.
2498 * Moved the load of the ssl_conf module to libcrypto, which helps
2499 loading engines that libssl uses before libssl is initialised.
2503 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2507 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2509 *Ingo Schwarze, Rich Salz*
2511 * Added output of accepting IP address and port for 'openssl s_server'
2515 * Added a new API for TLSv1.3 ciphersuites:
2516 SSL_CTX_set_ciphersuites()
2517 SSL_set_ciphersuites()
2521 * Memory allocation failures consistently add an error to the error
2526 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2527 in libcrypto when run as setuid/setgid.
2531 * Load any config file by default when libssl is used.
2535 * Added new public header file <openssl/rand_drbg.h> and documentation
2536 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2538 *Matthias St. Pierre*
2540 * QNX support removed (cannot find contributors to get their approval
2541 for the license change).
2545 * TLSv1.3 replay protection for early data has been implemented. See the
2546 SSL_read_early_data() man page for further details.
2550 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2551 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2552 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2553 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2554 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2555 configuration has been separated out. See the ciphers man page or the
2556 SSL_CTX_set_ciphersuites() man page for more information.
2560 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2561 in responder mode now supports the new "-multi" option, which
2562 spawns the specified number of child processes to handle OCSP
2563 requests. The "-timeout" option now also limits the OCSP
2564 responder's patience to wait to receive the full client request
2565 on a newly accepted connection. Child processes are respawned
2566 as needed, and the CA index file is automatically reloaded
2567 when changed. This makes it possible to run the "ocsp" responder
2568 as a long-running service, making the OpenSSL CA somewhat more
2569 feature-complete. In this mode, most diagnostic messages logged
2570 after entering the event loop are logged via syslog(3) rather than
2575 * Added support for X448 and Ed448. Heavily based on original work by
2580 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2581 objects loaded. This adds the functions OSSL_STORE_expect() and
2582 OSSL_STORE_find() as well as needed tools to construct searches and
2583 get the search data out of them.
2587 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2588 version of OpenSSL should review their configuration settings to ensure
2589 that they are still appropriate for TLSv1.3. For further information see:
2590 <https://wiki.openssl.org/index.php/TLS1.3>
2594 * Grand redesign of the OpenSSL random generator
2596 The default RAND method now utilizes an AES-CTR DRBG according to
2597 NIST standard SP 800-90Ar1. The new random generator is essentially
2598 a port of the default random generator from the OpenSSL FIPS 2.0
2599 object module. It is a hybrid deterministic random bit generator
2600 using an AES-CTR bit stream and which seeds and reseeds itself
2601 automatically using trusted system entropy sources.
2603 Some of its new features are:
2604 - Support for multiple DRBG instances with seed chaining.
2605 - The default RAND method makes use of a DRBG.
2606 - There is a public and private DRBG instance.
2607 - The DRBG instances are fork-safe.
2608 - Keep all global DRBG instances on the secure heap if it is enabled.
2609 - The public and private DRBG instance are per thread for lock free
2612 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2614 * Changed Configure so it only says what it does and doesn't dump
2615 so much data. Instead, ./configdata.pm should be used as a script
2616 to display all sorts of configuration data.
2620 * Added processing of "make variables" to Configure.
2624 * Added SHA512/224 and SHA512/256 algorithm support.
2628 * The last traces of Netware support, first removed in 1.1.0, have
2633 * Get rid of Makefile.shared, and in the process, make the processing
2634 of certain files (rc.obj, or the .def/.map/.opt files produced from
2635 the ordinal files) more visible and hopefully easier to trace and
2636 debug (or make silent).
2640 * Make it possible to have environment variable assignments as
2641 arguments to config / Configure.
2645 * Add multi-prime RSA (RFC 8017) support.
2649 * Add SM3 implemented according to GB/T 32905-2016
2650 *Jack Lloyd <jack.lloyd@ribose.com>,*
2651 *Ronald Tse <ronald.tse@ribose.com>,*
2652 *Erick Borsboom <erick.borsboom@ribose.com>*
2654 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2655 as documented in RFC6066.
2656 Based on a patch from Tomasz Moń
2658 *Filipe Raimundo da Silva*
2660 * Add SM4 implemented according to GB/T 32907-2016.
2661 *Jack Lloyd <jack.lloyd@ribose.com>,*
2662 *Ronald Tse <ronald.tse@ribose.com>,*
2663 *Erick Borsboom <erick.borsboom@ribose.com>*
2665 * Reimplement -newreq-nodes and ERR_error_string_n; the
2666 original author does not agree with the license change.
2670 * Add ARIA AEAD TLS support.
2674 * Some macro definitions to support VS6 have been removed. Visual
2675 Studio 6 has not worked since 1.1.0
2679 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2680 without clearing the errors.
2684 * Add "atfork" functions. If building on a system that without
2685 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2686 requirements. The RAND facility now uses/requires this.
2694 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2695 not possible to disable entirely. However, it's still possible to
2696 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2699 To disable, configure with 'no-ui-console'. 'no-ui' is still
2700 possible to use as an alias. Check at compile time with the
2701 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2702 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2706 * Add a STORE module, which implements a uniform and URI based reader of
2707 stores that can contain keys, certificates, CRLs and numerous other
2708 objects. The main API is loosely based on a few stdio functions,
2709 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2710 OSSL_STORE_error and OSSL_STORE_close.
2711 The implementation uses backends called "loaders" to implement arbitrary
2712 URI schemes. There is one built in "loader" for the 'file' scheme.
2716 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2717 then adjusted to work on FreeBSD 8.4 as well.
2718 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2719 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2723 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2724 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2725 error code calls like this:
2727 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2729 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2730 that can be encoded in C. For the foreseeable future, this will only
2733 *Richard Levitte and Tim Hudson*
2735 * Removed BSD cryptodev engine.
2739 * Add a build target 'build_all_generated', to build all generated files
2740 and only that. This can be used to prepare everything that requires
2741 things like perl for a system that lacks perl and then move everything
2742 to that system and do the rest of the build there.
2746 * In the UI interface, make it possible to duplicate the user data. This
2747 can be used by engines that need to retain the data for a longer time
2748 than just the call where this user data is passed.
2752 * Ignore the '-named_curve auto' value for compatibility of applications
2755 *Tomáš Mráz <tmraz@fedoraproject.org>*
2757 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2758 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2759 alerts across multiple records (some of which could be empty). In practice
2760 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2761 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2762 support this at all. Supporting it adds significant complexity to the
2763 record layer, and its removal is unlikely to cause interoperability
2768 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2769 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2770 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2775 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2776 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2778 *Richard Levitte, Andy Polyakov*
2780 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2785 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2786 platform rather than 'mingw'.
2790 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2791 success if they are asked to add an object which already exists
2792 in the store. This change cascades to other functions which load
2793 certificates and CRLs.
2797 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2798 facilitate stack unwinding even from assembly subroutines.
2802 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2803 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2807 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2808 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2809 which is the minimum version we support.
2813 * Certificate time validation (X509_cmp_time) enforces stricter
2814 compliance with RFC 5280. Fractional seconds and timezone offsets
2815 are no longer allowed.
2819 * Add support for ARIA
2823 * s_client will now send the Server Name Indication (SNI) extension by
2824 default unless the new "-noservername" option is used. The server name is
2825 based on the host provided to the "-connect" option unless overridden by
2826 using "-servername".
2830 * Add support for SipHash
2834 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2835 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2836 prevent issues where no progress is being made and the peer continually
2837 sends unrecognised record types, using up resources processing them.
2841 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2842 using the algorithm defined in
2843 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2847 * Heartbeat support has been removed; the ABI is changed for now.
2849 *Richard Levitte, Rich Salz*
2851 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2855 * The RSA "null" method, which was partially supported to avoid patent
2856 issues, has been replaced to always returns NULL.
2863 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2865 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2866 used even when parsing explicit parameters, when loading a encoded key
2867 or calling `EC_GROUP_new_from_ecpkparameters()`/
2868 `EC_GROUP_new_from_ecparameters()`.
2869 This prevents bypass of security hardening and performance gains,
2870 especially for curves with specialized EC_METHODs.
2871 By default, if a key encoded with explicit parameters is loaded and later
2872 encoded, the output is still encoded with explicit parameters, even if
2873 internally a "named" EC_GROUP is used for computation.
2877 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2878 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2879 NULL. After this change, only the cofactor parameter can be NULL. It also
2880 does some minimal sanity checks on the passed order.
2885 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2886 An attack is simple, if the first CMS_recipientInfo is valid but the
2887 second CMS_recipientInfo is chosen ciphertext. If the second
2888 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2889 encryption key will be replaced by garbage, and the message cannot be
2890 decoded, but if the RSA decryption fails, the correct encryption key is
2891 used and the recipient will not notice the attack.
2892 As a work around for this potential attack the length of the decrypted
2893 key must be equal to the cipher default key length, in case the
2894 certifiate is not given and all recipientInfo are tried out.
2895 The old behaviour can be re-enabled in the CMS code by setting the
2896 CMS_DEBUG_DECRYPT flag.
2901 * Use Windows installation paths in the mingw builds
2903 Mingw isn't a POSIX environment per se, which means that Windows
2904 paths should be used for installation.
2909 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2911 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2912 This changes the size when using the `genpkey` command when no size is given.
2913 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2914 generation commands to use 2048 bits by default.
2918 * Prevent over long nonces in ChaCha20-Poly1305.
2920 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2921 for every encryption operation. RFC 7539 specifies that the nonce value
2922 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2923 and front pads the nonce with 0 bytes if it is less than 12
2924 bytes. However it also incorrectly allows a nonce to be set of up to 16
2925 bytes. In this case only the last 12 bytes are significant and any
2926 additional leading bytes are ignored.
2928 It is a requirement of using this cipher that nonce values are
2929 unique. Messages encrypted using a reused nonce value are susceptible to
2930 serious confidentiality and integrity attacks. If an application changes
2931 the default nonce length to be longer than 12 bytes and then makes a
2932 change to the leading bytes of the nonce expecting the new value to be a
2933 new unique nonce then such an application could inadvertently encrypt
2934 messages with a reused nonce.
2936 Additionally the ignored bytes in a long nonce are not covered by the
2937 integrity guarantee of this cipher. Any application that relies on the
2938 integrity of these ignored leading bytes of a long nonce may be further
2939 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2940 is safe because no such use sets such a long nonce value. However user
2941 applications that use this cipher directly and set a non-default nonce
2942 length to be longer than 12 bytes may be vulnerable.
2944 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2950 * Added SCA hardening for modular field inversion in EC_GROUP through
2951 a new dedicated field_inv() pointer in EC_METHOD.
2952 This also addresses a leakage affecting conversions from projective
2953 to affine coordinates.
2955 *Billy Bob Brumley, Nicola Tuveri*
2957 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2958 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2962 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2966 * Remove the 'dist' target and add a tarball building script. The
2967 'dist' target has fallen out of use, and it shouldn't be
2968 necessary to configure just to create a source distribution.
2972 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2974 * Timing vulnerability in DSA signature generation
2976 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2977 timing side channel attack. An attacker could use variations in the signing
2978 algorithm to recover the private key.
2980 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2985 * Timing vulnerability in ECDSA signature generation
2987 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2988 timing side channel attack. An attacker could use variations in the signing
2989 algorithm to recover the private key.
2991 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2996 * Add coordinate blinding for EC_POINT and implement projective
2997 coordinate blinding for generic prime curves as a countermeasure to
2998 chosen point SCA attacks.
3000 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3002 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3004 * Client DoS due to large DH parameter
3006 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3007 malicious server can send a very large prime value to the client. This will
3008 cause the client to spend an unreasonably long period of time generating a
3009 key for this prime resulting in a hang until the client has finished. This
3010 could be exploited in a Denial Of Service attack.
3012 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3017 * Cache timing vulnerability in RSA Key Generation
3019 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3020 a cache timing side channel attack. An attacker with sufficient access to
3021 mount cache timing attacks during the RSA key generation process could
3022 recover the private key.
3024 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3025 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3030 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3031 parameter is no longer accepted, as it leads to a corrupt table. NULL
3032 pem_str is reserved for alias entries only.
3036 * Revert blinding in ECDSA sign and instead make problematic addition
3037 length-invariant. Switch even to fixed-length Montgomery multiplication.
3041 * Change generating and checking of primes so that the error rate of not
3042 being prime depends on the intended use based on the size of the input.
3043 For larger primes this will result in more rounds of Miller-Rabin.
3044 The maximal error rate for primes with more than 1080 bits is lowered
3047 *Kurt Roeckx, Annie Yousar*
3049 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3053 * Add blinding to ECDSA and DSA signatures to protect against side channel
3054 attacks discovered by Keegan Ryan (NCC Group).
3058 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3059 now allow empty (zero character) pass phrases.
3063 * Certificate time validation (X509_cmp_time) enforces stricter
3064 compliance with RFC 5280. Fractional seconds and timezone offsets
3065 are no longer allowed.
3069 * Fixed a text canonicalisation bug in CMS
3071 Where a CMS detached signature is used with text content the text goes
3072 through a canonicalisation process first prior to signing or verifying a
3073 signature. This process strips trailing space at the end of lines, converts
3074 line terminators to CRLF and removes additional trailing line terminators
3075 at the end of a file. A bug in the canonicalisation process meant that
3076 some characters, such as form-feed, were incorrectly treated as whitespace
3077 and removed. This is contrary to the specification (RFC5485). This fix
3078 could mean that detached text data signed with an earlier version of
3079 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3080 signed with a fixed OpenSSL may fail to verify with an earlier version of
3081 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3082 and use the "-binary" flag (for the "cms" command line application) or set
3083 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3087 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3089 * Constructed ASN.1 types with a recursive definition could exceed the stack
3091 Constructed ASN.1 types with a recursive definition (such as can be found
3092 in PKCS7) could eventually exceed the stack given malicious input with
3093 excessive recursion. This could result in a Denial Of Service attack. There
3094 are no such structures used within SSL/TLS that come from untrusted sources
3095 so this is considered safe.
3097 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3103 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3105 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3106 effectively reduced to only comparing the least significant bit of each
3107 byte. This allows an attacker to forge messages that would be considered as
3108 authenticated in an amount of tries lower than that guaranteed by the
3109 security claims of the scheme. The module can only be compiled by the
3110 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3112 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3118 * Add a build target 'build_all_generated', to build all generated files
3119 and only that. This can be used to prepare everything that requires
3120 things like perl for a system that lacks perl and then move everything
3121 to that system and do the rest of the build there.
3125 * Backport SSL_OP_NO_RENGOTIATION
3127 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3128 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3129 changes this is no longer possible in 1.1.0. Therefore the new
3130 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3131 1.1.0 to provide equivalent functionality.
3133 Note that if an application built against 1.1.0h headers (or above) is run
3134 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3135 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3139 * Removed the OS390-Unix config target. It relied on a script that doesn't
3144 * rsaz_1024_mul_avx2 overflow bug on x86_64
3146 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3147 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3148 Analysis suggests that attacks against RSA and DSA as a result of this
3149 defect would be very difficult to perform and are not believed likely.
3150 Attacks against DH1024 are considered just feasible, because most of the
3151 work necessary to deduce information about a private key may be performed
3152 offline. The amount of resources required for such an attack would be
3153 significant. However, for an attack on TLS to be meaningful, the server
3154 would have to share the DH1024 private key among multiple clients, which is
3155 no longer an option since CVE-2016-0701.
3157 This only affects processors that support the AVX2 but not ADX extensions
3158 like Intel Haswell (4th generation).
3160 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3161 was originally found via the OSS-Fuzz project.
3166 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3168 * bn_sqrx8x_internal carry bug on x86_64
3170 There is a carry propagating bug in the x86_64 Montgomery squaring
3171 procedure. No EC algorithms are affected. Analysis suggests that attacks
3172 against RSA and DSA as a result of this defect would be very difficult to
3173 perform and are not believed likely. Attacks against DH are considered just
3174 feasible (although very difficult) because most of the work necessary to
3175 deduce information about a private key may be performed offline. The amount
3176 of resources required for such an attack would be very significant and
3177 likely only accessible to a limited number of attackers. An attacker would
3178 additionally need online access to an unpatched system using the target
3179 private key in a scenario with persistent DH parameters and a private
3180 key that is shared between multiple clients.
3182 This only affects processors that support the BMI1, BMI2 and ADX extensions
3183 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3185 This issue was reported to OpenSSL by the OSS-Fuzz project.
3190 * Malformed X.509 IPAddressFamily could cause OOB read
3192 If an X.509 certificate has a malformed IPAddressFamily extension,
3193 OpenSSL could do a one-byte buffer overread. The most likely result
3194 would be an erroneous display of the certificate in text format.
3196 This issue was reported to OpenSSL by the OSS-Fuzz project.
3201 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3203 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3204 platform rather than 'mingw'.
3208 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3209 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3210 which is the minimum version we support.
3214 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3216 * Encrypt-Then-Mac renegotiation crash
3218 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3219 negotiated where it was not in the original handshake (or vice-versa) then
3220 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3221 and servers are affected.
3223 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3228 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3230 * Truncated packet could crash via OOB read
3232 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3233 cipher is being used, then a truncated packet can cause that host to
3234 perform an out-of-bounds read, usually resulting in a crash.
3236 This issue was reported to OpenSSL by Robert Święcki of Google.
3241 * Bad (EC)DHE parameters cause a client crash
3243 If a malicious server supplies bad parameters for a DHE or ECDHE key
3244 exchange then this can result in the client attempting to dereference a
3245 NULL pointer leading to a client crash. This could be exploited in a Denial
3248 This issue was reported to OpenSSL by Guido Vranken.
3253 * BN_mod_exp may produce incorrect results on x86_64
3255 There is a carry propagating bug in the x86_64 Montgomery squaring
3256 procedure. No EC algorithms are affected. Analysis suggests that attacks
3257 against RSA and DSA as a result of this defect would be very difficult to
3258 perform and are not believed likely. Attacks against DH are considered just
3259 feasible (although very difficult) because most of the work necessary to
3260 deduce information about a private key may be performed offline. The amount
3261 of resources required for such an attack would be very significant and
3262 likely only accessible to a limited number of attackers. An attacker would
3263 additionally need online access to an unpatched system using the target
3264 private key in a scenario with persistent DH parameters and a private
3265 key that is shared between multiple clients. For example this can occur by
3266 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3267 similar to CVE-2015-3193 but must be treated as a separate problem.
3269 This issue was reported to OpenSSL by the OSS-Fuzz project.
3274 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3276 * ChaCha20/Poly1305 heap-buffer-overflow
3278 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3279 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3280 crash. This issue is not considered to be exploitable beyond a DoS.
3282 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3287 * CMS Null dereference
3289 Applications parsing invalid CMS structures can crash with a NULL pointer
3290 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3291 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3292 structure callback if an attempt is made to free certain invalid encodings.
3293 Only CHOICE structures using a callback which do not handle NULL value are
3296 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3301 * Montgomery multiplication may produce incorrect results
3303 There is a carry propagating bug in the Broadwell-specific Montgomery
3304 multiplication procedure that handles input lengths divisible by, but
3305 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3306 and DH private keys are impossible. This is because the subroutine in
3307 question is not used in operations with the private key itself and an input
3308 of the attacker's direct choice. Otherwise the bug can manifest itself as
3309 transient authentication and key negotiation failures or reproducible
3310 erroneous outcome of public-key operations with specially crafted input.
3311 Among EC algorithms only Brainpool P-512 curves are affected and one
3312 presumably can attack ECDH key negotiation. Impact was not analyzed in
3313 detail, because pre-requisites for attack are considered unlikely. Namely
3314 multiple clients have to choose the curve in question and the server has to
3315 share the private key among them, neither of which is default behaviour.
3316 Even then only clients that chose the curve will be affected.
3318 This issue was publicly reported as transient failures and was not
3319 initially recognized as a security issue. Thanks to Richard Morgan for
3320 providing reproducible case.
3325 * Removed automatic addition of RPATH in shared libraries and executables,
3326 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3330 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3332 * Fix Use After Free for large message sizes
3334 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3335 message larger than approx 16k is received then the underlying buffer to
3336 store the incoming message is reallocated and moved. Unfortunately a
3337 dangling pointer to the old location is left which results in an attempt to
3338 write to the previously freed location. This is likely to result in a
3339 crash, however it could potentially lead to execution of arbitrary code.
3341 This issue only affects OpenSSL 1.1.0a.
3343 This issue was reported to OpenSSL by Robert Święcki.
3348 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3350 * OCSP Status Request extension unbounded memory growth
3352 A malicious client can send an excessively large OCSP Status Request
3353 extension. If that client continually requests renegotiation, sending a
3354 large OCSP Status Request extension each time, then there will be unbounded
3355 memory growth on the server. This will eventually lead to a Denial Of
3356 Service attack through memory exhaustion. Servers with a default
3357 configuration are vulnerable even if they do not support OCSP. Builds using
3358 the "no-ocsp" build time option are not affected.
3360 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3365 * SSL_peek() hang on empty record
3367 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3368 sends an empty record. This could be exploited by a malicious peer in a
3369 Denial Of Service attack.
3371 This issue was reported to OpenSSL by Alex Gaynor.
3376 * Excessive allocation of memory in tls_get_message_header() and
3377 dtls1_preprocess_fragment()
3379 A (D)TLS message includes 3 bytes for its length in the header for the
3380 message. This would allow for messages up to 16Mb in length. Messages of
3381 this length are excessive and OpenSSL includes a check to ensure that a
3382 peer is sending reasonably sized messages in order to avoid too much memory
3383 being consumed to service a connection. A flaw in the logic of version
3384 1.1.0 means that memory for the message is allocated too early, prior to
3385 the excessive message length check. Due to way memory is allocated in
3386 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3387 to service a connection. This could lead to a Denial of Service through
3388 memory exhaustion. However, the excessive message length check still takes
3389 place, and this would cause the connection to immediately fail. Assuming
3390 that the application calls SSL_free() on the failed connection in a timely
3391 manner then the 21Mb of allocated memory will then be immediately freed
3392 again. Therefore the excessive memory allocation will be transitory in
3393 nature. This then means that there is only a security impact if:
3395 1) The application does not call SSL_free() in a timely manner in the event
3396 that the connection fails
3398 2) The application is working in a constrained environment where there is
3399 very little free memory
3401 3) The attacker initiates multiple connection attempts such that there are
3402 multiple connections in a state where memory has been allocated for the
3403 connection; SSL_free() has not yet been called; and there is insufficient
3404 memory to service the multiple requests.
3406 Except in the instance of (1) above any Denial Of Service is likely to be
3407 transitory because as soon as the connection fails the memory is
3408 subsequently freed again in the SSL_free() call. However there is an
3409 increased risk during this period of application crashes due to the lack of
3410 memory - which would then mean a more serious Denial of Service.
3412 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3413 (CVE-2016-6307 and CVE-2016-6308)
3417 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3418 had to be removed. Primary reason is that vendor assembler can't
3419 assemble our modules with -KPIC flag. As result it, assembly
3420 support, was not even available as option. But its lack means
3421 lack of side-channel resistant code, which is incompatible with
3422 security by todays standards. Fortunately gcc is readily available
3423 prepackaged option, which we firmly point at...
3427 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3429 * Windows command-line tool supports UTF-8 opt-in option for arguments
3430 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3431 (to any value) allows Windows user to access PKCS#12 file generated
3432 with Windows CryptoAPI and protected with non-ASCII password, as well
3433 as files generated under UTF-8 locale on Linux also protected with
3438 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3439 have been disabled by default and removed from DEFAULT, just like RC4.
3440 See the RC4 item below to re-enable both.
3444 * The method for finding the storage location for the Windows RAND seed file
3445 has changed. First we check %RANDFILE%. If that is not set then we check
3446 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3447 all else fails we fall back to C:\.
3451 * The EVP_EncryptUpdate() function has had its return type changed from void
3452 to int. A return of 0 indicates and error while a return of 1 indicates
3457 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3458 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3459 off the constant time implementation for RSA, DSA and DH have been made
3460 no-ops and deprecated.
3464 * Windows RAND implementation was simplified to only get entropy by
3465 calling CryptGenRandom(). Various other RAND-related tickets
3468 *Joseph Wylie Yandle, Rich Salz*
3470 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3471 and `OPENSSL_LH_`, respectively. The old names are available
3472 with API compatibility. They new names are now completely documented.
3476 * Unify TYPE_up_ref(obj) methods signature.
3477 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3478 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3479 int (instead of void) like all others TYPE_up_ref() methods.
3480 So now these methods also check the return value of CRYPTO_atomic_add(),
3481 and the validity of object reference counter.
3483 *fdasilvayy@gmail.com*
3485 * With Windows Visual Studio builds, the .pdb files are installed
3486 alongside the installed libraries and executables. For a static
3487 library installation, ossl_static.pdb is the associate compiler
3488 generated .pdb file to be used when linking programs.
3492 * Remove openssl.spec. Packaging files belong with the packagers.
3496 * Automatic Darwin/OSX configuration has had a refresh, it will now
3497 recognise x86_64 architectures automatically. You can still decide
3498 to build for a different bitness with the environment variable
3499 KERNEL_BITS (can be 32 or 64), for example:
3501 KERNEL_BITS=32 ./config
3505 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3506 256 bit AES and HMAC with SHA256.
3510 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3514 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3518 * To enable users to have their own config files and build file templates,
3519 Configure looks in the directory indicated by the environment variable
3520 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3521 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3522 name and is used as is.
3526 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3527 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3528 X509_CERT_FILE_CTX was removed.
3532 * "shared" builds are now the default. To create only static libraries use
3533 the "no-shared" Configure option.
3537 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3538 All of these option have not worked for some while and are fundamental
3543 * Make various cleanup routines no-ops and mark them as deprecated. Most
3544 global cleanup functions are no longer required because they are handled
3545 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3546 Explicitly de-initing can cause problems (e.g. where a library that uses
3547 OpenSSL de-inits, but an application is still using it). The affected
3548 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3549 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3550 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3551 COMP_zlib_cleanup().
3555 * --strict-warnings no longer enables runtime debugging options
3556 such as REF_DEBUG. Instead, debug options are automatically
3557 enabled with '--debug' builds.
3559 *Andy Polyakov, Emilia Käsper*
3561 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3562 have been moved out of the public header files. New functions for managing
3563 these have been added.
3567 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3568 objects have been moved out of the public header files. New
3569 functions for managing these have been added.
3573 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3574 have been moved out of the public header files. New functions for managing
3575 these have been added.
3579 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3580 moved out of the public header files. New functions for managing these
3585 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3589 * Removed the mk1mf build scripts.
3593 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3594 it is always safe to #include a header now.
3598 * Removed the aged BC-32 config and all its supporting scripts
3602 * Removed support for Ultrix, Netware, and OS/2.
3606 * Add support for HKDF.
3608 *Alessandro Ghedini*
3610 * Add support for blake2b and blake2s
3614 * Added support for "pipelining". Ciphers that have the
3615 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3616 encryptions/decryptions simultaneously. There are currently no built-in
3617 ciphers with this property but the expectation is that engines will be able
3618 to offer it to significantly improve throughput. Support has been extended
3619 into libssl so that multiple records for a single connection can be
3620 processed in one go (for >=TLS 1.1).
3624 * Added the AFALG engine. This is an async capable engine which is able to
3625 offload work to the Linux kernel. In this initial version it only supports
3626 AES128-CBC. The kernel must be version 4.1.0 or greater.
3630 * OpenSSL now uses a new threading API. It is no longer necessary to
3631 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3632 are two supported threading models: pthreads and windows threads. It is
3633 also possible to configure OpenSSL at compile time for "no-threads". The
3634 old threading API should no longer be used. The functions have been
3635 replaced with "no-op" compatibility macros.
3637 *Alessandro Ghedini, Matt Caswell*
3639 * Modify behavior of ALPN to invoke callback after SNI/servername
3640 callback, such that updates to the SSL_CTX affect ALPN.
3644 * Add SSL_CIPHER queries for authentication and key-exchange.
3648 * Changes to the DEFAULT cipherlist:
3649 - Prefer (EC)DHE handshakes over plain RSA.
3650 - Prefer AEAD ciphers over legacy ciphers.
3651 - Prefer ECDSA over RSA when both certificates are available.
3652 - Prefer TLSv1.2 ciphers/PRF.
3653 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3658 * Change the ECC default curve list to be this, in order: x25519,
3659 secp256r1, secp521r1, secp384r1.
3663 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3664 disabled by default. They can be re-enabled using the
3665 enable-weak-ssl-ciphers option to Configure.
3669 * If the server has ALPN configured, but supports no protocols that the
3670 client advertises, send a fatal "no_application_protocol" alert.
3671 This behaviour is SHALL in RFC 7301, though it isn't universally
3672 implemented by other servers.
3676 * Add X25519 support.
3677 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3678 for public and private key encoding using the format documented in
3679 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3680 key generation and key derivation.
3682 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3687 * Deprecate SRP_VBASE_get_by_user.
3688 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3689 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3690 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3691 seed, even if the seed is configured.
3693 Users should use SRP_VBASE_get1_by_user instead. Note that in
3694 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3695 also that even though configuring the SRP seed attempts to hide
3696 invalid usernames by continuing the handshake with fake
3697 credentials, this behaviour is not constant time and no strong
3698 guarantees are made that the handshake is indistinguishable from
3699 that of a valid user.
3703 * Configuration change; it's now possible to build dynamic engines
3704 without having to build shared libraries and vice versa. This
3705 only applies to the engines in `engines/`, those in `crypto/engine/`
3706 will always be built into libcrypto (i.e. "static").
3708 Building dynamic engines is enabled by default; to disable, use
3709 the configuration option "disable-dynamic-engine".
3711 The only requirements for building dynamic engines are the
3712 presence of the DSO module and building with position independent
3713 code, so they will also automatically be disabled if configuring
3714 with "disable-dso" or "disable-pic".
3716 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3717 are also taken away from openssl/opensslconf.h, as they are
3722 * Configuration change; if there is a known flag to compile
3723 position independent code, it will always be applied on the
3724 libcrypto and libssl object files, and never on the application
3725 object files. This means other libraries that use routines from
3726 libcrypto / libssl can be made into shared libraries regardless
3727 of how OpenSSL was configured.
3729 If this isn't desirable, the configuration options "disable-pic"
3730 or "no-pic" can be used to disable the use of PIC. This will
3731 also disable building shared libraries and dynamic engines.
3735 * Removed JPAKE code. It was experimental and has no wide use.
3739 * The INSTALL_PREFIX Makefile variable has been renamed to
3740 DESTDIR. That makes for less confusion on what this variable
3741 is for. Also, the configuration option --install_prefix is
3746 * Heartbeat for TLS has been removed and is disabled by default
3747 for DTLS; configure with enable-heartbeats. Code that uses the
3748 old #define's might need to be updated.
3750 *Emilia Käsper, Rich Salz*
3752 * Rename REF_CHECK to REF_DEBUG.
3756 * New "unified" build system
3758 The "unified" build system is aimed to be a common system for all
3759 platforms we support. With it comes new support for VMS.
3761 This system builds supports building in a different directory tree
3762 than the source tree. It produces one Makefile (for unix family
3763 or lookalikes), or one descrip.mms (for VMS).
3765 The source of information to make the Makefile / descrip.mms is
3766 small files called 'build.info', holding the necessary
3767 information for each directory with source to compile, and a
3768 template in Configurations, like unix-Makefile.tmpl or
3771 With this change, the library names were also renamed on Windows
3772 and on VMS. They now have names that are closer to the standard
3773 on Unix, and include the major version number, and in certain
3774 cases, the architecture they are built for. See "Notes on shared
3775 libraries" in INSTALL.
3777 We rely heavily on the perl module Text::Template.
3781 * Added support for auto-initialisation and de-initialisation of the library.
3782 OpenSSL no longer requires explicit init or deinit routines to be called,
3783 except in certain circumstances. See the OPENSSL_init_crypto() and
3784 OPENSSL_init_ssl() man pages for further information.
3788 * The arguments to the DTLSv1_listen function have changed. Specifically the
3789 "peer" argument is now expected to be a BIO_ADDR object.
3791 * Rewrite of BIO networking library. The BIO library lacked consistent
3792 support of IPv6, and adding it required some more extensive
3793 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3794 which hold all types of addresses and chains of address information.
3795 It also introduces a new API, with functions like BIO_socket,
3796 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3797 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3798 have been adapted accordingly.
3802 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3807 * CRIME protection: disable compression by default, even if OpenSSL is
3808 compiled with zlib enabled. Applications can still enable compression
3809 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3810 using the SSL_CONF library to configure compression.
3814 * The signature of the session callback configured with
3815 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3816 was explicitly marked as `const unsigned char*` instead of
3821 * Always DPURIFY. Remove the use of uninitialized memory in the
3822 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3826 * Removed many obsolete configuration items, including
3827 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3828 MD2_CHAR, MD2_INT, MD2_LONG
3830 IDEA_SHORT, IDEA_LONG
3831 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3833 *Rich Salz, with advice from Andy Polyakov*
3835 * Many BN internals have been moved to an internal header file.
3837 *Rich Salz with help from Andy Polyakov*
3839 * Configuration and writing out the results from it has changed.
3840 Files such as Makefile include/openssl/opensslconf.h and are now
3841 produced through general templates, such as Makefile.in and
3842 crypto/opensslconf.h.in and some help from the perl module
3845 Also, the center of configuration information is no longer
3846 Makefile. Instead, Configure produces a perl module in
3847 configdata.pm which holds most of the config data (in the hash
3848 table %config), the target data that comes from the target
3849 configuration in one of the `Configurations/*.conf` files (in
3854 * To clarify their intended purposes, the Configure options
3855 --prefix and --openssldir change their semantics, and become more
3856 straightforward and less interdependent.
3858 --prefix shall be used exclusively to give the location INSTALLTOP
3859 where programs, scripts, libraries, include files and manuals are
3860 going to be installed. The default is now /usr/local.
3862 --openssldir shall be used exclusively to give the default
3863 location OPENSSLDIR where certificates, private keys, CRLs are
3864 managed. This is also where the default openssl.cnf gets
3866 If the directory given with this option is a relative path, the
3867 values of both the --prefix value and the --openssldir value will
3868 be combined to become OPENSSLDIR.
3869 The default for --openssldir is INSTALLTOP/ssl.
3871 Anyone who uses --openssldir to specify where OpenSSL is to be
3872 installed MUST change to use --prefix instead.
3876 * The GOST engine was out of date and therefore it has been removed. An up
3877 to date GOST engine is now being maintained in an external repository.
3878 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3879 support for GOST ciphersuites (these are only activated if a GOST engine
3884 * EGD is no longer supported by default; use enable-egd when
3887 *Ben Kaduk and Rich Salz*
3889 * The distribution now has Makefile.in files, which are used to
3890 create Makefile's when Configure is run. *Configure must be run
3891 before trying to build now.*
3895 * The return value for SSL_CIPHER_description() for error conditions
3900 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3902 Obtaining and performing DNSSEC validation of TLSA records is
3903 the application's responsibility. The application provides
3904 the TLSA records of its choice to OpenSSL, and these are then
3905 used to authenticate the peer.
3907 The TLSA records need not even come from DNS. They can, for
3908 example, be used to implement local end-entity certificate or
3909 trust-anchor "pinning", where the "pin" data takes the form
3910 of TLSA records, which can augment or replace verification
3911 based on the usual WebPKI public certification authorities.
3915 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3916 continues to support deprecated interfaces in default builds.
3917 However, applications are strongly advised to compile their
3918 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3919 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3920 or the 1.1.0 releases.
3922 In environments in which all applications have been ported to
3923 not use any deprecated interfaces OpenSSL's Configure script
3924 should be used with the --api=1.1.0 option to entirely remove
3925 support for the deprecated features from the library and
3926 unconditionally disable them in the installed headers.
3927 Essentially the same effect can be achieved with the "no-deprecated"
3928 argument to Configure, except that this will always restrict
3929 the build to just the latest API, rather than a fixed API
3932 As applications are ported to future revisions of the API,
3933 they should update their compile-time OPENSSL_API_COMPAT define
3934 accordingly, but in most cases should be able to continue to
3935 compile with later releases.
3937 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3938 0x10000000L and 0x00908000L, respectively. However those
3939 versions did not support the OPENSSL_API_COMPAT feature, and
3940 so applications are not typically tested for explicit support
3941 of just the undeprecated features of either release.
3945 * Add support for setting the minimum and maximum supported protocol.
3946 It can bet set via the SSL_set_min_proto_version() and
3947 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3948 MaxProtocol. It's recommended to use the new APIs to disable
3949 protocols instead of disabling individual protocols using
3950 SSL_set_options() or SSL_CONF's Protocol. This change also
3951 removes support for disabling TLS 1.2 in the OpenSSL TLS
3952 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3956 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3960 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3961 and integrates ECDSA and ECDH functionality into EC. Implementations can
3962 now redirect key generation and no longer need to convert to or from
3965 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3966 include the ec.h header file instead.
3970 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3971 ciphers who are no longer supported and drops support the ephemeral RSA key
3972 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3976 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3977 opaque. For HMAC_CTX, the following constructors and destructors
3980 HMAC_CTX *HMAC_CTX_new(void);
3981 void HMAC_CTX_free(HMAC_CTX *ctx);
3983 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3984 destroy such methods has been added. See EVP_MD_meth_new(3) and
3985 EVP_CIPHER_meth_new(3) for documentation.
3988 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3989 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3990 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3991 an already created structure.
3992 2) For consistency with the majority of our object creators and
3993 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3994 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3995 for deprecated builds.
3999 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4000 cryptographic operations to be performed asynchronously as long as an
4001 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4002 further details. Libssl has also had this capability integrated with the
4003 introduction of the new mode SSL_MODE_ASYNC and associated error
4004 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4005 pages. This work was developed in partnership with Intel Corp.
4009 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4010 always enabled now. If you want to disable the support you should
4011 exclude it using the list of supported ciphers. This also means that the
4012 "-no_ecdhe" option has been removed from s_server.
4016 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4017 SSL_{CTX_}set1_curves() which can set a list.
4021 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4022 curve you want to support using SSL_{CTX_}set1_curves().
4026 * State machine rewrite. The state machine code has been significantly
4027 refactored in order to remove much duplication of code and solve issues
4028 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4029 further details). This change does have some associated API changes.
4030 Notably the SSL_state() function has been removed and replaced by
4031 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4032 SSL_set_state() has been removed altogether. The previous handshake states
4033 defined in ssl.h and ssl3.h have also been removed.
4037 * All instances of the string "ssleay" in the public API were replaced
4038 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4039 Some error codes related to internal RSA_eay API's were renamed.
4043 * The demo files in crypto/threads were moved to demo/threads.
4047 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4050 *Matt Caswell, Rich Salz*
4052 * New ASN.1 embed macro.
4054 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4055 structure is not allocated: it is part of the parent. That is instead of
4063 This reduces memory fragmentation and make it impossible to accidentally
4064 set a mandatory field to NULL.
4066 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4067 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4068 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4073 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4077 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4078 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4079 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4080 DES and RC4 ciphersuites.
4084 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4085 This changes the decoding behaviour for some invalid messages,
4086 though the change is mostly in the more lenient direction, and
4087 legacy behaviour is preserved as much as possible.
4091 * Fix no-stdio build.
4092 *David Woodhouse <David.Woodhouse@intel.com> and also*
4093 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4095 * New testing framework
4096 The testing framework has been largely rewritten and is now using
4097 perl and the perl modules Test::Harness and an extended variant of
4098 Test::More called OpenSSL::Test to do its work. All test scripts in
4099 test/ have been rewritten into test recipes, and all direct calls to
4100 executables in test/Makefile have become individual recipes using the
4101 simplified testing OpenSSL::Test::Simple.
4103 For documentation on our testing modules, do:
4105 perldoc test/testlib/OpenSSL/Test/Simple.pm
4106 perldoc test/testlib/OpenSSL/Test.pm
4110 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4111 are used; the latter aborts on memory leaks (usually checked on exit).
4112 Some undocumented "set malloc, etc., hooks" functions were removed
4113 and others were changed. All are now documented.
4117 * In DSA_generate_parameters_ex, if the provided seed is too short,
4120 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4122 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4123 from RFC4279, RFC4785, RFC5487, RFC5489.
4125 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4126 original RSA_PSK patch.
4130 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4131 era flag was never set throughout the codebase (only read). Also removed
4132 SSL3_FLAGS_POP_BUFFER which was only used if
4133 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4137 * Changed the default name options in the "ca", "crl", "req" and "x509"
4138 to be "oneline" instead of "compat".
4142 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4143 not aware of clients that still exhibit this bug, and the workaround
4144 hasn't been working properly for a while.
4148 * The return type of BIO_number_read() and BIO_number_written() as well as
4149 the corresponding num_read and num_write members in the BIO structure has
4150 changed from unsigned long to uint64_t. On platforms where an unsigned
4151 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4156 * Given the pervasive nature of TLS extensions it is inadvisable to run
4157 OpenSSL without support for them. It also means that maintaining
4158 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4159 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4163 * Removed support for the two export grade static DH ciphersuites
4164 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4165 were newly added (along with a number of other static DH ciphersuites) to
4166 1.0.2. However the two export ones have *never* worked since they were
4167 introduced. It seems strange in any case to be adding new export
4168 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4172 * Version negotiation has been rewritten. In particular SSLv23_method(),
4173 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4174 and turned into macros which simply call the new preferred function names
4175 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4176 should use the new names instead. Also as part of this change the ssl23.h
4177 header file has been removed.
4181 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4182 code and the associated standard is no longer considered fit-for-purpose.
4186 * RT2547 was closed. When generating a private key, try to make the
4187 output file readable only by the owner. This behavior change might
4188 be noticeable when interacting with other software.
4190 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4195 * Added HTTP GET support to the ocsp command.
4199 * Changed default digest for the dgst and enc commands from MD5 to
4204 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4208 * Added support for TLS extended master secret from
4209 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4210 initial patch which was a great help during development.
4214 * All libssl internal structures have been removed from the public header
4215 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4216 now redundant). Users should not attempt to access internal structures
4217 directly. Instead they should use the provided API functions.
4221 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4222 Access to deprecated functions can be re-enabled by running config with
4223 "enable-deprecated". In addition applications wishing to use deprecated
4224 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4225 will, by default, disable some transitive includes that previously existed
4226 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4230 * Added support for OCB mode. OpenSSL has been granted a patent license
4231 compatible with the OpenSSL license for use of OCB. Details are available
4232 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4233 for OCB can be removed by calling config with no-ocb.
4237 * SSLv2 support has been removed. It still supports receiving a SSLv2
4238 compatible client hello.
4242 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4243 done while fixing the error code for the key-too-small case.
4245 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4247 * CA.sh has been removed; use CA.pl instead.
4251 * Removed old DES API.
4255 * Remove various unsupported platforms:
4261 Sinix/ReliantUNIX RM400
4266 16-bit platforms such as WIN16
4270 * Clean up OPENSSL_NO_xxx #define's
4271 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4272 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4273 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4274 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4275 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4276 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4277 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4278 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4279 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4280 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4284 * Cleaned up dead code
4285 Remove all but one '#ifdef undef' which is to be looked at.
4289 * Clean up calling of xxx_free routines.
4290 Just like free(), fix most of the xxx_free routines to accept
4291 NULL. Remove the non-null checks from callers. Save much code.
4295 * Add secure heap for storage of private keys (when possible).
4296 Add BIO_s_secmem(), CBIGNUM, etc.
4297 Contributed by Akamai Technologies under our Corporate CLA.
4301 * Experimental support for a new, fast, unbiased prime candidate generator,
4302 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4304 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4306 * New output format NSS in the sess_id command line tool. This allows
4307 exporting the session id and the master key in NSS keylog format.
4309 *Martin Kaiser <martin@kaiser.cx>*
4311 * Harmonize version and its documentation. -f flag is used to display
4314 *mancha <mancha1@zoho.com>*
4316 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4317 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4319 *mancha <mancha1@zoho.com>*
4321 * Fix some double frees. These are not thought to be exploitable.
4323 *mancha <mancha1@zoho.com>*
4325 * A missing bounds check in the handling of the TLS heartbeat extension
4326 can be used to reveal up to 64k of memory to a connected client or
4329 Thanks for Neel Mehta of Google Security for discovering this bug and to
4330 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4331 preparing the fix ([CVE-2014-0160])
4333 *Adam Langley, Bodo Moeller*
4335 * Fix for the attack described in the paper "Recovering OpenSSL
4336 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4337 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4338 <http://eprint.iacr.org/2014/140>
4340 Thanks to Yuval Yarom and Naomi Benger for discovering this
4341 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4343 *Yuval Yarom and Naomi Benger*
4345 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4346 this fixes a limitation in previous versions of OpenSSL.
4350 * Experimental encrypt-then-mac support.
4352 Experimental support for encrypt then mac from
4353 draft-gutmann-tls-encrypt-then-mac-02.txt
4355 To enable it set the appropriate extension number (0x42 for the test
4356 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4358 For non-compliant peers (i.e. just about everything) this should have no
4361 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4365 * Add EVP support for key wrapping algorithms, to avoid problems with
4366 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4367 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4368 algorithms and include tests cases.
4372 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4377 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4378 MGF1 digest and OAEP label.
4382 * Make openssl verify return errors.
4384 *Chris Palmer <palmer@google.com> and Ben Laurie*
4386 * New function ASN1_TIME_diff to calculate the difference between two
4387 ASN1_TIME structures or one structure and the current time.
4391 * Update fips_test_suite to support multiple command line options. New
4392 test to induce all self test errors in sequence and check expected
4397 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4398 sign or verify all in one operation.
4402 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4403 test programs and fips_test_suite. Includes functionality to parse
4404 the minimal script output of fipsalgest.pl directly.
4408 * Add authorisation parameter to FIPS_module_mode_set().
4412 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4416 * Use separate DRBG fields for internal and external flags. New function
4417 FIPS_drbg_health_check() to perform on demand health checking. Add
4418 generation tests to fips_test_suite with reduced health check interval to
4419 demonstrate periodic health checking. Add "nodh" option to
4420 fips_test_suite to skip very slow DH test.
4424 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4429 * More extensive health check for DRBG checking many more failure modes.
4430 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4431 combination: call this in fips_test_suite.
4435 * Add support for canonical generation of DSA parameter 'g'. See
4438 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4439 POST to handle HMAC cases.
4443 * Add functions FIPS_module_version() and FIPS_module_version_text()
4444 to return numerical and string versions of the FIPS module number.
4448 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4449 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4450 outside the validated module in the FIPS capable OpenSSL.
4454 * Minor change to DRBG entropy callback semantics. In some cases
4455 there is no multiple of the block length between min_len and
4456 max_len. Allow the callback to return more than max_len bytes
4457 of entropy but discard any extra: it is the callback's responsibility
4458 to ensure that the extra data discarded does not impact the
4459 requested amount of entropy.
4463 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4464 information in FIPS186-3, SP800-57 and SP800-131A.
4468 * CCM support via EVP. Interface is very similar to GCM case except we
4469 must supply all data in one chunk (i.e. no update, final) and the
4470 message length must be supplied if AAD is used. Add algorithm test
4475 * Initial version of POST overhaul. Add POST callback to allow the status
4476 of POST to be monitored and/or failures induced. Modify fips_test_suite
4477 to use callback. Always run all selftests even if one fails.
4481 * XTS support including algorithm test driver in the fips_gcmtest program.
4482 Note: this does increase the maximum key length from 32 to 64 bytes but
4483 there should be no binary compatibility issues as existing applications
4484 will never use XTS mode.
4488 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4489 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4490 performs algorithm blocking for unapproved PRNG types. Also do not
4491 set PRNG type in FIPS_mode_set(): leave this to the application.
4492 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4493 the standard OpenSSL PRNG: set additional data to a date time vector.
4497 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4498 This shouldn't present any incompatibility problems because applications
4499 shouldn't be using these directly and any that are will need to rethink
4500 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4504 * Extensive self tests and health checking required by SP800-90 DRBG.
4505 Remove strength parameter from FIPS_drbg_instantiate and always
4506 instantiate at maximum supported strength.
4510 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4514 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4518 * New function DH_compute_key_padded() to compute a DH key and pad with
4519 leading zeroes if needed: this complies with SP800-56A et al.
4523 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4524 anything, incomplete, subject to change and largely untested at present.
4528 * Modify fipscanisteronly build option to only build the necessary object
4529 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4533 * Add experimental option FIPSSYMS to give all symbols in
4534 fipscanister.o and FIPS or fips prefix. This will avoid
4535 conflicts with future versions of OpenSSL. Add perl script
4536 util/fipsas.pl to preprocess assembly language source files
4537 and rename any affected symbols.
4541 * Add selftest checks and algorithm block of non-fips algorithms in
4542 FIPS mode. Remove DES2 from selftests.
4546 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4547 return internal method without any ENGINE dependencies. Add new
4548 tiny fips sign and verify functions.
4552 * New build option no-ec2m to disable characteristic 2 code.
4556 * New build option "fipscanisteronly". This only builds fipscanister.o
4557 and (currently) associated fips utilities. Uses the file Makefile.fips
4558 instead of Makefile.org as the prototype.
4562 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4563 Update fips_gcmtest to use IV generator.
4567 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4568 setting output buffer to NULL. The `*Final` function must be
4569 called although it will not retrieve any additional data. The tag
4570 can be set or retrieved with a ctrl. The IV length is by default 12
4571 bytes (96 bits) but can be set to an alternative value. If the IV
4572 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4577 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4578 underlying do_cipher function handles all cipher semantics itself
4579 including padding and finalisation. This is useful if (for example)
4580 an ENGINE cipher handles block padding itself. The behaviour of
4581 do_cipher is subtly changed if this flag is set: the return value
4582 is the number of characters written to the output buffer (zero is
4583 no longer an error code) or a negative error code. Also if the
4584 input buffer is NULL and length 0 finalisation should be performed.
4588 * If a candidate issuer certificate is already part of the constructed
4589 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4593 * Improve forward-security support: add functions
4595 void SSL_CTX_set_not_resumable_session_callback(
4596 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4597 void SSL_set_not_resumable_session_callback(
4598 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4600 for use by SSL/TLS servers; the callback function will be called whenever a
4601 new session is created, and gets to decide whether the session may be
4602 cached to make it resumable (return 0) or not (return 1). (As by the
4603 SSL/TLS protocol specifications, the session_id sent by the server will be
4604 empty to indicate that the session is not resumable; also, the server will
4605 not generate RFC 4507 (RFC 5077) session tickets.)
4607 A simple reasonable callback implementation is to return is_forward_secure.
4608 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4609 by the SSL/TLS server library, indicating whether it can provide forward
4612 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4614 * New -verify_name option in command line utilities to set verification
4619 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4620 Add CMAC pkey methods.
4624 * Experimental renegotiation in s_server -www mode. If the client
4625 browses /reneg connection is renegotiated. If /renegcert it is
4626 renegotiated requesting a certificate.
4630 * Add an "external" session cache for debugging purposes to s_server. This
4631 should help trace issues which normally are only apparent in deployed
4632 multi-process servers.
4636 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4637 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4638 BIO_set_cipher() and some obscure PEM functions were changed so they
4639 can now return an error. The RAND changes required a change to the
4640 RAND_METHOD structure.
4644 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4645 a gcc attribute to warn if the result of a function is ignored. This
4646 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4647 whose return value is often ignored.
4651 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4652 These allow SCTs (signed certificate timestamps) to be requested and
4653 validated when establishing a connection.
4655 *Rob Percival <robpercival@google.com>*
4660 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4662 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4663 used even when parsing explicit parameters, when loading a encoded key
4664 or calling `EC_GROUP_new_from_ecpkparameters()`/
4665 `EC_GROUP_new_from_ecparameters()`.
4666 This prevents bypass of security hardening and performance gains,
4667 especially for curves with specialized EC_METHODs.
4668 By default, if a key encoded with explicit parameters is loaded and later
4669 encoded, the output is still encoded with explicit parameters, even if
4670 internally a "named" EC_GROUP is used for computation.
4674 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4675 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4676 NULL. After this change, only the cofactor parameter can be NULL. It also
4677 does some minimal sanity checks on the passed order.
4682 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4683 An attack is simple, if the first CMS_recipientInfo is valid but the
4684 second CMS_recipientInfo is chosen ciphertext. If the second
4685 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4686 encryption key will be replaced by garbage, and the message cannot be
4687 decoded, but if the RSA decryption fails, the correct encryption key is
4688 used and the recipient will not notice the attack.
4689 As a work around for this potential attack the length of the decrypted
4690 key must be equal to the cipher default key length, in case the
4691 certifiate is not given and all recipientInfo are tried out.
4692 The old behaviour can be re-enabled in the CMS code by setting the
4693 CMS_DEBUG_DECRYPT flag.
4698 * Document issue with installation paths in diverse Windows builds
4700 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4701 binaries and run-time config file.
4706 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4708 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4709 This changes the size when using the `genpkey` command when no size is given.
4710 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4711 generation commands to use 2048 bits by default.
4715 * Add FIPS support for Android Arm 64-bit
4717 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4718 Module in Version 2.0.10. For some reason, the corresponding target
4719 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4720 built with FIPS support on Android Arm 64-bit. This omission has been
4723 *Matthias St. Pierre*
4725 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4727 * 0-byte record padding oracle
4729 If an application encounters a fatal protocol error and then calls
4730 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4731 then OpenSSL can respond differently to the calling application if a 0 byte
4732 record is received with invalid padding compared to if a 0 byte record is
4733 received with an invalid MAC. If the application then behaves differently
4734 based on that in a way that is detectable to the remote peer, then this
4735 amounts to a padding oracle that could be used to decrypt data.
4737 In order for this to be exploitable "non-stitched" ciphersuites must be in
4738 use. Stitched ciphersuites are optimised implementations of certain
4739 commonly used ciphersuites. Also the application must call SSL_shutdown()
4740 twice even if a protocol error has occurred (applications should not do
4741 this but some do anyway).
4743 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4744 Aviram, with additional investigation by Steven Collison and Andrew
4745 Hourselt. It was reported to OpenSSL on 10th December 2018.
4750 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4754 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4756 * Microarchitecture timing vulnerability in ECC scalar multiplication
4758 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4759 shown to be vulnerable to a microarchitecture timing side channel attack.
4760 An attacker with sufficient access to mount local timing attacks during
4761 ECDSA signature generation could recover the private key.
4763 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4764 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4770 * Timing vulnerability in DSA signature generation
4772 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4773 timing side channel attack. An attacker could use variations in the signing
4774 algorithm to recover the private key.
4776 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4781 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4782 Module, accidentally introduced while backporting security fixes from the
4783 development branch and hindering the use of ECC in FIPS mode.
4787 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4789 * Client DoS due to large DH parameter
4791 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4792 malicious server can send a very large prime value to the client. This will
4793 cause the client to spend an unreasonably long period of time generating a
4794 key for this prime resulting in a hang until the client has finished. This
4795 could be exploited in a Denial Of Service attack.
4797 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4802 * Cache timing vulnerability in RSA Key Generation
4804 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4805 a cache timing side channel attack. An attacker with sufficient access to
4806 mount cache timing attacks during the RSA key generation process could
4807 recover the private key.
4809 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4810 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4815 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4816 parameter is no longer accepted, as it leads to a corrupt table. NULL
4817 pem_str is reserved for alias entries only.
4821 * Revert blinding in ECDSA sign and instead make problematic addition
4822 length-invariant. Switch even to fixed-length Montgomery multiplication.
4826 * Change generating and checking of primes so that the error rate of not
4827 being prime depends on the intended use based on the size of the input.
4828 For larger primes this will result in more rounds of Miller-Rabin.
4829 The maximal error rate for primes with more than 1080 bits is lowered
4832 *Kurt Roeckx, Annie Yousar*
4834 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4838 * Add blinding to ECDSA and DSA signatures to protect against side channel
4839 attacks discovered by Keegan Ryan (NCC Group).
4843 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4844 now allow empty (zero character) pass phrases.
4848 * Certificate time validation (X509_cmp_time) enforces stricter
4849 compliance with RFC 5280. Fractional seconds and timezone offsets
4850 are no longer allowed.
4854 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4856 * Constructed ASN.1 types with a recursive definition could exceed the stack
4858 Constructed ASN.1 types with a recursive definition (such as can be found
4859 in PKCS7) could eventually exceed the stack given malicious input with
4860 excessive recursion. This could result in a Denial Of Service attack. There
4861 are no such structures used within SSL/TLS that come from untrusted sources
4862 so this is considered safe.
4864 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4870 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4872 * Read/write after SSL object in error state
4874 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4875 mechanism. The intent was that if a fatal error occurred during a handshake
4876 then OpenSSL would move into the error state and would immediately fail if
4877 you attempted to continue the handshake. This works as designed for the
4878 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4879 SSL_connect()), however due to a bug it does not work correctly if
4880 SSL_read() or SSL_write() is called directly. In that scenario, if the
4881 handshake fails then a fatal error will be returned in the initial function
4882 call. If SSL_read()/SSL_write() is subsequently called by the application
4883 for the same SSL object then it will succeed and the data is passed without
4884 being decrypted/encrypted directly from the SSL/TLS record layer.
4886 In order to exploit this issue an application bug would have to be present
4887 that resulted in a call to SSL_read()/SSL_write() being issued after having
4888 already received a fatal error.
4890 This issue was reported to OpenSSL by David Benjamin (Google).
4895 * rsaz_1024_mul_avx2 overflow bug on x86_64
4897 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4898 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4899 Analysis suggests that attacks against RSA and DSA as a result of this
4900 defect would be very difficult to perform and are not believed likely.
4901 Attacks against DH1024 are considered just feasible, because most of the
4902 work necessary to deduce information about a private key may be performed
4903 offline. The amount of resources required for such an attack would be
4904 significant. However, for an attack on TLS to be meaningful, the server
4905 would have to share the DH1024 private key among multiple clients, which is
4906 no longer an option since CVE-2016-0701.
4908 This only affects processors that support the AVX2 but not ADX extensions
4909 like Intel Haswell (4th generation).
4911 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4912 was originally found via the OSS-Fuzz project.
4917 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4919 * bn_sqrx8x_internal carry bug on x86_64
4921 There is a carry propagating bug in the x86_64 Montgomery squaring
4922 procedure. No EC algorithms are affected. Analysis suggests that attacks
4923 against RSA and DSA as a result of this defect would be very difficult to
4924 perform and are not believed likely. Attacks against DH are considered just
4925 feasible (although very difficult) because most of the work necessary to
4926 deduce information about a private key may be performed offline. The amount
4927 of resources required for such an attack would be very significant and
4928 likely only accessible to a limited number of attackers. An attacker would
4929 additionally need online access to an unpatched system using the target
4930 private key in a scenario with persistent DH parameters and a private
4931 key that is shared between multiple clients.
4933 This only affects processors that support the BMI1, BMI2 and ADX extensions
4934 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4936 This issue was reported to OpenSSL by the OSS-Fuzz project.
4941 * Malformed X.509 IPAddressFamily could cause OOB read
4943 If an X.509 certificate has a malformed IPAddressFamily extension,
4944 OpenSSL could do a one-byte buffer overread. The most likely result
4945 would be an erroneous display of the certificate in text format.
4947 This issue was reported to OpenSSL by the OSS-Fuzz project.
4952 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4954 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4955 platform rather than 'mingw'.
4959 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4961 * Truncated packet could crash via OOB read
4963 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4964 cipher is being used, then a truncated packet can cause that host to
4965 perform an out-of-bounds read, usually resulting in a crash.
4967 This issue was reported to OpenSSL by Robert Święcki of Google.
4972 * BN_mod_exp may produce incorrect results on x86_64
4974 There is a carry propagating bug in the x86_64 Montgomery squaring
4975 procedure. No EC algorithms are affected. Analysis suggests that attacks
4976 against RSA and DSA as a result of this defect would be very difficult to
4977 perform and are not believed likely. Attacks against DH are considered just
4978 feasible (although very difficult) because most of the work necessary to
4979 deduce information about a private key may be performed offline. The amount
4980 of resources required for such an attack would be very significant and
4981 likely only accessible to a limited number of attackers. An attacker would
4982 additionally need online access to an unpatched system using the target
4983 private key in a scenario with persistent DH parameters and a private
4984 key that is shared between multiple clients. For example this can occur by
4985 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4986 similar to CVE-2015-3193 but must be treated as a separate problem.
4988 This issue was reported to OpenSSL by the OSS-Fuzz project.
4993 * Montgomery multiplication may produce incorrect results
4995 There is a carry propagating bug in the Broadwell-specific Montgomery
4996 multiplication procedure that handles input lengths divisible by, but
4997 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4998 and DH private keys are impossible. This is because the subroutine in
4999 question is not used in operations with the private key itself and an input
5000 of the attacker's direct choice. Otherwise the bug can manifest itself as
5001 transient authentication and key negotiation failures or reproducible
5002 erroneous outcome of public-key operations with specially crafted input.
5003 Among EC algorithms only Brainpool P-512 curves are affected and one
5004 presumably can attack ECDH key negotiation. Impact was not analyzed in
5005 detail, because pre-requisites for attack are considered unlikely. Namely
5006 multiple clients have to choose the curve in question and the server has to
5007 share the private key among them, neither of which is default behaviour.
5008 Even then only clients that chose the curve will be affected.
5010 This issue was publicly reported as transient failures and was not
5011 initially recognized as a security issue. Thanks to Richard Morgan for
5012 providing reproducible case.
5017 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5018 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5019 prevent issues where no progress is being made and the peer continually
5020 sends unrecognised record types, using up resources processing them.
5024 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5026 * Missing CRL sanity check
5028 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5029 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5030 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5032 This issue only affects the OpenSSL 1.0.2i
5037 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5039 * OCSP Status Request extension unbounded memory growth
5041 A malicious client can send an excessively large OCSP Status Request
5042 extension. If that client continually requests renegotiation, sending a
5043 large OCSP Status Request extension each time, then there will be unbounded
5044 memory growth on the server. This will eventually lead to a Denial Of
5045 Service attack through memory exhaustion. Servers with a default
5046 configuration are vulnerable even if they do not support OCSP. Builds using
5047 the "no-ocsp" build time option are not affected.
5049 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5054 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5057 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5063 * OOB write in MDC2_Update()
5065 An overflow can occur in MDC2_Update() either if called directly or
5066 through the EVP_DigestUpdate() function using MDC2. If an attacker
5067 is able to supply very large amounts of input data after a previous
5068 call to EVP_EncryptUpdate() with a partial block then a length check
5069 can overflow resulting in a heap corruption.
5071 The amount of data needed is comparable to SIZE_MAX which is impractical
5074 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5079 * Malformed SHA512 ticket DoS
5081 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5082 DoS attack where a malformed ticket will result in an OOB read which will
5085 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5086 a custom server callback and ticket lookup mechanism.
5088 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5093 * OOB write in BN_bn2dec()
5095 The function BN_bn2dec() does not check the return value of BN_div_word().
5096 This can cause an OOB write if an application uses this function with an
5097 overly large BIGNUM. This could be a problem if an overly large certificate
5098 or CRL is printed out from an untrusted source. TLS is not affected because
5099 record limits will reject an oversized certificate before it is parsed.
5101 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5106 * OOB read in TS_OBJ_print_bio()
5108 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5109 the total length the OID text representation would use and not the amount
5110 of data written. This will result in OOB reads when large OIDs are
5113 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5118 * Pointer arithmetic undefined behaviour
5120 Avoid some undefined pointer arithmetic
5122 A common idiom in the codebase is to check limits in the following manner:
5125 Where "p" points to some malloc'd data of SIZE bytes and
5128 "len" here could be from some externally supplied data (e.g. from a TLS
5131 The rules of C pointer arithmetic are such that "p + len" is only well
5132 defined where len <= SIZE. Therefore the above idiom is actually
5133 undefined behaviour.
5135 For example this could cause problems if some malloc implementation
5136 provides an address for "p" such that "p + len" actually overflows for
5137 values of len that are too big and therefore p + len < limit.
5139 This issue was reported to OpenSSL by Guido Vranken
5144 * Constant time flag not preserved in DSA signing
5146 Operations in the DSA signing algorithm should run in constant time in
5147 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5148 implementation means that a non-constant time codepath is followed for
5149 certain operations. This has been demonstrated through a cache-timing
5150 attack to be sufficient for an attacker to recover the private DSA key.
5152 This issue was reported by César Pereida (Aalto University), Billy Brumley
5153 (Tampere University of Technology), and Yuval Yarom (The University of
5154 Adelaide and NICTA).
5159 * DTLS buffered message DoS
5161 In a DTLS connection where handshake messages are delivered out-of-order
5162 those messages that OpenSSL is not yet ready to process will be buffered
5163 for later use. Under certain circumstances, a flaw in the logic means that
5164 those messages do not get removed from the buffer even though the handshake
5165 has been completed. An attacker could force up to approx. 15 messages to
5166 remain in the buffer when they are no longer required. These messages will
5167 be cleared when the DTLS connection is closed. The default maximum size for
5168 a message is 100k. Therefore the attacker could force an additional 1500k
5169 to be consumed per connection. By opening many simulataneous connections an
5170 attacker could cause a DoS attack through memory exhaustion.
5172 This issue was reported to OpenSSL by Quan Luo.
5177 * DTLS replay protection DoS
5179 A flaw in the DTLS replay attack protection mechanism means that records
5180 that arrive for future epochs update the replay protection "window" before
5181 the MAC for the record has been validated. This could be exploited by an
5182 attacker by sending a record for the next epoch (which does not have to
5183 decrypt or have a valid MAC), with a very large sequence number. This means
5184 that all subsequent legitimate packets are dropped causing a denial of
5185 service for a specific DTLS connection.
5187 This issue was reported to OpenSSL by the OCAP audit team.
5192 * Certificate message OOB reads
5194 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5195 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5196 theoretical DoS risk but this has not been observed in practice on common
5199 The messages affected are client certificate, client certificate request
5200 and server certificate. As a result the attack can only be performed
5201 against a client or a server which enables client authentication.
5203 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5208 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5210 * Prevent padding oracle in AES-NI CBC MAC check
5212 A MITM attacker can use a padding oracle attack to decrypt traffic
5213 when the connection uses an AES CBC cipher and the server support
5216 This issue was introduced as part of the fix for Lucky 13 padding
5217 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5218 constant time by making sure that always the same bytes are read and
5219 compared against either the MAC or padding bytes. But it no longer
5220 checked that there was enough data to have both the MAC and padding
5223 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5228 * Fix EVP_EncodeUpdate overflow
5230 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5231 Base64 encoding of binary data. If an attacker is able to supply very large
5232 amounts of input data then a length check can overflow resulting in a heap
5235 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5236 the `PEM_write_bio*` family of functions. These are mainly used within the
5237 OpenSSL command line applications, so any application which processes data
5238 from an untrusted source and outputs it as a PEM file should be considered
5239 vulnerable to this issue. User applications that call these APIs directly
5240 with large amounts of untrusted data may also be vulnerable.
5242 This issue was reported by Guido Vranken.
5247 * Fix EVP_EncryptUpdate overflow
5249 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5250 is able to supply very large amounts of input data after a previous call to
5251 EVP_EncryptUpdate() with a partial block then a length check can overflow
5252 resulting in a heap corruption. Following an analysis of all OpenSSL
5253 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5254 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5255 the first called function after an EVP_EncryptInit(), and therefore that
5256 specific call must be safe. The second form is where the length passed to
5257 EVP_EncryptUpdate() can be seen from the code to be some small value and
5258 therefore there is no possibility of an overflow. Since all instances are
5259 one of these two forms, it is believed that there can be no overflows in
5260 internal code due to this problem. It should be noted that
5261 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5262 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5263 of these calls have also been analysed too and it is believed there are no
5264 instances in internal usage where an overflow could occur.
5266 This issue was reported by Guido Vranken.
5271 * Prevent ASN.1 BIO excessive memory allocation
5273 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5274 a short invalid encoding can cause allocation of large amounts of memory
5275 potentially consuming excessive resources or exhausting memory.
5277 Any application parsing untrusted data through d2i BIO functions is
5278 affected. The memory based functions such as d2i_X509() are *not* affected.
5279 Since the memory based functions are used by the TLS library, TLS
5280 applications are not affected.
5282 This issue was reported by Brian Carpenter.
5289 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5290 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5291 in arbitrary stack data being returned in the buffer.
5293 This issue was reported by Guido Vranken.
5298 * Modify behavior of ALPN to invoke callback after SNI/servername
5299 callback, such that updates to the SSL_CTX affect ALPN.
5303 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5308 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5309 methods are enabled and ssl2 is disabled the methods return NULL.
5313 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5315 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5316 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5317 provide any "EXPORT" or "LOW" strength ciphers.
5321 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5322 is by default disabled at build-time. Builds that are not configured with
5323 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5324 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5325 will need to explicitly call either of:
5327 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5329 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5331 as appropriate. Even if either of those is used, or the application
5332 explicitly uses the version-specific SSLv2_method() or its client and
5333 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5334 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5335 ciphers, and SSLv2 56-bit DES are no longer available.
5340 * Fix a double-free in DSA code
5342 A double free bug was discovered when OpenSSL parses malformed DSA private
5343 keys and could lead to a DoS attack or memory corruption for applications
5344 that receive DSA private keys from untrusted sources. This scenario is
5347 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5353 * Disable SRP fake user seed to address a server memory leak.
5355 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5357 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5358 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5359 was changed to ignore the "fake user" SRP seed, even if the seed
5362 Users should use SRP_VBASE_get1_by_user instead. Note that in
5363 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5364 also that even though configuring the SRP seed attempts to hide
5365 invalid usernames by continuing the handshake with fake
5366 credentials, this behaviour is not constant time and no strong
5367 guarantees are made that the handshake is indistinguishable from
5368 that of a valid user.
5373 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5375 In the BN_hex2bn function the number of hex digits is calculated using an
5376 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5377 large values of `i` this can result in `bn_expand` not allocating any
5378 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5379 field as NULL leading to a subsequent NULL ptr deref. For very large values
5380 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5381 In this case memory is allocated to the internal BIGNUM data field, but it
5382 is insufficiently sized leading to heap corruption. A similar issue exists
5383 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5384 is ever called by user applications with very large untrusted hex/dec data.
5385 This is anticipated to be a rare occurrence.
5387 All OpenSSL internal usage of these functions use data that is not expected
5388 to be untrusted, e.g. config file data or application command line
5389 arguments. If user developed applications generate config file data based
5390 on untrusted data then it is possible that this could also lead to security
5391 consequences. This is also anticipated to be rare.
5393 This issue was reported to OpenSSL by Guido Vranken.
5398 * Fix memory issues in `BIO_*printf` functions
5400 The internal `fmtstr` function used in processing a "%s" format string in
5401 the `BIO_*printf` functions could overflow while calculating the length of a
5402 string and cause an OOB read when printing very long strings.
5404 Additionally the internal `doapr_outch` function can attempt to write to an
5405 OOB memory location (at an offset from the NULL pointer) in the event of a
5406 memory allocation failure. In 1.0.2 and below this could be caused where
5407 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5408 could be in processing a very long "%s" format string. Memory leaks can
5411 The first issue may mask the second issue dependent on compiler behaviour.
5412 These problems could enable attacks where large amounts of untrusted data
5413 is passed to the `BIO_*printf` functions. If applications use these functions
5414 in this way then they could be vulnerable. OpenSSL itself uses these
5415 functions when printing out human-readable dumps of ASN.1 data. Therefore
5416 applications that print this data could be vulnerable if the data is from
5417 untrusted sources. OpenSSL command line applications could also be
5418 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5419 as command line arguments.
5421 Libssl is not considered directly vulnerable. Additionally certificates etc
5422 received via remote connections via libssl are also unlikely to be able to
5423 trigger these issues because of message size limits enforced within libssl.
5425 This issue was reported to OpenSSL Guido Vranken.
5430 * Side channel attack on modular exponentiation
5432 A side-channel attack was found which makes use of cache-bank conflicts on
5433 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5434 of RSA keys. The ability to exploit this issue is limited as it relies on
5435 an attacker who has control of code in a thread running on the same
5436 hyper-threaded core as the victim thread which is performing decryptions.
5438 This issue was reported to OpenSSL by Yuval Yarom, The University of
5439 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5440 Nadia Heninger, University of Pennsylvania with more information at
5441 <http://cachebleed.info>.
5446 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5447 if no keysize is specified with default_bits. This fixes an
5448 omission in an earlier change that changed all RSA/DSA key generation
5449 commands to use 2048 bits by default.
5453 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5455 * DH small subgroups
5457 Historically OpenSSL only ever generated DH parameters based on "safe"
5458 primes. More recently (in version 1.0.2) support was provided for
5459 generating X9.42 style parameter files such as those required for RFC 5114
5460 support. The primes used in such files may not be "safe". Where an
5461 application is using DH configured with parameters based on primes that are
5462 not "safe" then an attacker could use this fact to find a peer's private
5463 DH exponent. This attack requires that the attacker complete multiple
5464 handshakes in which the peer uses the same private DH exponent. For example
5465 this could be used to discover a TLS server's private DH exponent if it's
5466 reusing the private DH exponent or it's using a static DH ciphersuite.
5468 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5469 TLS. It is not on by default. If the option is not set then the server
5470 reuses the same private DH exponent for the life of the server process and
5471 would be vulnerable to this attack. It is believed that many popular
5472 applications do set this option and would therefore not be at risk.
5474 The fix for this issue adds an additional check where a "q" parameter is
5475 available (as is the case in X9.42 based parameters). This detects the
5476 only known attack, and is the only possible defense for static DH
5477 ciphersuites. This could have some performance impact.
5479 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5480 default and cannot be disabled. This could have some performance impact.
5482 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5487 * SSLv2 doesn't block disabled ciphers
5489 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5490 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5491 been disabled, provided that the SSLv2 protocol was not also disabled via
5494 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5495 and Sebastian Schinzel.
5500 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5502 * BN_mod_exp may produce incorrect results on x86_64
5504 There is a carry propagating bug in the x86_64 Montgomery squaring
5505 procedure. No EC algorithms are affected. Analysis suggests that attacks
5506 against RSA and DSA as a result of this defect would be very difficult to
5507 perform and are not believed likely. Attacks against DH are considered just
5508 feasible (although very difficult) because most of the work necessary to
5509 deduce information about a private key may be performed offline. The amount
5510 of resources required for such an attack would be very significant and
5511 likely only accessible to a limited number of attackers. An attacker would
5512 additionally need online access to an unpatched system using the target
5513 private key in a scenario with persistent DH parameters and a private
5514 key that is shared between multiple clients. For example this can occur by
5515 default in OpenSSL DHE based SSL/TLS ciphersuites.
5517 This issue was reported to OpenSSL by Hanno Böck.
5522 * Certificate verify crash with missing PSS parameter
5524 The signature verification routines will crash with a NULL pointer
5525 dereference if presented with an ASN.1 signature using the RSA PSS
5526 algorithm and absent mask generation function parameter. Since these
5527 routines are used to verify certificate signature algorithms this can be
5528 used to crash any certificate verification operation and exploited in a
5529 DoS attack. Any application which performs certificate verification is
5530 vulnerable including OpenSSL clients and servers which enable client
5533 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5538 * X509_ATTRIBUTE memory leak
5540 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5541 memory. This structure is used by the PKCS#7 and CMS routines so any
5542 application which reads PKCS#7 or CMS data from untrusted sources is
5543 affected. SSL/TLS is not affected.
5545 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5551 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5552 This changes the decoding behaviour for some invalid messages,
5553 though the change is mostly in the more lenient direction, and
5554 legacy behaviour is preserved as much as possible.
5558 * In DSA_generate_parameters_ex, if the provided seed is too short,
5561 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5563 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5565 * Alternate chains certificate forgery
5567 During certificate verification, OpenSSL will attempt to find an
5568 alternative certificate chain if the first attempt to build such a chain
5569 fails. An error in the implementation of this logic can mean that an
5570 attacker could cause certain checks on untrusted certificates to be
5571 bypassed, such as the CA flag, enabling them to use a valid leaf
5572 certificate to act as a CA and "issue" an invalid certificate.
5574 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5579 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5581 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5582 incompatibility in the handling of HMAC. The previous ABI has now been
5587 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5589 * Malformed ECParameters causes infinite loop
5591 When processing an ECParameters structure OpenSSL enters an infinite loop
5592 if the curve specified is over a specially malformed binary polynomial
5595 This can be used to perform denial of service against any
5596 system which processes public keys, certificate requests or
5597 certificates. This includes TLS clients and TLS servers with
5598 client authentication enabled.
5600 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5605 * Exploitable out-of-bounds read in X509_cmp_time
5607 X509_cmp_time does not properly check the length of the ASN1_TIME
5608 string and can read a few bytes out of bounds. In addition,
5609 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5612 An attacker can use this to craft malformed certificates and CRLs of
5613 various sizes and potentially cause a segmentation fault, resulting in
5614 a DoS on applications that verify certificates or CRLs. TLS clients
5615 that verify CRLs are affected. TLS clients and servers with client
5616 authentication enabled may be affected if they use custom verification
5619 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5620 independently by Hanno Böck.
5625 * PKCS7 crash with missing EnvelopedContent
5627 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5628 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5629 with missing content and trigger a NULL pointer dereference on parsing.
5631 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5632 structures from untrusted sources are affected. OpenSSL clients and
5633 servers are not affected.
5635 This issue was reported to OpenSSL by Michal Zalewski (Google).
5640 * CMS verify infinite loop with unknown hash function
5642 When verifying a signedData message the CMS code can enter an infinite loop
5643 if presented with an unknown hash function OID. This can be used to perform
5644 denial of service against any system which verifies signedData messages using
5646 This issue was reported to OpenSSL by Johannes Bauer.
5651 * Race condition handling NewSessionTicket
5653 If a NewSessionTicket is received by a multi-threaded client when attempting to
5654 reuse a previous ticket then a race condition can occur potentially leading to
5655 a double free of the ticket data.
5660 * Only support 256-bit or stronger elliptic curves with the
5661 'ecdh_auto' setting (server) or by default (client). Of supported
5662 curves, prefer P-256 (both).
5666 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5668 * ClientHello sigalgs DoS fix
5670 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5671 invalid signature algorithms extension a NULL pointer dereference will
5672 occur. This can be exploited in a DoS attack against the server.
5674 This issue was was reported to OpenSSL by David Ramos of Stanford
5678 *Stephen Henson and Matt Caswell*
5680 * Multiblock corrupted pointer fix
5682 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5683 feature only applies on 64 bit x86 architecture platforms that support AES
5684 NI instructions. A defect in the implementation of "multiblock" can cause
5685 OpenSSL's internal write buffer to become incorrectly set to NULL when
5686 using non-blocking IO. Typically, when the user application is using a
5687 socket BIO for writing, this will only result in a failed connection.
5688 However if some other BIO is used then it is likely that a segmentation
5689 fault will be triggered, thus enabling a potential DoS attack.
5691 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5696 * Segmentation fault in DTLSv1_listen fix
5698 The DTLSv1_listen function is intended to be stateless and processes the
5699 initial ClientHello from many peers. It is common for user code to loop
5700 over the call to DTLSv1_listen until a valid ClientHello is received with
5701 an associated cookie. A defect in the implementation of DTLSv1_listen means
5702 that state is preserved in the SSL object from one invocation to the next
5703 that can lead to a segmentation fault. Errors processing the initial
5704 ClientHello can trigger this scenario. An example of such an error could be
5705 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5708 This issue was reported to OpenSSL by Per Allansson.
5713 * Segmentation fault in ASN1_TYPE_cmp fix
5715 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5716 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5717 certificate signature algorithm consistency this can be used to crash any
5718 certificate verification operation and exploited in a DoS attack. Any
5719 application which performs certificate verification is vulnerable including
5720 OpenSSL clients and servers which enable client authentication.
5725 * Segmentation fault for invalid PSS parameters fix
5727 The signature verification routines will crash with a NULL pointer
5728 dereference if presented with an ASN.1 signature using the RSA PSS
5729 algorithm and invalid parameters. Since these routines are used to verify
5730 certificate signature algorithms this can be used to crash any
5731 certificate verification operation and exploited in a DoS attack. Any
5732 application which performs certificate verification is vulnerable including
5733 OpenSSL clients and servers which enable client authentication.
5735 This issue was was reported to OpenSSL by Brian Carpenter.
5740 * ASN.1 structure reuse memory corruption fix
5742 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5743 memory corruption via an invalid write. Such reuse is and has been
5744 strongly discouraged and is believed to be rare.
5746 Applications that parse structures containing CHOICE or ANY DEFINED BY
5747 components may be affected. Certificate parsing (d2i_X509 and related
5748 functions) are however not affected. OpenSSL clients and servers are
5754 * PKCS7 NULL pointer dereferences fix
5756 The PKCS#7 parsing code does not handle missing outer ContentInfo
5757 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5758 missing content and trigger a NULL pointer dereference on parsing.
5760 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5761 otherwise parse PKCS#7 structures from untrusted sources are
5762 affected. OpenSSL clients and servers are not affected.
5764 This issue was reported to OpenSSL by Michal Zalewski (Google).
5769 * DoS via reachable assert in SSLv2 servers fix
5771 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5772 servers that both support SSLv2 and enable export cipher suites by sending
5773 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5775 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5776 (OpenSSL development team).
5781 * Empty CKE with client auth and DHE fix
5783 If client auth is used then a server can seg fault in the event of a DHE
5784 ciphersuite being selected and a zero length ClientKeyExchange message
5785 being sent by the client. This could be exploited in a DoS attack.
5790 * Handshake with unseeded PRNG fix
5792 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5793 with an unseeded PRNG. The conditions are:
5794 - The client is on a platform where the PRNG has not been seeded
5795 automatically, and the user has not seeded manually
5796 - A protocol specific client method version has been used (i.e. not
5797 SSL_client_methodv23)
5798 - A ciphersuite is used that does not require additional random data from
5799 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5801 If the handshake succeeds then the client random that has been used will
5802 have been generated from a PRNG with insufficient entropy and therefore the
5803 output may be predictable.
5805 For example using the following command with an unseeded openssl will
5806 succeed on an unpatched platform:
5808 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5813 * Use After Free following d2i_ECPrivatekey error fix
5815 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5816 could cause a use after free condition. This, in turn, could cause a double
5817 free in several private key parsing functions (such as d2i_PrivateKey
5818 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5819 for applications that receive EC private keys from untrusted
5820 sources. This scenario is considered rare.
5822 This issue was discovered by the BoringSSL project and fixed in their
5828 * X509_to_X509_REQ NULL pointer deref fix
5830 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5831 the certificate key is invalid. This function is rarely used in practice.
5833 This issue was discovered by Brian Carpenter.
5838 * Removed the export ciphers from the DEFAULT ciphers
5842 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5844 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5845 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5846 So far those who have to target multiple platforms would compromise
5847 and argue that binary targeting say ARMv5 would still execute on
5848 ARMv8. "Universal" build resolves this compromise by providing
5849 near-optimal performance even on newer platforms.
5853 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5854 (other platforms pending).
5856 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5858 * Add support for the SignedCertificateTimestampList certificate and
5859 OCSP response extensions from RFC6962.
5863 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5864 for corner cases. (Certain input points at infinity could lead to
5865 bogus results, with non-infinity inputs mapped to infinity too.)
5869 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5870 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5871 common cases are optimized and there still is room for further
5872 improvements. Vector Permutation AES for Altivec is also added.
5876 * Add support for little-endian ppc64 Linux target.
5878 *Marcelo Cerri (IBM)*
5880 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5881 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5882 are optimized and there still is room for further improvements.
5883 Both 32- and 64-bit modes are supported.
5885 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5887 * Improved ARMv7 NEON support.
5891 * Support for SPARC Architecture 2011 crypto extensions, first
5892 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5893 SHA256/512, MD5, GHASH and modular exponentiation.
5895 *Andy Polyakov, David Miller*
5897 * Accelerated modular exponentiation for Intel processors, a.k.a.
5900 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5902 * Support for new and upcoming Intel processors, including AVX2,
5903 BMI and SHA ISA extensions. This includes additional "stitched"
5904 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5907 This work was sponsored by Intel Corp.
5911 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5912 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5913 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5917 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5918 this fixes a limitation in previous versions of OpenSSL.
5922 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5923 MGF1 digest and OAEP label.
5927 * Add EVP support for key wrapping algorithms, to avoid problems with
5928 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5929 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5930 algorithms and include tests cases.
5934 * Add functions to allocate and set the fields of an ECDSA_METHOD
5937 *Douglas E. Engert, Steve Henson*
5939 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5940 difference in days and seconds between two tm or ASN1_TIME structures.
5944 * Add -rev test option to s_server to just reverse order of characters
5945 received by client and send back to server. Also prints an abbreviated
5946 summary of the connection parameters.
5950 * New option -brief for s_client and s_server to print out a brief summary
5951 of connection parameters.
5955 * Add callbacks for arbitrary TLS extensions.
5957 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5959 * New option -crl_download in several openssl utilities to download CRLs
5960 from CRLDP extension in certificates.
5964 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5968 * New function X509_CRL_diff to generate a delta CRL from the difference
5969 of two full CRLs. Add support to "crl" utility.
5973 * New functions to set lookup_crls function and to retrieve
5974 X509_STORE from X509_STORE_CTX.
5978 * Print out deprecated issuer and subject unique ID fields in
5983 * Extend OCSP I/O functions so they can be used for simple general purpose
5984 HTTP as well as OCSP. New wrapper function which can be used to download
5985 CRLs using the OCSP API.
5989 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5993 * `SSL_CONF*` functions. These provide a common framework for application
5994 configuration using configuration files or command lines.
5998 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5999 message callback and prints the results. Needs compile time option
6000 "enable-ssl-trace". New options to s_client and s_server to enable
6005 * New ctrl and macro to retrieve supported points extensions.
6006 Print out extension in s_server and s_client.
6010 * New functions to retrieve certificate signature and signature
6015 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6020 * New Suite B modes for TLS code. These use and enforce the requirements
6021 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6022 only use Suite B curves. The Suite B modes can be set by using the
6023 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6027 * New chain verification flags for Suite B levels of security. Check
6028 algorithms are acceptable when flags are set in X509_verify_cert.
6032 * Make tls1_check_chain return a set of flags indicating checks passed
6033 by a certificate chain. Add additional tests to handle client
6034 certificates: checks for matching certificate type and issuer name
6039 * If an attempt is made to use a signature algorithm not in the peer
6040 preference list abort the handshake. If client has no suitable
6041 signature algorithms in response to a certificate request do not
6042 use the certificate.
6046 * If server EC tmp key is not in client preference list abort handshake.
6050 * Add support for certificate stores in CERT structure. This makes it
6051 possible to have different stores per SSL structure or one store in
6052 the parent SSL_CTX. Include distinct stores for certificate chain
6053 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6054 to build and store a certificate chain in CERT structure: returning
6055 an error if the chain cannot be built: this will allow applications
6056 to test if a chain is correctly configured.
6058 Note: if the CERT based stores are not set then the parent SSL_CTX
6059 store is used to retain compatibility with existing behaviour.
6063 * New function ssl_set_client_disabled to set a ciphersuite disabled
6064 mask based on the current session, check mask when sending client
6065 hello and checking the requested ciphersuite.
6069 * New ctrls to retrieve and set certificate types in a certificate
6070 request message. Print out received values in s_client. If certificate
6071 types is not set with custom values set sensible values based on
6072 supported signature algorithms.
6076 * Support for distinct client and server supported signature algorithms.
6080 * Add certificate callback. If set this is called whenever a certificate
6081 is required by client or server. An application can decide which
6082 certificate chain to present based on arbitrary criteria: for example
6083 supported signature algorithms. Add very simple example to s_server.
6084 This fixes many of the problems and restrictions of the existing client
6085 certificate callback: for example you can now clear an existing
6086 certificate and specify the whole chain.
6090 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6091 the certificate can be used for (if anything). Set valid_flags field
6092 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6093 to have similar checks in it.
6095 Add new "cert_flags" field to CERT structure and include a "strict mode".
6096 This enforces some TLS certificate requirements (such as only permitting
6097 certificate signature algorithms contained in the supported algorithms
6098 extension) which some implementations ignore: this option should be used
6099 with caution as it could cause interoperability issues.
6103 * Update and tidy signature algorithm extension processing. Work out
6104 shared signature algorithms based on preferences and peer algorithms
6105 and print them out in s_client and s_server. Abort handshake if no
6106 shared signature algorithms.
6110 * Add new functions to allow customised supported signature algorithms
6111 for SSL and SSL_CTX structures. Add options to s_client and s_server
6116 * New function SSL_certs_clear() to delete all references to certificates
6117 from an SSL structure. Before this once a certificate had been added
6118 it couldn't be removed.
6122 * Integrate hostname, email address and IP address checking with certificate
6123 verification. New verify options supporting checking in openssl utility.
6127 * Fixes and wildcard matching support to hostname and email checking
6128 functions. Add manual page.
6130 *Florian Weimer (Red Hat Product Security Team)*
6132 * New functions to check a hostname email or IP address against a
6133 certificate. Add options x509 utility to print results of checks against
6138 * Fix OCSP checking.
6140 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6142 * Initial experimental support for explicitly trusted non-root CAs.
6143 OpenSSL still tries to build a complete chain to a root but if an
6144 intermediate CA has a trust setting included that is used. The first
6145 setting is used: whether to trust (e.g., -addtrust option to the x509
6150 * Add -trusted_first option which attempts to find certificates in the
6151 trusted store even if an untrusted chain is also supplied.
6155 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6156 platform support for Linux and Android.
6160 * Support for linux-x32, ILP32 environment in x86_64 framework.
6164 * Experimental multi-implementation support for FIPS capable OpenSSL.
6165 When in FIPS mode the approved implementations are used as normal,
6166 when not in FIPS mode the internal unapproved versions are used instead.
6167 This means that the FIPS capable OpenSSL isn't forced to use the
6168 (often lower performance) FIPS implementations outside FIPS mode.
6172 * Transparently support X9.42 DH parameters when calling
6173 PEM_read_bio_DHparameters. This means existing applications can handle
6174 the new parameter format automatically.
6178 * Initial experimental support for X9.42 DH parameter format: mainly
6179 to support use of 'q' parameter for RFC5114 parameters.
6183 * Add DH parameters from RFC5114 including test data to dhtest.
6187 * Support for automatic EC temporary key parameter selection. If enabled
6188 the most preferred EC parameters are automatically used instead of
6189 hardcoded fixed parameters. Now a server just has to call:
6190 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6191 support ECDH and use the most appropriate parameters.
6195 * Enhance and tidy EC curve and point format TLS extension code. Use
6196 static structures instead of allocation if default values are used.
6197 New ctrls to set curves we wish to support and to retrieve shared curves.
6198 Print out shared curves in s_server. New options to s_server and s_client
6199 to set list of supported curves.
6203 * New ctrls to retrieve supported signature algorithms and
6204 supported curve values as an array of NIDs. Extend openssl utility
6205 to print out received values.
6209 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6210 between NIDs and the more common NIST names such as "P-256". Enhance
6211 ecparam utility and ECC method to recognise the NIST names for curves.
6215 * Enhance SSL/TLS certificate chain handling to support different
6216 chains for each certificate instead of one chain in the parent SSL_CTX.
6220 * Support for fixed DH ciphersuite client authentication: where both
6221 server and client use DH certificates with common parameters.
6225 * Support for fixed DH ciphersuites: those requiring DH server
6230 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6232 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6233 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6234 X509_CINF_get_signature were reverted post internal team review.
6239 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6241 * OCSP Status Request extension unbounded memory growth
6243 A malicious client can send an excessively large OCSP Status Request
6244 extension. If that client continually requests renegotiation, sending a
6245 large OCSP Status Request extension each time, then there will be unbounded
6246 memory growth on the server. This will eventually lead to a Denial Of
6247 Service attack through memory exhaustion. Servers with a default
6248 configuration are vulnerable even if they do not support OCSP. Builds using
6249 the "no-ocsp" build time option are not affected.
6251 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6256 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6259 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6265 * OOB write in MDC2_Update()
6267 An overflow can occur in MDC2_Update() either if called directly or
6268 through the EVP_DigestUpdate() function using MDC2. If an attacker
6269 is able to supply very large amounts of input data after a previous
6270 call to EVP_EncryptUpdate() with a partial block then a length check
6271 can overflow resulting in a heap corruption.
6273 The amount of data needed is comparable to SIZE_MAX which is impractical
6276 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6281 * Malformed SHA512 ticket DoS
6283 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6284 DoS attack where a malformed ticket will result in an OOB read which will
6287 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6288 a custom server callback and ticket lookup mechanism.
6290 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6295 * OOB write in BN_bn2dec()
6297 The function BN_bn2dec() does not check the return value of BN_div_word().
6298 This can cause an OOB write if an application uses this function with an
6299 overly large BIGNUM. This could be a problem if an overly large certificate
6300 or CRL is printed out from an untrusted source. TLS is not affected because
6301 record limits will reject an oversized certificate before it is parsed.
6303 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6308 * OOB read in TS_OBJ_print_bio()
6310 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6311 the total length the OID text representation would use and not the amount
6312 of data written. This will result in OOB reads when large OIDs are
6315 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6320 * Pointer arithmetic undefined behaviour
6322 Avoid some undefined pointer arithmetic
6324 A common idiom in the codebase is to check limits in the following manner:
6327 Where "p" points to some malloc'd data of SIZE bytes and
6330 "len" here could be from some externally supplied data (e.g. from a TLS
6333 The rules of C pointer arithmetic are such that "p + len" is only well
6334 defined where len <= SIZE. Therefore the above idiom is actually
6335 undefined behaviour.
6337 For example this could cause problems if some malloc implementation
6338 provides an address for "p" such that "p + len" actually overflows for
6339 values of len that are too big and therefore p + len < limit.
6341 This issue was reported to OpenSSL by Guido Vranken
6346 * Constant time flag not preserved in DSA signing
6348 Operations in the DSA signing algorithm should run in constant time in
6349 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6350 implementation means that a non-constant time codepath is followed for
6351 certain operations. This has been demonstrated through a cache-timing
6352 attack to be sufficient for an attacker to recover the private DSA key.
6354 This issue was reported by César Pereida (Aalto University), Billy Brumley
6355 (Tampere University of Technology), and Yuval Yarom (The University of
6356 Adelaide and NICTA).
6361 * DTLS buffered message DoS
6363 In a DTLS connection where handshake messages are delivered out-of-order
6364 those messages that OpenSSL is not yet ready to process will be buffered
6365 for later use. Under certain circumstances, a flaw in the logic means that
6366 those messages do not get removed from the buffer even though the handshake
6367 has been completed. An attacker could force up to approx. 15 messages to
6368 remain in the buffer when they are no longer required. These messages will
6369 be cleared when the DTLS connection is closed. The default maximum size for
6370 a message is 100k. Therefore the attacker could force an additional 1500k
6371 to be consumed per connection. By opening many simulataneous connections an
6372 attacker could cause a DoS attack through memory exhaustion.
6374 This issue was reported to OpenSSL by Quan Luo.
6379 * DTLS replay protection DoS
6381 A flaw in the DTLS replay attack protection mechanism means that records
6382 that arrive for future epochs update the replay protection "window" before
6383 the MAC for the record has been validated. This could be exploited by an
6384 attacker by sending a record for the next epoch (which does not have to
6385 decrypt or have a valid MAC), with a very large sequence number. This means
6386 that all subsequent legitimate packets are dropped causing a denial of
6387 service for a specific DTLS connection.
6389 This issue was reported to OpenSSL by the OCAP audit team.
6394 * Certificate message OOB reads
6396 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6397 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6398 theoretical DoS risk but this has not been observed in practice on common
6401 The messages affected are client certificate, client certificate request
6402 and server certificate. As a result the attack can only be performed
6403 against a client or a server which enables client authentication.
6405 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6410 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6412 * Prevent padding oracle in AES-NI CBC MAC check
6414 A MITM attacker can use a padding oracle attack to decrypt traffic
6415 when the connection uses an AES CBC cipher and the server support
6418 This issue was introduced as part of the fix for Lucky 13 padding
6419 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6420 constant time by making sure that always the same bytes are read and
6421 compared against either the MAC or padding bytes. But it no longer
6422 checked that there was enough data to have both the MAC and padding
6425 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6430 * Fix EVP_EncodeUpdate overflow
6432 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6433 Base64 encoding of binary data. If an attacker is able to supply very large
6434 amounts of input data then a length check can overflow resulting in a heap
6437 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6438 the `PEM_write_bio*` family of functions. These are mainly used within the
6439 OpenSSL command line applications, so any application which processes data
6440 from an untrusted source and outputs it as a PEM file should be considered
6441 vulnerable to this issue. User applications that call these APIs directly
6442 with large amounts of untrusted data may also be vulnerable.
6444 This issue was reported by Guido Vranken.
6449 * Fix EVP_EncryptUpdate overflow
6451 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6452 is able to supply very large amounts of input data after a previous call to
6453 EVP_EncryptUpdate() with a partial block then a length check can overflow
6454 resulting in a heap corruption. Following an analysis of all OpenSSL
6455 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6456 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6457 the first called function after an EVP_EncryptInit(), and therefore that
6458 specific call must be safe. The second form is where the length passed to
6459 EVP_EncryptUpdate() can be seen from the code to be some small value and
6460 therefore there is no possibility of an overflow. Since all instances are
6461 one of these two forms, it is believed that there can be no overflows in
6462 internal code due to this problem. It should be noted that
6463 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6464 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6465 of these calls have also been analysed too and it is believed there are no
6466 instances in internal usage where an overflow could occur.
6468 This issue was reported by Guido Vranken.
6473 * Prevent ASN.1 BIO excessive memory allocation
6475 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6476 a short invalid encoding can casuse allocation of large amounts of memory
6477 potentially consuming excessive resources or exhausting memory.
6479 Any application parsing untrusted data through d2i BIO functions is
6480 affected. The memory based functions such as d2i_X509() are *not* affected.
6481 Since the memory based functions are used by the TLS library, TLS
6482 applications are not affected.
6484 This issue was reported by Brian Carpenter.
6491 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6492 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6493 in arbitrary stack data being returned in the buffer.
6495 This issue was reported by Guido Vranken.
6500 * Modify behavior of ALPN to invoke callback after SNI/servername
6501 callback, such that updates to the SSL_CTX affect ALPN.
6505 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6510 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6511 methods are enabled and ssl2 is disabled the methods return NULL.
6515 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6517 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6518 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6519 provide any "EXPORT" or "LOW" strength ciphers.
6523 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6524 is by default disabled at build-time. Builds that are not configured with
6525 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6526 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6527 will need to explicitly call either of:
6529 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6531 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6533 as appropriate. Even if either of those is used, or the application
6534 explicitly uses the version-specific SSLv2_method() or its client and
6535 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6536 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6537 ciphers, and SSLv2 56-bit DES are no longer available.
6542 * Fix a double-free in DSA code
6544 A double free bug was discovered when OpenSSL parses malformed DSA private
6545 keys and could lead to a DoS attack or memory corruption for applications
6546 that receive DSA private keys from untrusted sources. This scenario is
6549 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6555 * Disable SRP fake user seed to address a server memory leak.
6557 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6559 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6560 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6561 was changed to ignore the "fake user" SRP seed, even if the seed
6564 Users should use SRP_VBASE_get1_by_user instead. Note that in
6565 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6566 also that even though configuring the SRP seed attempts to hide
6567 invalid usernames by continuing the handshake with fake
6568 credentials, this behaviour is not constant time and no strong
6569 guarantees are made that the handshake is indistinguishable from
6570 that of a valid user.
6575 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6577 In the BN_hex2bn function the number of hex digits is calculated using an
6578 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6579 large values of `i` this can result in `bn_expand` not allocating any
6580 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6581 field as NULL leading to a subsequent NULL ptr deref. For very large values
6582 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6583 In this case memory is allocated to the internal BIGNUM data field, but it
6584 is insufficiently sized leading to heap corruption. A similar issue exists
6585 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6586 is ever called by user applications with very large untrusted hex/dec data.
6587 This is anticipated to be a rare occurrence.
6589 All OpenSSL internal usage of these functions use data that is not expected
6590 to be untrusted, e.g. config file data or application command line
6591 arguments. If user developed applications generate config file data based
6592 on untrusted data then it is possible that this could also lead to security
6593 consequences. This is also anticipated to be rare.
6595 This issue was reported to OpenSSL by Guido Vranken.
6600 * Fix memory issues in `BIO_*printf` functions
6602 The internal `fmtstr` function used in processing a "%s" format string in
6603 the `BIO_*printf` functions could overflow while calculating the length of a
6604 string and cause an OOB read when printing very long strings.
6606 Additionally the internal `doapr_outch` function can attempt to write to an
6607 OOB memory location (at an offset from the NULL pointer) in the event of a
6608 memory allocation failure. In 1.0.2 and below this could be caused where
6609 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6610 could be in processing a very long "%s" format string. Memory leaks can
6613 The first issue may mask the second issue dependent on compiler behaviour.
6614 These problems could enable attacks where large amounts of untrusted data
6615 is passed to the `BIO_*printf` functions. If applications use these functions
6616 in this way then they could be vulnerable. OpenSSL itself uses these
6617 functions when printing out human-readable dumps of ASN.1 data. Therefore
6618 applications that print this data could be vulnerable if the data is from
6619 untrusted sources. OpenSSL command line applications could also be
6620 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6621 as command line arguments.
6623 Libssl is not considered directly vulnerable. Additionally certificates etc
6624 received via remote connections via libssl are also unlikely to be able to
6625 trigger these issues because of message size limits enforced within libssl.
6627 This issue was reported to OpenSSL Guido Vranken.
6632 * Side channel attack on modular exponentiation
6634 A side-channel attack was found which makes use of cache-bank conflicts on
6635 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6636 of RSA keys. The ability to exploit this issue is limited as it relies on
6637 an attacker who has control of code in a thread running on the same
6638 hyper-threaded core as the victim thread which is performing decryptions.
6640 This issue was reported to OpenSSL by Yuval Yarom, The University of
6641 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6642 Nadia Heninger, University of Pennsylvania with more information at
6643 <http://cachebleed.info>.
6648 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6649 if no keysize is specified with default_bits. This fixes an
6650 omission in an earlier change that changed all RSA/DSA key generation
6651 commands to use 2048 bits by default.
6655 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6657 * Protection for DH small subgroup attacks
6659 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6660 switched on by default and cannot be disabled. This could have some
6665 * SSLv2 doesn't block disabled ciphers
6667 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6668 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6669 been disabled, provided that the SSLv2 protocol was not also disabled via
6672 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6673 and Sebastian Schinzel.
6678 * Reject DH handshakes with parameters shorter than 1024 bits.
6682 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6684 * Certificate verify crash with missing PSS parameter
6686 The signature verification routines will crash with a NULL pointer
6687 dereference if presented with an ASN.1 signature using the RSA PSS
6688 algorithm and absent mask generation function parameter. Since these
6689 routines are used to verify certificate signature algorithms this can be
6690 used to crash any certificate verification operation and exploited in a
6691 DoS attack. Any application which performs certificate verification is
6692 vulnerable including OpenSSL clients and servers which enable client
6695 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6700 * X509_ATTRIBUTE memory leak
6702 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6703 memory. This structure is used by the PKCS#7 and CMS routines so any
6704 application which reads PKCS#7 or CMS data from untrusted sources is
6705 affected. SSL/TLS is not affected.
6707 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6713 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6714 This changes the decoding behaviour for some invalid messages,
6715 though the change is mostly in the more lenient direction, and
6716 legacy behaviour is preserved as much as possible.
6720 * In DSA_generate_parameters_ex, if the provided seed is too short,
6721 use a random seed, as already documented.
6723 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6725 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6727 * Alternate chains certificate forgery
6729 During certificate verfification, OpenSSL will attempt to find an
6730 alternative certificate chain if the first attempt to build such a chain
6731 fails. An error in the implementation of this logic can mean that an
6732 attacker could cause certain checks on untrusted certificates to be
6733 bypassed, such as the CA flag, enabling them to use a valid leaf
6734 certificate to act as a CA and "issue" an invalid certificate.
6736 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6742 * Race condition handling PSK identify hint
6744 If PSK identity hints are received by a multi-threaded client then
6745 the values are wrongly updated in the parent SSL_CTX structure. This can
6746 result in a race condition potentially leading to a double free of the
6752 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6754 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6755 incompatibility in the handling of HMAC. The previous ABI has now been
6758 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6760 * Malformed ECParameters causes infinite loop
6762 When processing an ECParameters structure OpenSSL enters an infinite loop
6763 if the curve specified is over a specially malformed binary polynomial
6766 This can be used to perform denial of service against any
6767 system which processes public keys, certificate requests or
6768 certificates. This includes TLS clients and TLS servers with
6769 client authentication enabled.
6771 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6776 * Exploitable out-of-bounds read in X509_cmp_time
6778 X509_cmp_time does not properly check the length of the ASN1_TIME
6779 string and can read a few bytes out of bounds. In addition,
6780 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6783 An attacker can use this to craft malformed certificates and CRLs of
6784 various sizes and potentially cause a segmentation fault, resulting in
6785 a DoS on applications that verify certificates or CRLs. TLS clients
6786 that verify CRLs are affected. TLS clients and servers with client
6787 authentication enabled may be affected if they use custom verification
6790 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6791 independently by Hanno Böck.
6796 * PKCS7 crash with missing EnvelopedContent
6798 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6799 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6800 with missing content and trigger a NULL pointer dereference on parsing.
6802 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6803 structures from untrusted sources are affected. OpenSSL clients and
6804 servers are not affected.
6806 This issue was reported to OpenSSL by Michal Zalewski (Google).
6811 * CMS verify infinite loop with unknown hash function
6813 When verifying a signedData message the CMS code can enter an infinite loop
6814 if presented with an unknown hash function OID. This can be used to perform
6815 denial of service against any system which verifies signedData messages using
6817 This issue was reported to OpenSSL by Johannes Bauer.
6822 * Race condition handling NewSessionTicket
6824 If a NewSessionTicket is received by a multi-threaded client when attempting to
6825 reuse a previous ticket then a race condition can occur potentially leading to
6826 a double free of the ticket data.
6831 * Reject DH handshakes with parameters shorter than 768 bits.
6833 *Kurt Roeckx and Emilia Kasper*
6835 * dhparam: generate 2048-bit parameters by default.
6837 *Kurt Roeckx and Emilia Kasper*
6839 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6841 * Segmentation fault in ASN1_TYPE_cmp fix
6843 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6844 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6845 certificate signature algorithm consistency this can be used to crash any
6846 certificate verification operation and exploited in a DoS attack. Any
6847 application which performs certificate verification is vulnerable including
6848 OpenSSL clients and servers which enable client authentication.
6853 * ASN.1 structure reuse memory corruption fix
6855 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6856 memory corruption via an invalid write. Such reuse is and has been
6857 strongly discouraged and is believed to be rare.
6859 Applications that parse structures containing CHOICE or ANY DEFINED BY
6860 components may be affected. Certificate parsing (d2i_X509 and related
6861 functions) are however not affected. OpenSSL clients and servers are
6867 * PKCS7 NULL pointer dereferences fix
6869 The PKCS#7 parsing code does not handle missing outer ContentInfo
6870 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6871 missing content and trigger a NULL pointer dereference on parsing.
6873 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6874 otherwise parse PKCS#7 structures from untrusted sources are
6875 affected. OpenSSL clients and servers are not affected.
6877 This issue was reported to OpenSSL by Michal Zalewski (Google).
6882 * DoS via reachable assert in SSLv2 servers fix
6884 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6885 servers that both support SSLv2 and enable export cipher suites by sending
6886 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6888 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6889 (OpenSSL development team).
6894 * Use After Free following d2i_ECPrivatekey error fix
6896 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6897 could cause a use after free condition. This, in turn, could cause a double
6898 free in several private key parsing functions (such as d2i_PrivateKey
6899 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6900 for applications that receive EC private keys from untrusted
6901 sources. This scenario is considered rare.
6903 This issue was discovered by the BoringSSL project and fixed in their
6909 * X509_to_X509_REQ NULL pointer deref fix
6911 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6912 the certificate key is invalid. This function is rarely used in practice.
6914 This issue was discovered by Brian Carpenter.
6919 * Removed the export ciphers from the DEFAULT ciphers
6923 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6925 * Build fixes for the Windows and OpenVMS platforms
6927 *Matt Caswell and Richard Levitte*
6929 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6931 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6932 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6933 dereference. This could lead to a Denial Of Service attack. Thanks to
6934 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6939 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6940 dtls1_buffer_record function under certain conditions. In particular this
6941 could occur if an attacker sent repeated DTLS records with the same
6942 sequence number but for the next epoch. The memory leak could be exploited
6943 by an attacker in a Denial of Service attack through memory exhaustion.
6944 Thanks to Chris Mueller for reporting this issue.
6949 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6950 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6951 method would be set to NULL which could later result in a NULL pointer
6952 dereference. Thanks to Frank Schmirler for reporting this issue.
6957 * Abort handshake if server key exchange message is omitted for ephemeral
6960 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6961 reporting this issue.
6966 * Remove non-export ephemeral RSA code on client and server. This code
6967 violated the TLS standard by allowing the use of temporary RSA keys in
6968 non-export ciphersuites and could be used by a server to effectively
6969 downgrade the RSA key length used to a value smaller than the server
6970 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6971 INRIA or reporting this issue.
6976 * Fixed issue where DH client certificates are accepted without verification.
6977 An OpenSSL server will accept a DH certificate for client authentication
6978 without the certificate verify message. This effectively allows a client to
6979 authenticate without the use of a private key. This only affects servers
6980 which trust a client certificate authority which issues certificates
6981 containing DH keys: these are extremely rare and hardly ever encountered.
6982 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6988 * Ensure that the session ID context of an SSL is updated when its
6989 SSL_CTX is updated via SSL_set_SSL_CTX.
6991 The session ID context is typically set from the parent SSL_CTX,
6992 and can vary with the CTX.
6996 * Fix various certificate fingerprint issues.
6998 By using non-DER or invalid encodings outside the signed portion of a
6999 certificate the fingerprint can be changed without breaking the signature.
7000 Although no details of the signed portion of the certificate can be changed
7001 this can cause problems with some applications: e.g. those using the
7002 certificate fingerprint for blacklists.
7004 1. Reject signatures with non zero unused bits.
7006 If the BIT STRING containing the signature has non zero unused bits reject
7007 the signature. All current signature algorithms require zero unused bits.
7009 2. Check certificate algorithm consistency.
7011 Check the AlgorithmIdentifier inside TBS matches the one in the
7012 certificate signature. NB: this will result in signature failure
7013 errors for some broken certificates.
7015 Thanks to Konrad Kraszewski from Google for reporting this issue.
7017 3. Check DSA/ECDSA signatures use DER.
7019 Re-encode DSA/ECDSA signatures and compare with the original received
7020 signature. Return an error if there is a mismatch.
7022 This will reject various cases including garbage after signature
7023 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7024 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7025 (negative or with leading zeroes).
7027 Further analysis was conducted and fixes were developed by Stephen Henson
7028 of the OpenSSL core team.
7034 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7035 results on some platforms, including x86_64. This bug occurs at random
7036 with a very low probability, and is not known to be exploitable in any
7037 way, though its exact impact is difficult to determine. Thanks to Pieter
7038 Wuille (Blockstream) who reported this issue and also suggested an initial
7039 fix. Further analysis was conducted by the OpenSSL development team and
7040 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7041 the OpenSSL core team.
7046 * Do not resume sessions on the server if the negotiated protocol
7047 version does not match the session's version. Resuming with a different
7048 version, while not strictly forbidden by the RFC, is of questionable
7049 sanity and breaks all known clients.
7051 *David Benjamin, Emilia Käsper*
7053 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7054 early CCS messages during renegotiation. (Note that because
7055 renegotiation is encrypted, this early CCS was not exploitable.)
7059 * Tighten client-side session ticket handling during renegotiation:
7060 ensure that the client only accepts a session ticket if the server sends
7061 the extension anew in the ServerHello. Previously, a TLS client would
7062 reuse the old extension state and thus accept a session ticket if one was
7063 announced in the initial ServerHello.
7065 Similarly, ensure that the client requires a session ticket if one
7066 was advertised in the ServerHello. Previously, a TLS client would
7067 ignore a missing NewSessionTicket message.
7071 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7075 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7076 sends a carefully crafted handshake message, to cause OpenSSL to fail
7077 to free up to 64k of memory causing a memory leak. This could be
7078 exploited in a Denial Of Service attack. This issue affects OpenSSL
7079 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7080 whether SRTP is used or configured. Implementations of OpenSSL that
7081 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7083 The fix was developed by the OpenSSL team.
7088 * Session Ticket Memory Leak.
7090 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7091 integrity of that ticket is first verified. In the event of a session
7092 ticket integrity check failing, OpenSSL will fail to free memory
7093 causing a memory leak. By sending a large number of invalid session
7094 tickets an attacker could exploit this issue in a Denial Of Service
7100 * Build option no-ssl3 is incomplete.
7102 When OpenSSL is configured with "no-ssl3" as a build option, servers
7103 could accept and complete a SSL 3.0 handshake, and clients could be
7104 configured to send them.
7107 *Akamai and the OpenSSL team*
7109 * Add support for TLS_FALLBACK_SCSV.
7110 Client applications doing fallback retries should call
7111 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7114 *Adam Langley, Bodo Moeller*
7116 * Add additional DigestInfo checks.
7118 Re-encode DigestInto in DER and check against the original when
7119 verifying RSA signature: this will reject any improperly encoded
7120 DigestInfo structures.
7122 Note: this is a precautionary measure and no attacks are currently known.
7126 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7128 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7129 SRP code can be overrun an internal buffer. Add sanity check that
7130 g, A, B < N to SRP code.
7132 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7133 Group for discovering this issue.
7138 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7139 TLS 1.0 instead of higher protocol versions when the ClientHello message
7140 is badly fragmented. This allows a man-in-the-middle attacker to force a
7141 downgrade to TLS 1.0 even if both the server and the client support a
7142 higher protocol version, by modifying the client's TLS records.
7144 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7145 researching this issue.
7150 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7151 to a denial of service attack. A malicious server can crash the client
7152 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7153 ciphersuite and sending carefully crafted handshake messages.
7155 Thanks to Felix Gröbert (Google) for discovering and researching this
7161 * By sending carefully crafted DTLS packets an attacker could cause openssl
7162 to leak memory. This can be exploited through a Denial of Service attack.
7163 Thanks to Adam Langley for discovering and researching this issue.
7168 * An attacker can force openssl to consume large amounts of memory whilst
7169 processing DTLS handshake messages. This can be exploited through a
7170 Denial of Service attack.
7171 Thanks to Adam Langley for discovering and researching this issue.
7176 * An attacker can force an error condition which causes openssl to crash
7177 whilst processing DTLS packets due to memory being freed twice. This
7178 can be exploited through a Denial of Service attack.
7179 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7185 * If a multithreaded client connects to a malicious server using a resumed
7186 session and the server sends an ec point format extension it could write
7187 up to 255 bytes to freed memory.
7189 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7195 * A malicious server can crash an OpenSSL client with a null pointer
7196 dereference (read) by specifying an SRP ciphersuite even though it was not
7197 properly negotiated with the client. This can be exploited through a
7198 Denial of Service attack.
7200 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7201 discovering and researching this issue.
7206 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7207 X509_name_oneline, X509_name_print_ex et al. to leak some information
7208 from the stack. Applications may be affected if they echo pretty printing
7209 output to the attacker.
7211 Thanks to Ivan Fratric (Google) for discovering this issue.
7214 *Emilia Käsper, and Steve Henson*
7216 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7217 for corner cases. (Certain input points at infinity could lead to
7218 bogus results, with non-infinity inputs mapped to infinity too.)
7222 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7224 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7225 handshake can force the use of weak keying material in OpenSSL
7226 SSL/TLS clients and servers.
7228 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7229 researching this issue. ([CVE-2014-0224])
7231 *KIKUCHI Masashi, Steve Henson*
7233 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7234 OpenSSL DTLS client the code can be made to recurse eventually crashing
7237 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7240 *Imre Rad, Steve Henson*
7242 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7243 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7244 client or server. This is potentially exploitable to run arbitrary
7245 code on a vulnerable client or server.
7247 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7249 *Jüri Aedla, Steve Henson*
7251 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7252 are subject to a denial of service attack.
7254 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7255 this issue. ([CVE-2014-3470])
7257 *Felix Gröbert, Ivan Fratric, Steve Henson*
7259 * Harmonize version and its documentation. -f flag is used to display
7262 *mancha <mancha1@zoho.com>*
7264 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7265 in i2d_ECPrivateKey.
7267 *mancha <mancha1@zoho.com>*
7269 * Fix some double frees. These are not thought to be exploitable.
7271 *mancha <mancha1@zoho.com>*
7273 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7275 * A missing bounds check in the handling of the TLS heartbeat extension
7276 can be used to reveal up to 64k of memory to a connected client or
7279 Thanks for Neel Mehta of Google Security for discovering this bug and to
7280 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7281 preparing the fix ([CVE-2014-0160])
7283 *Adam Langley, Bodo Moeller*
7285 * Fix for the attack described in the paper "Recovering OpenSSL
7286 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7287 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7288 <http://eprint.iacr.org/2014/140>
7290 Thanks to Yuval Yarom and Naomi Benger for discovering this
7291 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7293 *Yuval Yarom and Naomi Benger*
7295 * TLS pad extension: draft-agl-tls-padding-03
7297 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7298 TLS client Hello record length value would otherwise be > 255 and
7299 less that 512 pad with a dummy extension containing zeroes so it
7300 is at least 512 bytes long.
7302 *Adam Langley, Steve Henson*
7304 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7306 * Fix for TLS record tampering bug. A carefully crafted invalid
7307 handshake could crash OpenSSL with a NULL pointer exception.
7308 Thanks to Anton Johansson for reporting this issues.
7311 * Keep original DTLS digest and encryption contexts in retransmission
7312 structures so we can use the previous session parameters if they need
7313 to be resent. ([CVE-2013-6450])
7317 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7318 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7319 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7320 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7321 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7322 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7324 *Rob Stradling, Adam Langley*
7326 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7328 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7329 supporting platforms or when small records were transferred.
7331 *Andy Polyakov, Steve Henson*
7333 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7335 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7337 This addresses the flaw in CBC record processing discovered by
7338 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7339 at: <http://www.isg.rhul.ac.uk/tls/>
7341 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7342 Security Group at Royal Holloway, University of London
7343 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7344 Emilia Käsper for the initial patch.
7347 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7349 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7350 ciphersuites which can be exploited in a denial of service attack.
7351 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7352 and detecting this bug and to Wolfgang Ettlinger
7353 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7358 * Return an error when checking OCSP signatures when key is NULL.
7359 This fixes a DoS attack. ([CVE-2013-0166])
7363 * Make openssl verify return errors.
7365 *Chris Palmer <palmer@google.com> and Ben Laurie*
7367 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7368 the right response is stapled. Also change SSL_get_certificate()
7369 so it returns the certificate actually sent.
7370 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7372 *Rob Stradling <rob.stradling@comodo.com>*
7374 * Fix possible deadlock when decoding public keys.
7378 * Don't use TLS 1.0 record version number in initial client hello
7383 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7385 * Sanity check record length before skipping explicit IV in TLS
7386 1.2, 1.1 and DTLS to fix DoS attack.
7388 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7389 fuzzing as a service testing platform.
7394 * Initialise tkeylen properly when encrypting CMS messages.
7395 Thanks to Solar Designer of Openwall for reporting this issue.
7399 * In FIPS mode don't try to use composite ciphers as they are not
7404 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7406 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7407 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7408 mean any application compiled against OpenSSL 1.0.0 headers setting
7409 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7410 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7411 0x10000000L Any application which was previously compiled against
7412 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7413 will need to be recompiled as a result. Letting be results in
7414 inability to disable specifically TLS 1.1 and in client context,
7415 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7419 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7420 disable just protocol X, but all protocols above X *if* there are
7421 protocols *below* X still enabled. In more practical terms it means
7422 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7423 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7424 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7429 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7431 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7432 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7433 in CRYPTO_realloc_clean.
7435 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7436 issue and to Adam Langley <agl@chromium.org> for fixing it.
7439 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7441 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7445 * Workarounds for some broken servers that "hang" if a client hello
7446 record length exceeds 255 bytes.
7448 1. Do not use record version number > TLS 1.0 in initial client
7449 hello: some (but not all) hanging servers will now work.
7450 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7451 the number of ciphers sent in the client hello. This should be
7452 set to an even number, such as 50, for example by passing:
7453 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7454 Most broken servers should now work.
7455 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7456 TLS 1.2 client support entirely.
7460 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7464 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7466 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7467 STRING form instead of a DigestInfo.
7471 * The format used for MDC2 RSA signatures is inconsistent between EVP
7472 and the RSA_sign/RSA_verify functions. This was made more apparent when
7473 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7474 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7475 the correct format in RSA_verify so both forms transparently work.
7479 * Some servers which support TLS 1.0 can choke if we initially indicate
7480 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7481 encrypted premaster secret. As a workaround use the maximum permitted
7482 client version in client hello, this should keep such servers happy
7483 and still work with previous versions of OpenSSL.
7487 * Add support for TLS/DTLS heartbeats.
7489 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7491 * Add support for SCTP.
7493 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7495 * Improved PRNG seeding for VOS.
7497 *Paul Green <Paul.Green@stratus.com>*
7499 * Extensive assembler packs updates, most notably:
7501 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7502 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7503 - x86_64: bit-sliced AES implementation;
7504 - ARM: NEON support, contemporary platforms optimizations;
7505 - s390x: z196 support;
7506 - `*`: GHASH and GF(2^m) multiplication implementations;
7510 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7511 (removal of unnecessary code)
7513 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7515 * Add TLS key material exporter from RFC 5705.
7519 * Add DTLS-SRTP negotiation from RFC 5764.
7523 * Add Next Protocol Negotiation,
7524 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7525 disabled with a no-npn flag to config or Configure. Code donated
7528 *Adam Langley <agl@google.com> and Ben Laurie*
7530 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7531 NIST-P256, NIST-P521, with constant-time single point multiplication on
7532 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7533 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7534 Code made available under Apache License version 2.0.
7536 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7537 line to include this in your build of OpenSSL, and run "make depend" (or
7538 "make update"). This enables the following EC_METHODs:
7540 EC_GFp_nistp224_method()
7541 EC_GFp_nistp256_method()
7542 EC_GFp_nistp521_method()
7544 EC_GROUP_new_by_curve_name() will automatically use these (while
7545 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7548 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7550 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7551 all platforms. Move ssize_t definition from e_os.h to the public
7552 header file e_os2.h as it now appears in public header file cms.h
7556 * New -sigopt option to the ca, req and x509 utilities. Additional
7557 signature parameters can be passed using this option and in
7562 * Add RSA PSS signing function. This will generate and set the
7563 appropriate AlgorithmIdentifiers for PSS based on those in the
7564 corresponding EVP_MD_CTX structure. No application support yet.
7568 * Support for companion algorithm specific ASN1 signing routines.
7569 New function ASN1_item_sign_ctx() signs a pre-initialised
7570 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7571 the appropriate parameters.
7575 * Add new algorithm specific ASN1 verification initialisation function
7576 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7577 handling will be the same no matter what EVP_PKEY_METHOD is used.
7578 Add a PSS handler to support verification of PSS signatures: checked
7579 against a number of sample certificates.
7583 * Add signature printing for PSS. Add PSS OIDs.
7585 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7587 * Add algorithm specific signature printing. An individual ASN1 method
7588 can now print out signatures instead of the standard hex dump.
7590 More complex signatures (e.g. PSS) can print out more meaningful
7591 information. Include DSA version that prints out the signature
7596 * Password based recipient info support for CMS library: implementing
7601 * Split password based encryption into PBES2 and PBKDF2 functions. This
7602 neatly separates the code into cipher and PBE sections and is required
7603 for some algorithms that split PBES2 into separate pieces (such as
7604 password based CMS).
7608 * Session-handling fixes:
7609 - Fix handling of connections that are resuming with a session ID,
7610 but also support Session Tickets.
7611 - Fix a bug that suppressed issuing of a new ticket if the client
7612 presented a ticket with an expired session.
7613 - Try to set the ticket lifetime hint to something reasonable.
7614 - Make tickets shorter by excluding irrelevant information.
7615 - On the client side, don't ignore renewed tickets.
7617 *Adam Langley, Bodo Moeller (Google)*
7619 * Fix PSK session representation.
7623 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7625 This work was sponsored by Intel.
7629 * Add GCM support to TLS library. Some custom code is needed to split
7630 the IV between the fixed (from PRF) and explicit (from TLS record)
7631 portions. This adds all GCM ciphersuites supported by RFC5288 and
7632 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7633 add a special AESGCM string for GCM only.
7637 * Expand range of ctrls for AES GCM. Permit setting invocation
7638 field on decrypt and retrieval of invocation field only on encrypt.
7642 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7643 As required by RFC5289 these ciphersuites cannot be used if for
7644 versions of TLS earlier than 1.2.
7648 * For FIPS capable OpenSSL interpret a NULL default public key method
7649 as unset and return the appropriate default but do *not* set the default.
7650 This means we can return the appropriate method in applications that
7651 switch between FIPS and non-FIPS modes.
7655 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7656 ENGINE is used then we cannot handle that in the FIPS module so we
7657 keep original code iff non-FIPS operations are allowed.
7661 * Add -attime option to openssl utilities.
7663 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7665 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7669 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7670 FIPS EC methods unconditionally for now.
7674 * New build option no-ec2m to disable characteristic 2 code.
7678 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7679 all cases can be covered as some introduce binary incompatibilities.
7683 * Redirect RSA operations to FIPS module including keygen,
7684 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7688 * Add similar low-level API blocking to ciphers.
7692 * low-level digest APIs are not approved in FIPS mode: any attempt
7693 to use these will cause a fatal error. Applications that *really* want
7694 to use them can use the `private_*` version instead.
7698 * Redirect cipher operations to FIPS module for FIPS builds.
7702 * Redirect digest operations to FIPS module for FIPS builds.
7706 * Update build system to add "fips" flag which will link in fipscanister.o
7707 for static and shared library builds embedding a signature if needed.
7711 * Output TLS supported curves in preference order instead of numerical
7712 order. This is currently hardcoded for the highest order curves first.
7713 This should be configurable so applications can judge speed vs strength.
7717 * Add TLS v1.2 server support for client authentication.
7721 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7726 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7727 FIPS modules versions.
7731 * Add TLS v1.2 client side support for client authentication. Keep cache
7732 of handshake records longer as we don't know the hash algorithm to use
7733 until after the certificate request message is received.
7737 * Initial TLS v1.2 client support. Add a default signature algorithms
7738 extension including all the algorithms we support. Parse new signature
7739 format in client key exchange. Relax some ECC signing restrictions for
7740 TLS v1.2 as indicated in RFC5246.
7744 * Add server support for TLS v1.2 signature algorithms extension. Switch
7745 to new signature format when needed using client digest preference.
7746 All server ciphersuites should now work correctly in TLS v1.2. No client
7747 support yet and no support for client certificates.
7751 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7752 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7753 ciphersuites. At present only RSA key exchange ciphersuites work with
7754 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7755 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7756 and version checking.
7760 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7761 with this defined it will not be affected by any changes to ssl internal
7762 structures. Add several utility functions to allow openssl application
7763 to work with OPENSSL_NO_SSL_INTERN defined.
7767 * A long standing patch to add support for SRP from EdelWeb (Peter
7768 Sylvester and Christophe Renou) was integrated.
7769 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7770 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7773 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7777 * Permit abbreviated handshakes when renegotiating using the function
7778 SSL_renegotiate_abbreviated().
7780 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7782 * Add call to ENGINE_register_all_complete() to
7783 ENGINE_load_builtin_engines(), so some implementations get used
7784 automatically instead of needing explicit application support.
7788 * Add support for TLS key exporter as described in RFC5705.
7790 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7792 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7793 a few changes are required:
7795 Add SSL_OP_NO_TLSv1_1 flag.
7796 Add TLSv1_1 methods.
7797 Update version checking logic to handle version 1.1.
7798 Add explicit IV handling (ported from DTLS code).
7799 Add command line options to s_client/s_server.
7806 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7808 * X509_ATTRIBUTE memory leak
7810 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7811 memory. This structure is used by the PKCS#7 and CMS routines so any
7812 application which reads PKCS#7 or CMS data from untrusted sources is
7813 affected. SSL/TLS is not affected.
7815 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7821 * Race condition handling PSK identify hint
7823 If PSK identity hints are received by a multi-threaded client then
7824 the values are wrongly updated in the parent SSL_CTX structure. This can
7825 result in a race condition potentially leading to a double free of the
7831 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7833 * Malformed ECParameters causes infinite loop
7835 When processing an ECParameters structure OpenSSL enters an infinite loop
7836 if the curve specified is over a specially malformed binary polynomial
7839 This can be used to perform denial of service against any
7840 system which processes public keys, certificate requests or
7841 certificates. This includes TLS clients and TLS servers with
7842 client authentication enabled.
7844 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7849 * Exploitable out-of-bounds read in X509_cmp_time
7851 X509_cmp_time does not properly check the length of the ASN1_TIME
7852 string and can read a few bytes out of bounds. In addition,
7853 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7856 An attacker can use this to craft malformed certificates and CRLs of
7857 various sizes and potentially cause a segmentation fault, resulting in
7858 a DoS on applications that verify certificates or CRLs. TLS clients
7859 that verify CRLs are affected. TLS clients and servers with client
7860 authentication enabled may be affected if they use custom verification
7863 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7864 independently by Hanno Böck.
7869 * PKCS7 crash with missing EnvelopedContent
7871 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7872 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7873 with missing content and trigger a NULL pointer dereference on parsing.
7875 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7876 structures from untrusted sources are affected. OpenSSL clients and
7877 servers are not affected.
7879 This issue was reported to OpenSSL by Michal Zalewski (Google).
7884 * CMS verify infinite loop with unknown hash function
7886 When verifying a signedData message the CMS code can enter an infinite loop
7887 if presented with an unknown hash function OID. This can be used to perform
7888 denial of service against any system which verifies signedData messages using
7890 This issue was reported to OpenSSL by Johannes Bauer.
7895 * Race condition handling NewSessionTicket
7897 If a NewSessionTicket is received by a multi-threaded client when attempting to
7898 reuse a previous ticket then a race condition can occur potentially leading to
7899 a double free of the ticket data.
7904 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7906 * Segmentation fault in ASN1_TYPE_cmp fix
7908 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7909 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7910 certificate signature algorithm consistency this can be used to crash any
7911 certificate verification operation and exploited in a DoS attack. Any
7912 application which performs certificate verification is vulnerable including
7913 OpenSSL clients and servers which enable client authentication.
7918 * ASN.1 structure reuse memory corruption fix
7920 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7921 memory corruption via an invalid write. Such reuse is and has been
7922 strongly discouraged and is believed to be rare.
7924 Applications that parse structures containing CHOICE or ANY DEFINED BY
7925 components may be affected. Certificate parsing (d2i_X509 and related
7926 functions) are however not affected. OpenSSL clients and servers are
7932 * PKCS7 NULL pointer dereferences fix
7934 The PKCS#7 parsing code does not handle missing outer ContentInfo
7935 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7936 missing content and trigger a NULL pointer dereference on parsing.
7938 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7939 otherwise parse PKCS#7 structures from untrusted sources are
7940 affected. OpenSSL clients and servers are not affected.
7942 This issue was reported to OpenSSL by Michal Zalewski (Google).
7947 * DoS via reachable assert in SSLv2 servers fix
7949 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7950 servers that both support SSLv2 and enable export cipher suites by sending
7951 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7953 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7954 (OpenSSL development team).
7959 * Use After Free following d2i_ECPrivatekey error fix
7961 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7962 could cause a use after free condition. This, in turn, could cause a double
7963 free in several private key parsing functions (such as d2i_PrivateKey
7964 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7965 for applications that receive EC private keys from untrusted
7966 sources. This scenario is considered rare.
7968 This issue was discovered by the BoringSSL project and fixed in their
7974 * X509_to_X509_REQ NULL pointer deref fix
7976 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7977 the certificate key is invalid. This function is rarely used in practice.
7979 This issue was discovered by Brian Carpenter.
7984 * Removed the export ciphers from the DEFAULT ciphers
7988 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7990 * Build fixes for the Windows and OpenVMS platforms
7992 *Matt Caswell and Richard Levitte*
7994 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7996 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7997 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7998 dereference. This could lead to a Denial Of Service attack. Thanks to
7999 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8004 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8005 dtls1_buffer_record function under certain conditions. In particular this
8006 could occur if an attacker sent repeated DTLS records with the same
8007 sequence number but for the next epoch. The memory leak could be exploited
8008 by an attacker in a Denial of Service attack through memory exhaustion.
8009 Thanks to Chris Mueller for reporting this issue.
8014 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8015 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8016 method would be set to NULL which could later result in a NULL pointer
8017 dereference. Thanks to Frank Schmirler for reporting this issue.
8022 * Abort handshake if server key exchange message is omitted for ephemeral
8025 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8026 reporting this issue.
8031 * Remove non-export ephemeral RSA code on client and server. This code
8032 violated the TLS standard by allowing the use of temporary RSA keys in
8033 non-export ciphersuites and could be used by a server to effectively
8034 downgrade the RSA key length used to a value smaller than the server
8035 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8036 INRIA or reporting this issue.
8041 * Fixed issue where DH client certificates are accepted without verification.
8042 An OpenSSL server will accept a DH certificate for client authentication
8043 without the certificate verify message. This effectively allows a client to
8044 authenticate without the use of a private key. This only affects servers
8045 which trust a client certificate authority which issues certificates
8046 containing DH keys: these are extremely rare and hardly ever encountered.
8047 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8053 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8054 results on some platforms, including x86_64. This bug occurs at random
8055 with a very low probability, and is not known to be exploitable in any
8056 way, though its exact impact is difficult to determine. Thanks to Pieter
8057 Wuille (Blockstream) who reported this issue and also suggested an initial
8058 fix. Further analysis was conducted by the OpenSSL development team and
8059 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8060 the OpenSSL core team.
8065 * Fix various certificate fingerprint issues.
8067 By using non-DER or invalid encodings outside the signed portion of a
8068 certificate the fingerprint can be changed without breaking the signature.
8069 Although no details of the signed portion of the certificate can be changed
8070 this can cause problems with some applications: e.g. those using the
8071 certificate fingerprint for blacklists.
8073 1. Reject signatures with non zero unused bits.
8075 If the BIT STRING containing the signature has non zero unused bits reject
8076 the signature. All current signature algorithms require zero unused bits.
8078 2. Check certificate algorithm consistency.
8080 Check the AlgorithmIdentifier inside TBS matches the one in the
8081 certificate signature. NB: this will result in signature failure
8082 errors for some broken certificates.
8084 Thanks to Konrad Kraszewski from Google for reporting this issue.
8086 3. Check DSA/ECDSA signatures use DER.
8088 Reencode DSA/ECDSA signatures and compare with the original received
8089 signature. Return an error if there is a mismatch.
8091 This will reject various cases including garbage after signature
8092 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8093 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8094 (negative or with leading zeroes).
8096 Further analysis was conducted and fixes were developed by Stephen Henson
8097 of the OpenSSL core team.
8103 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8105 * Session Ticket Memory Leak.
8107 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8108 integrity of that ticket is first verified. In the event of a session
8109 ticket integrity check failing, OpenSSL will fail to free memory
8110 causing a memory leak. By sending a large number of invalid session
8111 tickets an attacker could exploit this issue in a Denial Of Service
8117 * Build option no-ssl3 is incomplete.
8119 When OpenSSL is configured with "no-ssl3" as a build option, servers
8120 could accept and complete a SSL 3.0 handshake, and clients could be
8121 configured to send them.
8124 *Akamai and the OpenSSL team*
8126 * Add support for TLS_FALLBACK_SCSV.
8127 Client applications doing fallback retries should call
8128 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8131 *Adam Langley, Bodo Moeller*
8133 * Add additional DigestInfo checks.
8135 Reencode DigestInto in DER and check against the original when
8136 verifying RSA signature: this will reject any improperly encoded
8137 DigestInfo structures.
8139 Note: this is a precautionary measure and no attacks are currently known.
8143 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8145 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8146 to a denial of service attack. A malicious server can crash the client
8147 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8148 ciphersuite and sending carefully crafted handshake messages.
8150 Thanks to Felix Gröbert (Google) for discovering and researching this
8156 * By sending carefully crafted DTLS packets an attacker could cause openssl
8157 to leak memory. This can be exploited through a Denial of Service attack.
8158 Thanks to Adam Langley for discovering and researching this issue.
8163 * An attacker can force openssl to consume large amounts of memory whilst
8164 processing DTLS handshake messages. This can be exploited through a
8165 Denial of Service attack.
8166 Thanks to Adam Langley for discovering and researching this issue.
8171 * An attacker can force an error condition which causes openssl to crash
8172 whilst processing DTLS packets due to memory being freed twice. This
8173 can be exploited through a Denial of Service attack.
8174 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8180 * If a multithreaded client connects to a malicious server using a resumed
8181 session and the server sends an ec point format extension it could write
8182 up to 255 bytes to freed memory.
8184 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8190 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8191 X509_name_oneline, X509_name_print_ex et al. to leak some information
8192 from the stack. Applications may be affected if they echo pretty printing
8193 output to the attacker.
8195 Thanks to Ivan Fratric (Google) for discovering this issue.
8198 *Emilia Käsper, and Steve Henson*
8200 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8201 for corner cases. (Certain input points at infinity could lead to
8202 bogus results, with non-infinity inputs mapped to infinity too.)
8206 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8208 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8209 handshake can force the use of weak keying material in OpenSSL
8210 SSL/TLS clients and servers.
8212 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8213 researching this issue. ([CVE-2014-0224])
8215 *KIKUCHI Masashi, Steve Henson*
8217 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8218 OpenSSL DTLS client the code can be made to recurse eventually crashing
8221 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8224 *Imre Rad, Steve Henson*
8226 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8227 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8228 client or server. This is potentially exploitable to run arbitrary
8229 code on a vulnerable client or server.
8231 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8233 *Jüri Aedla, Steve Henson*
8235 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8236 are subject to a denial of service attack.
8238 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8239 this issue. ([CVE-2014-3470])
8241 *Felix Gröbert, Ivan Fratric, Steve Henson*
8243 * Harmonize version and its documentation. -f flag is used to display
8246 *mancha <mancha1@zoho.com>*
8248 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8249 in i2d_ECPrivateKey.
8251 *mancha <mancha1@zoho.com>*
8253 * Fix some double frees. These are not thought to be exploitable.
8255 *mancha <mancha1@zoho.com>*
8257 * Fix for the attack described in the paper "Recovering OpenSSL
8258 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8259 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8260 <http://eprint.iacr.org/2014/140>
8262 Thanks to Yuval Yarom and Naomi Benger for discovering this
8263 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8265 *Yuval Yarom and Naomi Benger*
8267 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8269 * Keep original DTLS digest and encryption contexts in retransmission
8270 structures so we can use the previous session parameters if they need
8271 to be resent. ([CVE-2013-6450])
8275 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8276 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8277 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8278 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8279 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8280 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8282 *Rob Stradling, Adam Langley*
8284 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8286 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8288 This addresses the flaw in CBC record processing discovered by
8289 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8290 at: <http://www.isg.rhul.ac.uk/tls/>
8292 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8293 Security Group at Royal Holloway, University of London
8294 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8295 Emilia Käsper for the initial patch.
8298 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8300 * Return an error when checking OCSP signatures when key is NULL.
8301 This fixes a DoS attack. ([CVE-2013-0166])
8305 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8306 the right response is stapled. Also change SSL_get_certificate()
8307 so it returns the certificate actually sent.
8308 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8309 (This is a backport)
8311 *Rob Stradling <rob.stradling@comodo.com>*
8313 * Fix possible deadlock when decoding public keys.
8317 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8319 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8322 * Sanity check record length before skipping explicit IV in DTLS
8325 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8326 fuzzing as a service testing platform.
8331 * Initialise tkeylen properly when encrypting CMS messages.
8332 Thanks to Solar Designer of Openwall for reporting this issue.
8336 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8338 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8339 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8340 in CRYPTO_realloc_clean.
8342 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8343 issue and to Adam Langley <agl@chromium.org> for fixing it.
8346 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8348 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8350 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8351 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8352 content decryption and always return the same error. Note: this attack
8353 needs on average 2^20 messages so it only affects automated senders. The
8354 old behaviour can be re-enabled in the CMS code by setting the
8355 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8356 an MMA defence is not necessary.
8357 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8358 this issue. ([CVE-2012-0884])
8362 * Fix CVE-2011-4619: make sure we really are receiving a
8363 client hello before rejecting multiple SGC restarts. Thanks to
8364 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8368 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8370 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8371 Thanks to Antonio Martin, Enterprise Secure Access Research and
8372 Development, Cisco Systems, Inc. for discovering this bug and
8373 preparing a fix. ([CVE-2012-0050])
8377 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8379 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8380 of the Vaudenay padding oracle attack on CBC mode encryption
8381 which enables an efficient plaintext recovery attack against
8382 the OpenSSL implementation of DTLS. Their attack exploits timing
8383 differences arising during decryption processing. A research
8384 paper describing this attack can be found at:
8385 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8386 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8387 Security Group at Royal Holloway, University of London
8388 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8389 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8390 for preparing the fix. ([CVE-2011-4108])
8392 *Robin Seggelmann, Michael Tuexen*
8394 * Clear bytes used for block padding of SSL 3.0 records.
8397 *Adam Langley (Google)*
8399 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8400 Kadianakis <desnacked@gmail.com> for discovering this issue and
8401 Adam Langley for preparing the fix. ([CVE-2011-4619])
8403 *Adam Langley (Google)*
8405 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8407 *Andrey Kulikov <amdeich@gmail.com>*
8409 * Prevent malformed RFC3779 data triggering an assertion failure.
8410 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8411 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8413 *Rob Austein <sra@hactrn.net>*
8415 * Improved PRNG seeding for VOS.
8417 *Paul Green <Paul.Green@stratus.com>*
8419 * Fix ssl_ciph.c set-up race.
8421 *Adam Langley (Google)*
8423 * Fix spurious failures in ecdsatest.c.
8425 *Emilia Käsper (Google)*
8427 * Fix the BIO_f_buffer() implementation (which was mixing different
8428 interpretations of the `..._len` fields).
8430 *Adam Langley (Google)*
8432 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8433 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8434 threads won't reuse the same blinding coefficients.
8436 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8437 lock to call BN_BLINDING_invert_ex, and avoids one use of
8438 BN_BLINDING_update for each BN_BLINDING structure (previously,
8439 the last update always remained unused).
8441 *Emilia Käsper (Google)*
8443 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8445 *Bob Buckholz (Google)*
8447 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8449 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8450 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8452 *Kaspar Brand <ossl@velox.ch>*
8454 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8455 for multi-threaded use of ECDH. ([CVE-2011-3210])
8457 *Adam Langley (Google)*
8459 * Fix x509_name_ex_d2i memory leak on bad inputs.
8463 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8464 signature public key algorithm by using OID xref utilities instead.
8465 Before this you could only use some ECC ciphersuites with SHA1 only.
8469 * Add protection against ECDSA timing attacks as mentioned in the paper
8470 by Billy Bob Brumley and Nicola Tuveri, see:
8471 <http://eprint.iacr.org/2011/232.pdf>
8473 *Billy Bob Brumley and Nicola Tuveri*
8475 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8477 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8479 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8481 * Fix bug in string printing code: if *any* escaping is enabled we must
8482 escape the escape character (backslash) or the resulting string is
8487 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8489 * Disable code workaround for ancient and obsolete Netscape browsers
8490 and servers: an attacker can use it in a ciphersuite downgrade attack.
8491 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8495 * Fixed J-PAKE implementation error, originally discovered by
8496 Sebastien Martini, further info and confirmation from Stefan
8497 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8501 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8503 * Fix extension code to avoid race conditions which can result in a buffer
8504 overrun vulnerability: resumed sessions must not be modified as they can
8505 be shared by multiple threads. CVE-2010-3864
8509 * Fix WIN32 build system to correctly link an ENGINE directory into
8514 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8516 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8519 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8521 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8523 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8524 context. The operation can be customised via the ctrl mechanism in
8525 case ENGINEs want to include additional functionality.
8529 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8533 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8534 output hashes compatible with older versions of OpenSSL.
8536 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8538 * Fix compression algorithm handling: if resuming a session use the
8539 compression algorithm of the resumed session instead of determining
8540 it from client hello again. Don't allow server to change algorithm.
8544 * Add load_crls() function to commands tidying load_certs() too. Add option
8545 to verify utility to allow additional CRLs to be included.
8549 * Update OCSP request code to permit adding custom headers to the request:
8550 some responders need this.
8554 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8557 *Julia Lawall <julia@diku.dk>*
8559 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8560 needlessly dereferenced structures, used obsolete functions and
8561 didn't handle all updated verify codes correctly.
8565 * Disable MD2 in the default configuration.
8569 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8570 indicate the initial BIO being pushed or popped. This makes it possible
8571 to determine whether the BIO is the one explicitly called or as a result
8572 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8573 it handles reference counts correctly and doesn't zero out the I/O bio
8574 when it is not being explicitly popped. WARNING: applications which
8575 included workarounds for the old buggy behaviour will need to be modified
8576 or they could free up already freed BIOs.
8580 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8581 renaming to all platforms (within the 0.9.8 branch, this was
8582 done conditionally on Netware platforms to avoid a name clash).
8584 *Guenter <lists@gknw.net>*
8586 * Add ECDHE and PSK support to DTLS.
8588 *Michael Tuexen <tuexen@fh-muenster.de>*
8590 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8595 * Add "missing" function EVP_MD_flags() (without this the only way to
8596 retrieve a digest flags is by accessing the structure directly. Update
8597 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8598 or cipher is registered as in the "from" argument. Print out all
8599 registered digests in the dgst usage message instead of manually
8600 attempting to work them out.
8604 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8605 this allows the use of compression and extensions. Change default cipher
8606 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8607 by default unless an application cipher string requests it.
8611 * Alter match criteria in PKCS12_parse(). It used to try to use local
8612 key ids to find matching certificates and keys but some PKCS#12 files
8613 don't follow the (somewhat unwritten) rules and this strategy fails.
8614 Now just gather all certificates together and the first private key
8615 then look for the first certificate that matches the key.
8619 * Support use of registered digest and cipher names for dgst and cipher
8620 commands instead of having to add each one as a special case. So now
8627 openssl dgst -sha256 foo
8629 and this works for ENGINE based algorithms too.
8633 * Update Gost ENGINE to support parameter files.
8635 *Victor B. Wagner <vitus@cryptocom.ru>*
8637 * Support GeneralizedTime in ca utility.
8639 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8641 * Enhance the hash format used for certificate directory links. The new
8642 form uses the canonical encoding (meaning equivalent names will work
8643 even if they aren't identical) and uses SHA1 instead of MD5. This form
8644 is incompatible with the older format and as a result c_rehash should
8645 be used to rebuild symbolic links.
8649 * Make PKCS#8 the default write format for private keys, replacing the
8650 traditional format. This form is standardised, more secure and doesn't
8651 include an implicit MD5 dependency.
8655 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8656 committed to OpenSSL should pass this lot as a minimum.
8660 * Add session ticket override functionality for use by EAP-FAST.
8662 *Jouni Malinen <j@w1.fi>*
8664 * Modify HMAC functions to return a value. Since these can be implemented
8665 in an ENGINE errors can occur.
8669 * Type-checked OBJ_bsearch_ex.
8673 * Type-checked OBJ_bsearch. Also some constification necessitated
8674 by type-checking. Still to come: TXT_DB, bsearch(?),
8675 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8680 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8681 seconds to a tm structure directly, instead of going through OS
8682 specific date routines. This avoids any issues with OS routines such
8683 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8684 and X509_time_adj_ex() to cover the extended range. The existing
8685 X509_time_adj() is still usable and will no longer have any date issues.
8689 * Delta CRL support. New use deltas option which will attempt to locate
8690 and search any appropriate delta CRLs available.
8692 This work was sponsored by Google.
8696 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8697 code and add additional score elements. Validate alternate CRL paths
8698 as part of the CRL checking and indicate a new error "CRL path validation
8699 error" in this case. Applications wanting additional details can use
8700 the verify callback and check the new "parent" field. If this is not
8701 NULL CRL path validation is taking place. Existing applications won't
8702 see this because it requires extended CRL support which is off by
8705 This work was sponsored by Google.
8709 * Support for freshest CRL extension.
8711 This work was sponsored by Google.
8715 * Initial indirect CRL support. Currently only supported in the CRLs
8716 passed directly and not via lookup. Process certificate issuer
8717 CRL entry extension and lookup CRL entries by bother issuer name
8718 and serial number. Check and process CRL issuer entry in IDP extension.
8720 This work was sponsored by Google.
8724 * Add support for distinct certificate and CRL paths. The CRL issuer
8725 certificate is validated separately in this case. Only enabled if
8726 an extended CRL support flag is set: this flag will enable additional
8727 CRL functionality in future.
8729 This work was sponsored by Google.
8733 * Add support for policy mappings extension.
8735 This work was sponsored by Google.
8739 * Fixes to pathlength constraint, self issued certificate handling,
8740 policy processing to align with RFC3280 and PKITS tests.
8742 This work was sponsored by Google.
8746 * Support for name constraints certificate extension. DN, email, DNS
8747 and URI types are currently supported.
8749 This work was sponsored by Google.
8753 * To cater for systems that provide a pointer-based thread ID rather
8754 than numeric, deprecate the current numeric thread ID mechanism and
8755 replace it with a structure and associated callback type. This
8756 mechanism allows a numeric "hash" to be extracted from a thread ID in
8757 either case, and on platforms where pointers are larger than 'long',
8758 mixing is done to help ensure the numeric 'hash' is usable even if it
8759 can't be guaranteed unique. The default mechanism is to use "&errno"
8760 as a pointer-based thread ID to distinguish between threads.
8762 Applications that want to provide their own thread IDs should now use
8763 CRYPTO_THREADID_set_callback() to register a callback that will call
8764 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8766 Note that ERR_remove_state() is now deprecated, because it is tied
8767 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8768 to free the current thread's error state should be replaced by
8769 ERR_remove_thread_state(NULL).
8771 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8772 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8773 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8774 application was previously providing a numeric thread callback that
8775 was inappropriate for distinguishing threads, then uniqueness might
8776 have been obtained with &errno that happened immediately in the
8777 intermediate development versions of OpenSSL; this is no longer the
8778 case, the numeric thread callback will now override the automatic use
8781 *Geoff Thorpe, with help from Bodo Moeller*
8783 * Initial support for different CRL issuing certificates. This covers a
8784 simple case where the self issued certificates in the chain exist and
8785 the real CRL issuer is higher in the existing chain.
8787 This work was sponsored by Google.
8791 * Removed effectively defunct crypto/store from the build.
8795 * Revamp of STACK to provide stronger type-checking. Still to come:
8796 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8797 ASN1_STRING, CONF_VALUE.
8801 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8802 RAM on SSL connections. This option can save about 34k per idle SSL.
8806 * Revamp of LHASH to provide stronger type-checking. Still to come:
8807 STACK, TXT_DB, bsearch, qsort.
8811 * Initial support for Cryptographic Message Syntax (aka CMS) based
8812 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8813 support for data, signedData, compressedData, digestedData and
8814 encryptedData, envelopedData types included. Scripts to check against
8815 RFC4134 examples draft and interop and consistency checks of many
8816 content types and variants.
8820 * Add options to enc utility to support use of zlib compression BIO.
8824 * Extend mk1mf to support importing of options and assembly language
8825 files from Configure script, currently only included in VC-WIN32.
8826 The assembly language rules can now optionally generate the source
8827 files from the associated perl scripts.
8831 * Implement remaining functionality needed to support GOST ciphersuites.
8832 Interop testing has been performed using CryptoPro implementations.
8834 *Victor B. Wagner <vitus@cryptocom.ru>*
8836 * s390x assembler pack.
8840 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8845 * Implement Opaque PRF Input TLS extension as specified in
8846 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8847 official specification yet and no extension type assignment by
8848 IANA exists, this extension (for now) will have to be explicitly
8849 enabled when building OpenSSL by providing the extension number
8850 to use. For example, specify an option
8852 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8854 to the "config" or "Configure" script to enable the extension,
8855 assuming extension number 0x9527 (which is a completely arbitrary
8856 and unofficial assignment based on the MD5 hash of the Internet
8857 Draft). Note that by doing so, you potentially lose
8858 interoperability with other TLS implementations since these might
8859 be using the same extension number for other purposes.
8861 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8862 opaque PRF input value to use in the handshake. This will create
8863 an internal copy of the length-'len' string at 'src', and will
8864 return non-zero for success.
8866 To get more control and flexibility, provide a callback function
8869 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8870 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8874 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8877 Callback function 'cb' will be called in handshakes, and is
8878 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8879 Argument 'arg' is for application purposes (the value as given to
8880 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8881 be provided to the callback function). The callback function
8882 has to return non-zero to report success: usually 1 to use opaque
8883 PRF input just if possible, or 2 to enforce use of the opaque PRF
8884 input. In the latter case, the library will abort the handshake
8885 if opaque PRF input is not successfully negotiated.
8887 Arguments 'peerinput' and 'len' given to the callback function
8888 will always be NULL and 0 in the case of a client. A server will
8889 see the client's opaque PRF input through these variables if
8890 available (NULL and 0 otherwise). Note that if the server
8891 provides an opaque PRF input, the length must be the same as the
8892 length of the client's opaque PRF input.
8894 Note that the callback function will only be called when creating
8895 a new session (session resumption can resume whatever was
8896 previously negotiated), and will not be called in SSL 2.0
8897 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8898 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8899 for applications that need to enforce opaque PRF input.
8903 * Update ssl code to support digests other than SHA1+MD5 for handshake
8906 *Victor B. Wagner <vitus@cryptocom.ru>*
8908 * Add RFC4507 support to OpenSSL. This includes the corrections in
8909 RFC4507bis. The encrypted ticket format is an encrypted encoded
8910 SSL_SESSION structure, that way new session features are automatically
8913 If a client application caches session in an SSL_SESSION structure
8914 support is transparent because tickets are now stored in the encoded
8917 The SSL_CTX structure automatically generates keys for ticket
8918 protection in servers so again support should be possible
8919 with no application modification.
8921 If a client or server wishes to disable RFC4507 support then the option
8922 SSL_OP_NO_TICKET can be set.
8924 Add a TLS extension debugging callback to allow the contents of any client
8925 or server extensions to be examined.
8927 This work was sponsored by Google.
8931 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8932 OpenSSL should now compile cleanly on gcc 4.2
8934 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8936 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8937 support including streaming MAC support: this is required for GOST
8938 ciphersuite support.
8940 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8942 * Add option -stream to use PKCS#7 streaming in smime utility. New
8943 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8944 to output in BER and PEM format.
8948 * Experimental support for use of HMAC via EVP_PKEY interface. This
8949 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8950 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8951 ENGINE support for HMAC keys which are unextractable. New -mac and
8952 -macopt options to dgst utility.
8956 * New option -sigopt to dgst utility. Update dgst to use
8957 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8958 alternative signing parameters such as X9.31 or PSS in the dgst
8963 * Change ssl_cipher_apply_rule(), the internal function that does
8964 the work each time a ciphersuite string requests enabling
8965 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8966 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8967 the order of disabled ciphersuites such that those ciphersuites
8968 that most recently went from enabled to disabled not only stay
8969 in order with respect to each other, but also have higher priority
8970 than other disabled ciphersuites the next time ciphersuites are
8973 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8974 the same ciphersuites as with "HIGH" alone, but in a specific
8975 order where the PSK ciphersuites come first (since they are the
8976 most recently disabled ciphersuites when "HIGH" is parsed).
8978 Also, change ssl_create_cipher_list() (using this new
8979 functionality) such that between otherwise identical
8980 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8985 * Change ssl_create_cipher_list() so that it automatically
8986 arranges the ciphersuites in reasonable order before starting
8987 to process the rule string. Thus, the definition for "DEFAULT"
8988 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8989 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8990 This makes it much easier to arrive at a reasonable default order
8991 in applications for which anonymous ciphers are OK (meaning
8992 that you can't actually use DEFAULT).
8994 *Bodo Moeller; suggested by Victor Duchovni*
8996 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8997 processing) into multiple integers instead of setting
8998 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8999 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9000 (These masks as well as the individual bit definitions are hidden
9001 away into the non-exported interface ssl/ssl_locl.h, so this
9002 change to the definition of the SSL_CIPHER structure shouldn't
9003 affect applications.) This give us more bits for each of these
9004 categories, so there is no longer a need to coagulate AES128 and
9005 AES256 into a single algorithm bit, and to coagulate Camellia128
9006 and Camellia256 into a single algorithm bit, which has led to all
9009 Thus, among other things, the kludge introduced in 0.9.7m and
9010 0.9.8e for masking out AES256 independently of AES128 or masking
9011 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9013 With the change, we also introduce new ciphersuite aliases that
9014 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9019 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9020 Use the leftmost N bytes of the signature input if the input is
9021 larger than the prime q (with N being the size in bytes of q).
9025 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9026 it yet and it is largely untested.
9030 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9034 * Initial incomplete changes to avoid need for function casts in OpenSSL
9035 some compilers (gcc 4.2 and later) reject their use. Safestack is
9036 reimplemented. Update ASN1 to avoid use of legacy functions.
9040 * Win32/64 targets are linked with Winsock2.
9044 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9045 to external functions. This can be used to increase CRL handling
9046 efficiency especially when CRLs are very large by (for example) storing
9047 the CRL revoked certificates in a database.
9051 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9052 new CRLs added to a directory can be used. New command line option
9053 -verify_return_error to s_client and s_server. This causes real errors
9054 to be returned by the verify callback instead of carrying on no matter
9055 what. This reflects the way a "real world" verify callback would behave.
9059 * GOST engine, supporting several GOST algorithms and public key formats.
9060 Kindly donated by Cryptocom.
9064 * Partial support for Issuing Distribution Point CRL extension. CRLs
9065 partitioned by DP are handled but no indirect CRL or reason partitioning
9066 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9067 selected via a scoring technique which handles IDP and AKID in CRLs.
9071 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9072 will ultimately be used for all verify operations: this will remove the
9073 X509_STORE dependency on certificate verification and allow alternative
9074 lookup methods. X509_STORE based implementations of these two callbacks.
9078 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9079 Modify get_crl() to find a valid (unexpired) CRL if possible.
9083 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9084 this would be called X509_CRL_cmp() but that name is already used by
9085 a function that just compares CRL issuer names. Cache several CRL
9086 extensions in X509_CRL structure and cache CRLDP in X509.
9090 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9091 this maps equivalent X509_NAME structures into a consistent structure.
9092 Name comparison can then be performed rapidly using memcmp().
9096 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9101 * Allow digests to supply their own micalg string for S/MIME type using
9102 the ctrl EVP_MD_CTRL_MICALG.
9106 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9107 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9108 ctrl. It can then customise the structure before and/or after signing
9113 * New function OBJ_add_sigid() to allow application defined signature OIDs
9114 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9115 to free up any added signature OIDs.
9119 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9120 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9121 digest and cipher tables. New options added to openssl utility:
9122 list-message-digest-algorithms and list-cipher-algorithms.
9126 * Change the array representation of binary polynomials: the list
9127 of degrees of non-zero coefficients is now terminated with -1.
9128 Previously it was terminated with 0, which was also part of the
9129 value; thus, the array representation was not applicable to
9130 polynomials where t^0 has coefficient zero. This change makes
9131 the array representation useful in a more general context.
9135 * Various modifications and fixes to SSL/TLS cipher string
9136 handling. For ECC, the code now distinguishes between fixed ECDH
9137 with RSA certificates on the one hand and with ECDSA certificates
9138 on the other hand, since these are separate ciphersuites. The
9139 unused code for Fortezza ciphersuites has been removed.
9141 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9142 (not "ECDHE"). For consistency with the code for DH
9143 certificates, use of ECDH certificates is now considered ECDH
9144 authentication, not RSA or ECDSA authentication (the latter is
9145 merely the CA's signing algorithm and not actively used in the
9148 The temporary ciphersuite alias "ECCdraft" is no longer
9149 available, and ECC ciphersuites are no longer excluded from "ALL"
9150 and "DEFAULT". The following aliases now exist for RFC 4492
9151 ciphersuites, most of these by analogy with the DH case:
9153 kECDHr - ECDH cert, signed with RSA
9154 kECDHe - ECDH cert, signed with ECDSA
9155 kECDH - ECDH cert (signed with either RSA or ECDSA)
9156 kEECDH - ephemeral ECDH
9157 ECDH - ECDH cert or ephemeral ECDH
9163 AECDH - anonymous ECDH
9164 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9168 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9169 Use correct micalg parameters depending on digest(s) in signed message.
9173 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9174 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9178 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9179 an engine to register a method. Add ENGINE lookups for methods and
9180 functional reference processing.
9184 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9185 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9190 * New -resign option to smime utility. This adds one or more signers
9191 to an existing PKCS#7 signedData structure. Also -md option to use an
9192 alternative message digest algorithm for signing.
9196 * Tidy up PKCS#7 routines and add new functions to make it easier to
9197 create PKCS7 structures containing multiple signers. Update smime
9198 application to support multiple signers.
9202 * New -macalg option to pkcs12 utility to allow setting of an alternative
9207 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9208 Reorganize PBE internals to lookup from a static table using NIDs,
9209 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9210 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9211 PRF which will be automatically used with PBES2.
9215 * Replace the algorithm specific calls to generate keys in "req" with the
9220 * Update PKCS#7 enveloped data routines to use new API. This is now
9221 supported by any public key method supporting the encrypt operation. A
9222 ctrl is added to allow the public key algorithm to examine or modify
9223 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9228 * Add a ctrl to asn1 method to allow a public key algorithm to express
9229 a default digest type to use. In most cases this will be SHA1 but some
9230 algorithms (such as GOST) need to specify an alternative digest. The
9231 return value indicates how strong the preference is 1 means optional and
9232 2 is mandatory (that is it is the only supported type). Modify
9233 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9234 use the default md. Update openssl utilities to use the default digest
9235 type for signing if it is not explicitly indicated.
9239 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9240 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9241 signing method from the key type. This effectively removes the link
9242 between digests and public key types.
9246 * Add an OID cross reference table and utility functions. Its purpose is to
9247 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9248 rsaEncryption. This will allow some of the algorithm specific hackery
9249 needed to use the correct OID to be removed.
9253 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9254 structures for PKCS7_sign(). They are now set up by the relevant public
9259 * Add provisional EC pkey method with support for ECDSA and ECDH.
9263 * Add support for key derivation (agreement) in the API, DH method and
9268 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9269 public and private key formats. As a side effect these add additional
9270 command line functionality not previously available: DSA signatures can be
9271 generated and verified using pkeyutl and DH key support and generation in
9278 *Oliver Tappe <zooey@hirschkaefer.de>*
9280 * New make target "install_html_docs" installs HTML renditions of the
9283 *Oliver Tappe <zooey@hirschkaefer.de>*
9285 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9286 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9287 support key and parameter generation and add initial key generation
9288 functionality for RSA.
9292 * Add functions for main EVP_PKEY_method operations. The undocumented
9293 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9294 `EVP_PKEY_{encrypt,decrypt}_old`.
9298 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9299 key API, doesn't do much yet.
9303 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9304 public key algorithms. New option to openssl utility:
9305 "list-public-key-algorithms" to print out info.
9309 * Implement the Supported Elliptic Curves Extension for
9310 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9314 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9315 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9319 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9320 utilities such as rsa, dsa, dsaparam etc except they process any key
9325 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9326 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9327 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9332 * Initial support for pluggable public key ASN1.
9333 De-spaghettify the public key ASN1 handling. Move public and private
9334 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9335 algorithm specific handling to a single module within the relevant
9336 algorithm directory. Add functions to allow (near) opaque processing
9337 of public and private key structures.
9341 * Implement the Supported Point Formats Extension for
9342 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9346 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9347 for the psk identity [hint] and the psk callback functions to the
9348 SSL_SESSION, SSL and SSL_CTX structure.
9351 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9355 SSL_CTX_use_psk_identity_hint
9356 SSL_get_psk_identity_hint
9357 SSL_get_psk_identity
9358 SSL_use_psk_identity_hint
9360 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9362 * Add RFC 3161 compliant time stamp request creation, response generation
9363 and response verification functionality.
9365 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9367 * Add initial support for TLS extensions, specifically for the server_name
9368 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9369 have new members for a host name. The SSL data structure has an
9370 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9371 stored in that context to allow for session resumption, even after the
9372 SSL has been switched to a new SSL_CTX in reaction to a client's
9373 server_name extension.
9375 New functions (subject to change):
9377 SSL_get_servername()
9378 SSL_get_servername_type()
9381 New CTRL codes and macros (subject to change):
9383 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9384 - SSL_CTX_set_tlsext_servername_callback()
9385 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9386 - SSL_CTX_set_tlsext_servername_arg()
9387 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9389 openssl s_client has a new '-servername ...' option.
9391 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9392 '-key2 ...', '-servername_fatal' (subject to change). This allows
9393 testing the HostName extension for a specific single host name ('-cert'
9394 and '-key' remain fallbacks for handshakes without HostName
9395 negotiation). If the unrecognized_name alert has to be sent, this by
9396 default is a warning; it becomes fatal with the '-servername_fatal'
9399 *Peter Sylvester, Remy Allais, Christophe Renou*
9401 * Whirlpool hash implementation is added.
9405 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9406 bn(64,32). Because of instruction set limitations it doesn't have
9407 any negative impact on performance. This was done mostly in order
9408 to make it possible to share assembler modules, such as bn_mul_mont
9409 implementations, between 32- and 64-bit builds without hassle.
9413 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9414 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9419 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9420 dedicated Montgomery multiplication procedure, is introduced.
9421 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9422 "64-bit" performance on certain 32-bit targets.
9426 * New option SSL_OP_NO_COMP to disable use of compression selectively
9427 in SSL structures. New SSL ctrl to set maximum send fragment size.
9428 Save memory by setting the I/O buffer sizes dynamically instead of
9429 using the maximum available value.
9433 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9434 in addition to the text details.
9438 * Very, very preliminary EXPERIMENTAL support for printing of general
9439 ASN1 structures. This currently produces rather ugly output and doesn't
9440 handle several customised structures at all.
9444 * Integrated support for PVK file format and some related formats such
9445 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9446 these in the 'rsa' and 'dsa' utilities.
9450 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9454 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9455 place for the (very old) "NETSCAPE" format certificates which are now
9456 handled using new ASN1 code equivalents.
9460 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9461 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9462 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9466 * Modify CRL distribution points extension code to print out previously
9467 unsupported fields. Enhance extension setting code to allow setting of
9472 * Add print and set support for Issuing Distribution Point CRL extension.
9476 * Change 'Configure' script to enable Camellia by default.
9483 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9485 * When rejecting SSL/TLS records due to an incorrect version number, never
9486 update s->server with a new major version number. As of
9487 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9488 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9489 the previous behavior could result in a read attempt at NULL when
9490 receiving specific incorrect SSL/TLS records once record payload
9491 protection is active. ([CVE-2010-0740])
9493 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9495 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9496 could be crashed if the relevant tables were not present (e.g. chrooted).
9498 *Tomas Hoger <thoger@redhat.com>*
9500 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9502 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9504 *Martin Olsson, Neel Mehta*
9506 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9507 accommodate for stack sorting, always a write lock!).
9511 * On some versions of WIN32 Heap32Next is very slow. This can cause
9512 excessive delays in the RAND_poll(): over a minute. As a workaround
9513 include a time check in the inner Heap32Next loop too.
9517 * The code that handled flushing of data in SSL/TLS originally used the
9518 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9519 the problem outlined in PR#1949. The fix suggested there however can
9520 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9521 of Apache). So instead simplify the code to flush unconditionally.
9522 This should be fine since flushing with no data to flush is a no op.
9526 * Handle TLS versions 2.0 and later properly and correctly use the
9527 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9528 off ancient servers have a habit of sticking around for a while...
9532 * Modify compression code so it frees up structures without using the
9533 ex_data callbacks. This works around a problem where some applications
9534 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9535 restarting) then use compression (e.g. SSL with compression) later.
9536 This results in significant per-connection memory leaks and
9537 has caused some security issues including CVE-2008-1678 and
9542 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9543 change when encrypting or decrypting.
9547 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9548 connect and renegotiate with servers which do not support RI.
9549 Until RI is more widely deployed this option is enabled by default.
9553 * Add "missing" ssl ctrls to clear options and mode.
9557 * If client attempts to renegotiate and doesn't support RI respond with
9558 a no_renegotiation alert as required by RFC5746. Some renegotiating
9559 TLS clients will continue a connection gracefully when they receive
9560 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9561 waiting for a server hello which it will never receive. Now we treat a
9562 received no_renegotiation alert as a fatal error. This is because
9563 applications requesting a renegotiation might well expect it to succeed
9564 and would have no code in place to handle the server denying it so the
9565 only safe thing to do is to terminate the connection.
9569 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9570 peer supports secure renegotiation and 0 otherwise. Print out peer
9571 renegotiation support in s_client/s_server.
9575 * Replace the highly broken and deprecated SPKAC certification method with
9576 the updated NID creation version. This should correctly handle UTF8.
9580 * Implement RFC5746. Re-enable renegotiation but require the extension
9581 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9582 turns out to be a bad idea. It has been replaced by
9583 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9584 SSL_CTX_set_options(). This is really not recommended unless you
9585 know what you are doing.
9587 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9589 * Fixes to stateless session resumption handling. Use initial_ctx when
9590 issuing and attempting to decrypt tickets in case it has changed during
9591 servername handling. Use a non-zero length session ID when attempting
9592 stateless session resumption: this makes it possible to determine if
9593 a resumption has occurred immediately after receiving server hello
9594 (several places in OpenSSL subtly assume this) instead of later in
9599 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9600 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9601 fixes for a few places where the return code is not checked
9604 *Julia Lawall <julia@diku.dk>*
9606 * Add --strict-warnings option to Configure script to include devteam
9607 warnings in other configurations.
9611 * Add support for --libdir option and LIBDIR variable in makefiles. This
9612 makes it possible to install openssl libraries in locations which
9613 have names other than "lib", for example "/usr/lib64" which some
9616 *Steve Henson, based on patch from Jeremy Utley*
9618 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9619 X690 8.9.12 and can produce some misleading textual output of OIDs.
9621 *Steve Henson, reported by Dan Kaminsky*
9623 * Delete MD2 from algorithm tables. This follows the recommendation in
9624 several standards that it is not used in new applications due to
9625 several cryptographic weaknesses. For binary compatibility reasons
9626 the MD2 API is still compiled in by default.
9630 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9635 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9636 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9639 *Guenter <lists@gknw.net>*
9641 * Fix the server certificate chain building code to use X509_verify_cert(),
9642 it used to have an ad-hoc builder which was unable to cope with anything
9643 other than a simple chain.
9645 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9647 * Don't check self signed certificate signatures in X509_verify_cert()
9648 by default (a flag can override this): it just wastes time without
9649 adding any security. As a useful side effect self signed root CAs
9650 with non-FIPS digests are now usable in FIPS mode.
9654 * In dtls1_process_out_of_seq_message() the check if the current message
9655 is already buffered was missing. For every new message was memory
9656 allocated, allowing an attacker to perform an denial of service attack
9657 with sending out of seq handshake messages until there is no memory
9658 left. Additionally every future message was buffered, even if the
9659 sequence number made no sense and would be part of another handshake.
9660 So only messages with sequence numbers less than 10 in advance will be
9661 buffered. ([CVE-2009-1378])
9663 *Robin Seggelmann, discovered by Daniel Mentz*
9665 * Records are buffered if they arrive with a future epoch to be
9666 processed after finishing the corresponding handshake. There is
9667 currently no limitation to this buffer allowing an attacker to perform
9668 a DOS attack with sending records with future epochs until there is no
9669 memory left. This patch adds the pqueue_size() function to determine
9670 the size of a buffer and limits the record buffer to 100 entries.
9673 *Robin Seggelmann, discovered by Daniel Mentz*
9675 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9676 parent structure is freed. ([CVE-2009-1379])
9680 * Handle non-blocking I/O properly in SSL_shutdown() call.
9682 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9684 * Add `2.5.4.*` OIDs
9686 *Ilya O. <vrghost@gmail.com>*
9688 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9690 * Disable renegotiation completely - this fixes a severe security
9691 problem ([CVE-2009-3555]) at the cost of breaking all
9692 renegotiation. Renegotiation can be re-enabled by setting
9693 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9694 run-time. This is really not recommended unless you know what
9699 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9701 * Don't set val to NULL when freeing up structures, it is freed up by
9702 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9703 zeroing past the valid field. ([CVE-2009-0789])
9705 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9707 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9708 checked correctly. This would allow some invalid signed attributes to
9709 appear to verify correctly. ([CVE-2009-0591])
9711 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9713 * Reject UniversalString and BMPString types with invalid lengths. This
9714 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9715 a legal length. ([CVE-2009-0590])
9719 * Set S/MIME signing as the default purpose rather than setting it
9720 unconditionally. This allows applications to override it at the store
9725 * Permit restricted recursion of ASN1 strings. This is needed in practice
9726 to handle some structures.
9730 * Improve efficiency of mem_gets: don't search whole buffer each time
9733 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9735 * New -hex option for openssl rand.
9739 * Print out UTF8String and NumericString when parsing ASN1.
9743 * Support NumericString type for name components.
9747 * Allow CC in the environment to override the automatically chosen
9748 compiler. Note that nothing is done to ensure flags work with the
9753 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9755 * Properly check EVP_VerifyFinal() and similar return values
9758 *Ben Laurie, Bodo Moeller, Google Security Team*
9760 * Enable TLS extensions by default.
9764 * Allow the CHIL engine to be loaded, whether the application is
9765 multithreaded or not. (This does not release the developer from the
9766 obligation to set up the dynamic locking callbacks.)
9768 *Sander Temme <sander@temme.net>*
9770 * Use correct exit code if there is an error in dgst command.
9772 *Steve Henson; problem pointed out by Roland Dirlewanger*
9774 * Tweak Configure so that you need to say "experimental-jpake" to enable
9775 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9779 * Add experimental JPAKE support, including demo authentication in
9780 s_client and s_server.
9784 * Set the comparison function in v3_addr_canonize().
9786 *Rob Austein <sra@hactrn.net>*
9788 * Add support for XMPP STARTTLS in s_client.
9790 *Philip Paeps <philip@freebsd.org>*
9792 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9793 to ensure that even with this option, only ciphersuites in the
9794 server's preference list will be accepted. (Note that the option
9795 applies only when resuming a session, so the earlier behavior was
9796 just about the algorithm choice for symmetric cryptography.)
9800 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9802 * Fix NULL pointer dereference if a DTLS server received
9803 ChangeCipherSpec as first record ([CVE-2009-1386]).
9807 * Fix a state transition in s3_srvr.c and d1_srvr.c
9808 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9812 * The fix in 0.9.8c that supposedly got rid of unsafe
9813 double-checked locking was incomplete for RSA blinding,
9814 addressing just one layer of what turns out to have been
9815 doubly unsafe triple-checked locking.
9817 So now fix this for real by retiring the MONT_HELPER macro
9818 in crypto/rsa/rsa_eay.c.
9820 *Bodo Moeller; problem pointed out by Marius Schilder*
9822 * Various precautionary measures:
9824 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9826 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9827 (NB: This would require knowledge of the secret session ticket key
9828 to exploit, in which case you'd be SOL either way.)
9830 - Change bn_nist.c so that it will properly handle input BIGNUMs
9831 outside the expected range.
9833 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9836 *Neel Mehta, Bodo Moeller*
9838 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9839 the load fails. Useful for distros.
9841 *Ben Laurie and the FreeBSD team*
9843 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9847 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9851 * Expand ENGINE to support engine supplied SSL client certificate functions.
9853 This work was sponsored by Logica.
9857 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9858 keystores. Support for SSL/TLS client authentication too.
9859 Not compiled unless enable-capieng specified to Configure.
9861 This work was sponsored by Logica.
9865 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9866 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9867 attribute creation routines such as certificate requests and PKCS#12
9872 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9874 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9875 handshake which could lead to a client crash as found using the
9876 Codenomicon TLS test suite ([CVE-2008-1672])
9878 *Steve Henson, Mark Cox*
9880 * Fix double free in TLS server name extensions which could lead to
9881 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9885 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9887 Clear the error queue to ensure that error entries left from
9888 older function calls do not interfere with the correct operation.
9890 *Lutz Jaenicke, Erik de Castro Lopo*
9892 * Remove root CA certificates of commercial CAs:
9894 The OpenSSL project does not recommend any specific CA and does not
9895 have any policy with respect to including or excluding any CA.
9896 Therefore it does not make any sense to ship an arbitrary selection
9897 of root CA certificates with the OpenSSL software.
9901 * RSA OAEP patches to fix two separate invalid memory reads.
9902 The first one involves inputs when 'lzero' is greater than
9903 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9904 before the beginning of from). The second one involves inputs where
9905 the 'db' section contains nothing but zeroes (there is a one-byte
9906 invalid read after the end of 'db').
9908 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9910 * Partial backport from 0.9.9-dev:
9912 Introduce bn_mul_mont (dedicated Montgomery multiplication
9913 procedure) as a candidate for BIGNUM assembler implementation.
9914 While 0.9.9-dev uses assembler for various architectures, only
9915 x86_64 is available by default here in the 0.9.8 branch, and
9916 32-bit x86 is available through a compile-time setting.
9918 To try the 32-bit x86 assembler implementation, use Configure
9919 option "enable-montasm" (which exists only for this backport).
9921 As "enable-montasm" for 32-bit x86 disclaims code stability
9922 anyway, in this constellation we activate additional code
9923 backported from 0.9.9-dev for further performance improvements,
9924 namely BN_from_montgomery_word. (To enable this otherwise,
9925 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9927 *Andy Polyakov (backport partially by Bodo Moeller)*
9929 * Add TLS session ticket callback. This allows an application to set
9930 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9931 values. This is useful for key rollover for example where several key
9932 sets may exist with different names.
9936 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9937 This was broken until now in 0.9.8 releases, such that the only way
9938 a registered ENGINE could be used (assuming it initialises
9939 successfully on the host) was to explicitly set it as the default
9940 for the relevant algorithms. This is in contradiction with 0.9.7
9941 behaviour and the documentation. With this fix, when an ENGINE is
9942 registered into a given algorithm's table of implementations, the
9943 'uptodate' flag is reset so that auto-discovery will be used next
9944 time a new context for that algorithm attempts to select an
9947 *Ian Lister (tweaked by Geoff Thorpe)*
9949 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9950 implementation in the following ways:
9952 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9955 Lack of BER streaming support means one pass streaming processing is
9956 only supported if data is detached: setting the streaming flag is
9957 ignored for embedded content.
9959 CMS support is disabled by default and must be explicitly enabled
9960 with the enable-cms configuration option.
9964 * Update the GMP engine glue to do direct copies between BIGNUM and
9965 mpz_t when openssl and GMP use the same limb size. Otherwise the
9966 existing "conversion via a text string export" trick is still used.
9968 *Paul Sheer <paulsheer@gmail.com>*
9970 * Zlib compression BIO. This is a filter BIO which compressed and
9971 uncompresses any data passed through it.
9975 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9976 RFC3394 compatible AES key wrapping.
9980 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9981 sets string data without copying. X509_ALGOR_set0() and
9982 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9983 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9984 from an X509_ATTRIBUTE structure optionally checking it occurs only
9985 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9990 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9991 to get the expected BN_FLG_CONSTTIME behavior.
9993 *Bodo Moeller (Google)*
9997 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9998 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9999 - added some more tests to do_tests.pl
10000 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10001 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10002 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10003 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10004 - various changes to netware.pl to enable gcc-cross builds on Win32
10006 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10007 - various changes to fix missing prototype warnings
10008 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10009 - added AES, WHIRLPOOL and CPUID assembler code to build files
10010 - added missing AES assembler make rules to mk1mf.pl
10011 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10013 *Guenter Knauf <eflash@gmx.net>*
10015 * Implement certificate status request TLS extension defined in RFC3546.
10016 A client can set the appropriate parameters and receive the encoded
10017 OCSP response via a callback. A server can query the supplied parameters
10018 and set the encoded OCSP response in the callback. Add simplified examples
10019 to s_client and s_server.
10023 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10025 * Fix various bugs:
10026 + Binary incompatibility of ssl_ctx_st structure
10027 + DTLS interoperation with non-compliant servers
10028 + Don't call get_session_cb() without proposed session
10029 + Fix ia64 assembler code
10031 *Andy Polyakov, Steve Henson*
10033 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10035 * DTLS Handshake overhaul. There were longstanding issues with
10036 OpenSSL DTLS implementation, which were making it impossible for
10037 RFC 4347 compliant client to communicate with OpenSSL server.
10038 Unfortunately just fixing these incompatibilities would "cut off"
10039 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10040 server keeps tolerating non RFC compliant syntax. The opposite is
10041 not true, 0.9.8f client can not communicate with earlier server.
10042 This update even addresses CVE-2007-4995.
10046 * Changes to avoid need for function casts in OpenSSL: some compilers
10047 (gcc 4.2 and later) reject their use.
10048 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10051 * Add RFC4507 support to OpenSSL. This includes the corrections in
10052 RFC4507bis. The encrypted ticket format is an encrypted encoded
10053 SSL_SESSION structure, that way new session features are automatically
10056 If a client application caches session in an SSL_SESSION structure
10057 support is transparent because tickets are now stored in the encoded
10060 The SSL_CTX structure automatically generates keys for ticket
10061 protection in servers so again support should be possible
10062 with no application modification.
10064 If a client or server wishes to disable RFC4507 support then the option
10065 SSL_OP_NO_TICKET can be set.
10067 Add a TLS extension debugging callback to allow the contents of any client
10068 or server extensions to be examined.
10070 This work was sponsored by Google.
10074 * Add initial support for TLS extensions, specifically for the server_name
10075 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10076 have new members for a host name. The SSL data structure has an
10077 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10078 stored in that context to allow for session resumption, even after the
10079 SSL has been switched to a new SSL_CTX in reaction to a client's
10080 server_name extension.
10082 New functions (subject to change):
10084 SSL_get_servername()
10085 SSL_get_servername_type()
10088 New CTRL codes and macros (subject to change):
10090 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10091 - SSL_CTX_set_tlsext_servername_callback()
10092 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10093 - SSL_CTX_set_tlsext_servername_arg()
10094 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10096 openssl s_client has a new '-servername ...' option.
10098 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10099 '-key2 ...', '-servername_fatal' (subject to change). This allows
10100 testing the HostName extension for a specific single host name ('-cert'
10101 and '-key' remain fallbacks for handshakes without HostName
10102 negotiation). If the unrecognized_name alert has to be sent, this by
10103 default is a warning; it becomes fatal with the '-servername_fatal'
10106 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10108 * Add AES and SSE2 assembly language support to VC++ build.
10112 * Mitigate attack on final subtraction in Montgomery reduction.
10116 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10117 (which previously caused an internal error).
10121 * Squeeze another 10% out of IGE mode when in != out.
10125 * AES IGE mode speedup.
10127 *Dean Gaudet (Google)*
10129 * Add the Korean symmetric 128-bit cipher SEED (see
10130 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10131 add SEED ciphersuites from RFC 4162:
10133 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10134 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10135 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10136 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10138 To minimize changes between patchlevels in the OpenSSL 0.9.8
10139 series, SEED remains excluded from compilation unless OpenSSL
10140 is configured with 'enable-seed'.
10142 *KISA, Bodo Moeller*
10144 * Mitigate branch prediction attacks, which can be practical if a
10145 single processor is shared, allowing a spy process to extract
10146 information. For detailed background information, see
10147 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10148 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10149 and Necessary Software Countermeasures"). The core of the change
10150 are new versions BN_div_no_branch() and
10151 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10152 respectively, which are slower, but avoid the security-relevant
10153 conditional branches. These are automatically called by BN_div()
10154 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10155 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10156 remove a conditional branch.
10158 BN_FLG_CONSTTIME is the new name for the previous
10159 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10160 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10161 in the exponent causes BN_mod_exp_mont() to use the alternative
10162 implementation in BN_mod_exp_mont_consttime().) The old name
10163 remains as a deprecated alias.
10165 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10166 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10167 constant-time implementations for more than just exponentiation.
10168 Here too the old name is kept as a deprecated alias.
10170 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10171 the BN_BLINDING structure gets an independent copy of the
10172 modulus. This means that the previous `BIGNUM *m` argument to
10173 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10174 essentially becomes `const BIGNUM *m`, although we can't actually
10175 change this in the header file before 0.9.9. It allows
10176 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10177 enable BN_FLG_CONSTTIME.
10179 *Matthew D Wood (Intel Corp)*
10181 * In the SSL/TLS server implementation, be strict about session ID
10182 context matching (which matters if an application uses a single
10183 external cache for different purposes). Previously,
10184 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10185 set. This did ensure strict client verification, but meant that,
10186 with applications using a single external cache for quite
10187 different requirements, clients could circumvent ciphersuite
10188 restrictions for a given session ID context by starting a session
10189 in a different context.
10193 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10194 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10195 authentication-only ciphersuites.
10199 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10200 not complete and could lead to a possible single byte overflow
10201 ([CVE-2007-5135]) [Ben Laurie]
10203 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10205 * Since AES128 and AES256 (and similarly Camellia128 and
10206 Camellia256) share a single mask bit in the logic of
10207 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10208 kludge to work properly if AES128 is available and AES256 isn't
10209 (or if Camellia128 is available and Camellia256 isn't).
10213 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10214 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10215 When a point or a seed is encoded in a BIT STRING, we need to
10216 prevent the removal of trailing zero bits to get the proper DER
10217 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10218 of a NamedBitList, for which trailing 0 bits need to be removed.)
10222 * Have SSL/TLS server implementation tolerate "mismatched" record
10223 protocol version while receiving ClientHello even if the
10224 ClientHello is fragmented. (The server can't insist on the
10225 particular protocol version it has chosen before the ServerHello
10226 message has informed the client about his choice.)
10230 * Add RFC 3779 support.
10232 *Rob Austein for ARIN, Ben Laurie*
10234 * Load error codes if they are not already present instead of using a
10235 static variable. This allows them to be cleanly unloaded and reloaded.
10236 Improve header file function name parsing.
10240 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10241 or CAPABILITY handshake as required by RFCs.
10243 *Goetz Babin-Ebell*
10245 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10247 * Introduce limits to prevent malicious keys being able to
10248 cause a denial of service. ([CVE-2006-2940])
10250 *Steve Henson, Bodo Moeller*
10252 * Fix ASN.1 parsing of certain invalid structures that can result
10253 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10255 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10256 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10258 * Fix SSL client code which could crash if connecting to a
10259 malicious SSLv2 server. ([CVE-2006-4343])
10261 *Tavis Ormandy and Will Drewry, Google Security Team*
10263 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10264 match only those. Before that, "AES256-SHA" would be interpreted
10265 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10266 the same strength classification in 0.9.7h) as we currently only
10267 have a single AES bit in the ciphersuite description bitmap.
10268 That change, however, also applied to ciphersuite strings such as
10269 "RC4-MD5" that intentionally matched multiple ciphersuites --
10270 namely, SSL 2.0 ciphersuites in addition to the more common ones
10271 from SSL 3.0/TLS 1.0.
10273 So we change the selection algorithm again: Naming an explicit
10274 ciphersuite selects this one ciphersuite, and any other similar
10275 ciphersuite (same bitmap) from *other* protocol versions.
10276 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10277 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10279 Since SSL 2.0 does not have any ciphersuites for which the
10280 128/256 bit distinction would be relevant, this works for now.
10281 The proper fix will be to use different bits for AES128 and
10282 AES256, which would have avoided the problems from the beginning;
10283 however, bits are scarce, so we can only do this in a new release
10284 (not just a patchlevel) when we can change the SSL_CIPHER
10285 definition to split the single 'unsigned long mask' bitmap into
10286 multiple values to extend the available space.
10290 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10292 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10293 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10295 * Add AES IGE and biIGE modes.
10299 * Change the Unix randomness entropy gathering to use poll() when
10300 possible instead of select(), since the latter has some
10301 undesirable limitations.
10303 *Darryl Miles via Richard Levitte and Bodo Moeller*
10305 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10306 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10307 cannot be implicitly activated as part of, e.g., the "AES" alias.
10308 However, please upgrade to OpenSSL 0.9.9[-dev] for
10309 non-experimental use of the ECC ciphersuites to get TLS extension
10310 support, which is required for curve and point format negotiation
10311 to avoid potential handshake problems.
10315 * Disable rogue ciphersuites:
10317 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10318 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10319 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10321 The latter two were purportedly from
10322 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10325 Also deactivate the remaining ciphersuites from
10326 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10327 unofficial, and the ID has long expired.
10331 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10332 dual-core machines) and other potential thread-safety issues.
10336 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10337 versions), which is now available for royalty-free use
10338 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10339 Also, add Camellia TLS ciphersuites from RFC 4132.
10341 To minimize changes between patchlevels in the OpenSSL 0.9.8
10342 series, Camellia remains excluded from compilation unless OpenSSL
10343 is configured with 'enable-camellia'.
10347 * Disable the padding bug check when compression is in use. The padding
10348 bug check assumes the first packet is of even length, this is not
10349 necessarily true if compression is enabled and can result in false
10350 positives causing handshake failure. The actual bug test is ancient
10351 code so it is hoped that implementations will either have fixed it by
10352 now or any which still have the bug do not support compression.
10356 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10358 * When applying a cipher rule check to see if string match is an explicit
10359 cipher suite and only match that one cipher suite if it is.
10363 * Link in manifests for VC++ if needed.
10365 *Austin Ziegler <halostatue@gmail.com>*
10367 * Update support for ECC-based TLS ciphersuites according to
10368 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10369 TLS extensions, which are supported starting with the 0.9.9
10370 branch, not in the OpenSSL 0.9.8 branch).
10374 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10375 opaque EVP_CIPHER_CTX handling.
10379 * Fixes and enhancements to zlib compression code. We now only use
10380 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10381 to conform with the standards mentioned here:
10382 <http://www.zlib.net/DLL_FAQ.txt>
10383 Static zlib linking now works on Windows and the new --with-zlib-include
10384 --with-zlib-lib options to Configure can be used to supply the location
10385 of the headers and library. Gracefully handle case where zlib library
10390 * Several fixes and enhancements to the OID generation code. The old code
10391 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10392 handle numbers larger than ULONG_MAX, truncated printing and had a
10393 non standard OBJ_obj2txt() behaviour.
10397 * Add support for building of engines under engine/ as shared libraries
10398 under VC++ build system.
10402 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10403 Hopefully, we will not see any false combination of paths any more.
10407 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10409 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10410 (part of SSL_OP_ALL). This option used to disable the
10411 countermeasure against man-in-the-middle protocol-version
10412 rollback in the SSL 2.0 server implementation, which is a bad
10413 idea. ([CVE-2005-2969])
10415 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10416 for Information Security, National Institute of Advanced Industrial
10417 Science and Technology [AIST], Japan)*
10419 * Add two function to clear and return the verify parameter flags.
10423 * Keep cipherlists sorted in the source instead of sorting them at
10424 runtime, thus removing the need for a lock.
10428 * Avoid some small subgroup attacks in Diffie-Hellman.
10430 *Nick Mathewson and Ben Laurie*
10432 * Add functions for well-known primes.
10436 * Extended Windows CE support.
10438 *Satoshi Nakamura and Andy Polyakov*
10440 * Initialize SSL_METHOD structures at compile time instead of during
10441 runtime, thus removing the need for a lock.
10445 * Make PKCS7_decrypt() work even if no certificate is supplied by
10446 attempting to decrypt each encrypted key in turn. Add support to
10451 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10453 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10456 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10460 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10461 key into the same file any more.
10465 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10469 * Add -utf8 command line and config file option to 'ca'.
10471 *Stefan <stf@udoma.org*
10473 * Removed the macro des_crypt(), as it seems to conflict with some
10474 libraries. Use DES_crypt().
10478 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10479 involves renaming the source and generated shared-libs for
10480 both. The engines will accept the corrected or legacy ids
10481 ('ncipher' and '4758_cca' respectively) when binding. NB,
10482 this only applies when building 'shared'.
10484 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10486 * Add attribute functions to EVP_PKEY structure. Modify
10487 PKCS12_create() to recognize a CSP name attribute and
10488 use it. Make -CSP option work again in pkcs12 utility.
10492 * Add new functionality to the bn blinding code:
10493 - automatic re-creation of the BN_BLINDING parameters after
10494 a fixed number of uses (currently 32)
10495 - add new function for parameter creation
10496 - introduce flags to control the update behaviour of the
10497 BN_BLINDING parameters
10498 - hide BN_BLINDING structure
10499 Add a second BN_BLINDING slot to the RSA structure to improve
10500 performance when a single RSA object is shared among several
10505 * Add support for DTLS.
10507 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10509 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10510 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10514 * Remove buggy and incomplete DH cert support from
10515 ssl/ssl_rsa.c and ssl/s3_both.c
10519 * Use SHA-1 instead of MD5 as the default digest algorithm for
10520 the `apps/openssl` commands.
10524 * Compile clean with "-Wall -Wmissing-prototypes
10525 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10526 DEBUG_SAFESTACK must also be set.
10530 * Change ./Configure so that certain algorithms can be disabled by default.
10531 The new counterpiece to "no-xxx" is "enable-xxx".
10533 The patented RC5 and MDC2 algorithms will now be disabled unless
10534 "enable-rc5" and "enable-mdc2", respectively, are specified.
10536 (IDEA remains enabled despite being patented. This is because IDEA
10537 is frequently required for interoperability, and there is no license
10538 fee for non-commercial use. As before, "no-idea" can be used to
10539 avoid this algorithm.)
10543 * Add processing of proxy certificates (see RFC 3820). This work was
10544 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10545 EGEE (Enabling Grids for E-science in Europe).
10549 * RC4 performance overhaul on modern architectures/implementations, such
10550 as Intel P4, IA-64 and AMD64.
10554 * New utility extract-section.pl. This can be used specify an alternative
10555 section number in a pod file instead of having to treat each file as
10556 a separate case in Makefile. This can be done by adding two lines to the
10559 =for comment openssl_section:XXX
10561 The blank line is mandatory.
10565 * New arguments -certform, -keyform and -pass for s_client and s_server
10566 to allow alternative format key and certificate files and passphrase
10571 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10572 update associated structures and add various utility functions.
10574 Add new policy related verify parameters, include policy checking in
10575 standard verify code. Enhance 'smime' application with extra parameters
10576 to support policy checking and print out.
10580 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10581 Nehemiah processors. These extensions support AES encryption in hardware
10582 as well as RNG (though RNG support is currently disabled).
10584 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10586 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10590 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10592 *Andy Polyakov and a number of other people*
10594 * Improved PowerPC platform support. Most notably BIGNUM assembler
10595 implementation contributed by IBM.
10597 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10599 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10600 exponent rather than 'unsigned long'. There is a corresponding change to
10601 the new 'rsa_keygen' element of the RSA_METHOD structure.
10603 *Jelte Jansen, Geoff Thorpe*
10605 * Functionality for creating the initial serial number file is now
10606 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10608 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10609 number file to 1, which is bound to cause problems. To avoid
10610 the problems while respecting compatibility between different 0.9.7
10611 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10612 CA.pl for serial number initialization. With the new release 0.9.8,
10613 we can fix the problem directly in the 'ca' utility.)
10617 * Reduced header interdependencies by declaring more opaque objects in
10618 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10619 give fewer recursive includes, which could break lazy source code - so
10620 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10621 developers should define this symbol when building and using openssl to
10622 ensure they track the recommended behaviour, interfaces, [etc], but
10623 backwards-compatible behaviour prevails when this isn't defined.
10627 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10631 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10632 This will generate a random key of the appropriate length based on the
10633 cipher context. The EVP_CIPHER can provide its own random key generation
10634 routine to support keys of a specific form. This is used in the des and
10635 3des routines to generate a key of the correct parity. Update S/MIME
10636 code to use new functions and hence generate correct parity DES keys.
10637 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10638 valid (weak or incorrect parity).
10642 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10643 as looking them up. This is useful when the verified structure may contain
10644 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10645 present unless the new PKCS7_NO_CRL flag is asserted.
10649 * Extend ASN1 oid configuration module. It now additionally accepts the
10652 shortName = some long name, 1.2.3.4
10656 * Reimplemented the BN_CTX implementation. There is now no more static
10657 limitation on the number of variables it can handle nor the depth of the
10658 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10659 information can now expand as required, and rather than having a single
10660 static array of bignums, BN_CTX now uses a linked-list of such arrays
10661 allowing it to expand on demand whilst maintaining the usefulness of
10662 BN_CTX's "bundling".
10666 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10667 to allow all RSA operations to function using a single BN_CTX.
10671 * Preliminary support for certificate policy evaluation and checking. This
10672 is initially intended to pass the tests outlined in "Conformance Testing
10673 of Relying Party Client Certificate Path Processing Logic" v1.07.
10677 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10678 remained unused and not that useful. A variety of other little bignum
10679 tweaks and fixes have also been made continuing on from the audit (see
10684 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10685 associated ASN1, EVP and SSL functions and old ASN1 macros.
10689 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10690 and this should never fail. So the return value from the use of
10691 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10692 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10696 * BN_CTX_get() should return zero-valued bignums, providing the same
10697 initialised value as BN_new().
10699 *Geoff Thorpe, suggested by Ulf Möller*
10701 * Support for inhibitAnyPolicy certificate extension.
10705 * An audit of the BIGNUM code is underway, for which debugging code is
10706 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10707 is considered valid when processing BIGNUMs, and causes execution to
10708 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10709 further steps are taken to deliberately pollute unused data in BIGNUM
10710 structures to try and expose faulty code further on. For now, openssl will
10711 (in its default mode of operation) continue to tolerate the inconsistent
10712 forms that it has tolerated in the past, but authors and packagers should
10713 consider trying openssl and their own applications when compiled with
10714 these debugging symbols defined. It will help highlight potential bugs in
10715 their own code, and will improve the test coverage for OpenSSL itself. At
10716 some point, these tighter rules will become openssl's default to improve
10717 maintainability, though the assert()s and other overheads will remain only
10718 in debugging configurations. See bn.h for more details.
10720 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10722 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10723 that can only be obtained through BN_CTX_new() (which implicitly
10724 initialises it). The presence of this function only made it possible
10725 to overwrite an existing structure (and cause memory leaks).
10729 * Because of the callback-based approach for implementing LHASH as a
10730 template type, lh_insert() adds opaque objects to hash-tables and
10731 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10732 to clean up those corresponding objects before destroying the hash table
10733 (and losing the object pointers). So some over-zealous constifications in
10734 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10735 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10736 prototyped to have "const" restrictions on the object pointers they are
10737 given (and so aren't required to cast them away any more).
10741 * The tmdiff.h API was so ugly and minimal that our own timing utility
10742 (speed) prefers to use its own implementation. The two implementations
10743 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10744 its object type properly exposed (MS_TM) instead of casting to/from
10745 `char *`. This may still change yet if someone realises MS_TM and
10747 aren't necessarily the greatest nomenclatures - but this is what was used
10748 internally to the implementation so I've used that for now.
10752 * Ensure that deprecated functions do not get compiled when
10753 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10754 the self-tests were still using deprecated key-generation functions so
10755 these have been updated also.
10759 * Reorganise PKCS#7 code to separate the digest location functionality
10760 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10761 New function PKCS7_set_digest() to set the digest type for PKCS#7
10762 digestedData type. Add additional code to correctly generate the
10763 digestedData type and add support for this type in PKCS7 initialization
10768 * New function PKCS7_set0_type_other() this initializes a PKCS7
10769 structure of type "other".
10773 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10774 sure the loop does correctly stop and breaking ("division by zero")
10775 modulus operations are not performed. The (pre-generated) prime
10776 table crypto/bn/bn_prime.h was already correct, but it could not be
10777 re-generated on some platforms because of the "division by zero"
10778 situation in the script.
10780 *Ralf S. Engelschall*
10782 * Update support for ECC-based TLS ciphersuites according to
10783 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10784 SHA-1 now is only used for "small" curves (where the
10785 representation of a field element takes up to 24 bytes); for
10786 larger curves, the field element resulting from ECDH is directly
10787 used as premaster secret.
10789 *Douglas Stebila (Sun Microsystems Laboratories)*
10791 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10792 curve secp160r1 to the tests.
10794 *Douglas Stebila (Sun Microsystems Laboratories)*
10796 * Add the possibility to load symbols globally with DSO.
10798 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10800 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10801 control of the error stack.
10805 * Add support for STORE in ENGINE.
10809 * Add the STORE type. The intention is to provide a common interface
10810 to certificate and key stores, be they simple file-based stores, or
10811 HSM-type store, or LDAP stores, or...
10812 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10816 * Add a generic structure called OPENSSL_ITEM. This can be used to
10817 pass a list of arguments to any function as well as provide a way
10818 for a function to pass data back to the caller.
10822 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10823 works like BUF_strdup() but can be used to duplicate a portion of
10824 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10829 * Add the function sk_find_ex() which works like sk_find(), but will
10830 return an index to an element even if an exact match couldn't be
10831 found. The index is guaranteed to point at the element where the
10832 searched-for key would be inserted to preserve sorting order.
10836 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10837 takes an extra flags argument for optional functionality. Currently,
10838 the following flags are defined:
10840 OBJ_BSEARCH_VALUE_ON_NOMATCH
10841 This one gets OBJ_bsearch_ex() to return a pointer to the first
10842 element where the comparing function returns a negative or zero
10845 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10846 This one gets OBJ_bsearch_ex() to return a pointer to the first
10847 element where the comparing function returns zero. This is useful
10848 if there are more than one element where the comparing function
10853 * Make it possible to create self-signed certificates with 'openssl ca'
10854 in such a way that the self-signed certificate becomes part of the
10855 CA database and uses the same mechanisms for serial number generation
10856 as all other certificate signing. The new flag '-selfsign' enables
10857 this functionality. Adapt CA.sh and CA.pl.in.
10861 * Add functionality to check the public key of a certificate request
10862 against a given private. This is useful to check that a certificate
10863 request can be signed by that key (self-signing).
10867 * Make it possible to have multiple active certificates with the same
10868 subject in the CA index file. This is done only if the keyword
10869 'unique_subject' is set to 'no' in the main CA section (default
10870 if 'CA_default') of the configuration file. The value is saved
10871 with the database itself in a separate index attribute file,
10872 named like the index file with '.attr' appended to the name.
10876 * Generate multi-valued AVAs using '+' notation in config files for
10881 * Support for nameConstraints certificate extension.
10885 * Support for policyConstraints certificate extension.
10889 * Support for policyMappings certificate extension.
10893 * Make sure the default DSA_METHOD implementation only uses its
10894 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10895 and change its own handlers to be NULL so as to remove unnecessary
10896 indirection. This lets alternative implementations fallback to the
10897 default implementation more easily.
10901 * Support for directoryName in GeneralName related extensions
10906 * Make it possible to link applications using Makefile.shared.
10907 Make that possible even when linking against static libraries!
10911 * Support for single pass processing for S/MIME signing. This now
10912 means that S/MIME signing can be done from a pipe, in addition
10913 cleartext signing (multipart/signed type) is effectively streaming
10914 and the signed data does not need to be all held in memory.
10916 This is done with a new flag PKCS7_STREAM. When this flag is set
10917 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10918 is done after the data is output (and digests calculated) in
10919 SMIME_write_PKCS7().
10923 * Add full support for -rpath/-R, both in shared libraries and
10924 applications, at least on the platforms where it's known how
10929 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10930 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10931 will now compute a table of multiples of the generator that
10932 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10933 faster (notably in the case of a single point multiplication,
10934 scalar * generator).
10936 *Nils Larsch, Bodo Moeller*
10938 * IPv6 support for certificate extensions. The various extensions
10939 which use the IP:a.b.c.d can now take IPv6 addresses using the
10940 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10945 * Added an ENGINE that implements RSA by performing private key
10946 exponentiations with the GMP library. The conversions to and from
10947 GMP's mpz_t format aren't optimised nor are any montgomery forms
10948 cached, and on x86 it appears OpenSSL's own performance has caught up.
10949 However there are likely to be other architectures where GMP could
10950 provide a boost. This ENGINE is not built in by default, but it can be
10951 specified at Configure time and should be accompanied by the necessary
10952 linker additions, eg;
10953 ./config -DOPENSSL_USE_GMP -lgmp
10957 * "openssl engine" will not display ENGINE/DSO load failure errors when
10958 testing availability of engines with "-t" - the old behaviour is
10959 produced by increasing the feature's verbosity with "-tt".
10963 * ECDSA routines: under certain error conditions uninitialized BN objects
10964 could be freed. Solution: make sure initialization is performed early
10965 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10970 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10971 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10972 software implementations. For DSA and DH, parameter generation can
10973 also be overridden by providing the appropriate method callbacks.
10977 * Change the "progress" mechanism used in key-generation and
10978 primality testing to functions that take a new BN_GENCB pointer in
10979 place of callback/argument pairs. The new API functions have `_ex`
10980 postfixes and the older functions are reimplemented as wrappers for
10981 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10982 declarations of the old functions to help (graceful) attempts to
10983 migrate to the new functions. Also, the new key-generation API
10984 functions operate on a caller-supplied key-structure and return
10985 success/failure rather than returning a key or NULL - this is to
10986 help make "keygen" another member function of RSA_METHOD etc.
10988 Example for using the new callback interface:
10990 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10991 void *my_arg = ...;
10994 BN_GENCB_set(&my_cb, my_callback, my_arg);
10996 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10997 /* For the meaning of a, b in calls to my_callback(), see the
10998 * documentation of the function that calls the callback.
10999 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11000 * my_callback should return 1 if it wants BN_is_prime_ex()
11001 * to continue, or 0 to stop.
11006 * Change the ZLIB compression method to be stateful, and make it
11007 available to TLS with the number defined in
11008 draft-ietf-tls-compression-04.txt.
11012 * Add the ASN.1 structures and functions for CertificatePair, which
11013 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11015 CertificatePair ::= SEQUENCE {
11016 forward [0] Certificate OPTIONAL,
11017 reverse [1] Certificate OPTIONAL,
11018 -- at least one of the pair shall be present -- }
11020 Also implement the PEM functions to read and write certificate
11021 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11023 This needed to be defined, mostly for the sake of the LDAP
11024 attribute crossCertificatePair, but may prove useful elsewhere as
11029 * Make it possible to inhibit symlinking of shared libraries in
11030 Makefile.shared, for Cygwin's sake.
11034 * Extend the BIGNUM API by creating a function
11035 void BN_set_negative(BIGNUM *a, int neg);
11036 and a macro that behave like
11037 int BN_is_negative(const BIGNUM *a);
11039 to avoid the need to access 'a->neg' directly in applications.
11043 * Implement fast modular reduction for pseudo-Mersenne primes
11044 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11045 EC_GROUP_new_curve_GFp() will now automatically use this
11048 *Nils Larsch <nla@trustcenter.de>*
11050 * Add new lock type (CRYPTO_LOCK_BN).
11054 * Change the ENGINE framework to automatically load engines
11055 dynamically from specific directories unless they could be
11056 found to already be built in or loaded. Move all the
11057 current engines except for the cryptodev one to a new
11058 directory engines/.
11059 The engines in engines/ are built as shared libraries if
11060 the "shared" options was given to ./Configure or ./config.
11061 Otherwise, they are inserted in libcrypto.a.
11062 /usr/local/ssl/engines is the default directory for dynamic
11063 engines, but that can be overridden at configure time through
11064 the usual use of --prefix and/or --openssldir, and at run
11065 time with the environment variable OPENSSL_ENGINES.
11067 *Geoff Thorpe and Richard Levitte*
11069 * Add Makefile.shared, a helper makefile to build shared
11070 libraries. Adapt Makefile.org.
11074 * Add version info to Win32 DLLs.
11076 *Peter 'Luna' Runestig" <peter@runestig.com>*
11078 * Add new 'medium level' PKCS#12 API. Certificates and keys
11079 can be added using this API to created arbitrary PKCS#12
11080 files while avoiding the low-level API.
11082 New options to PKCS12_create(), key or cert can be NULL and
11083 will then be omitted from the output file. The encryption
11084 algorithm NIDs can be set to -1 for no encryption, the mac
11085 iteration count can be set to 0 to omit the mac.
11087 Enhance pkcs12 utility by making the -nokeys and -nocerts
11088 options work when creating a PKCS#12 file. New option -nomac
11089 to omit the mac, NONE can be set for an encryption algorithm.
11090 New code is modified to use the enhanced PKCS12_create()
11091 instead of the low-level API.
11095 * Extend ASN1 encoder to support indefinite length constructed
11096 encoding. This can output sequences tags and octet strings in
11097 this form. Modify pk7_asn1.c to support indefinite length
11098 encoding. This is experimental and needs additional code to
11099 be useful, such as an ASN1 bio and some enhanced streaming
11102 Extend template encode functionality so that tagging is passed
11103 down to the template encoder.
11107 * Let 'openssl req' fail if an argument to '-newkey' is not
11108 recognized instead of using RSA as a default.
11112 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11113 As these are not official, they are not included in "ALL";
11114 the "ECCdraft" ciphersuite group alias can be used to select them.
11116 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11118 * Add ECDH engine support.
11120 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11122 * Add ECDH in new directory crypto/ecdh/.
11124 *Douglas Stebila (Sun Microsystems Laboratories)*
11126 * Let BN_rand_range() abort with an error after 100 iterations
11127 without success (which indicates a broken PRNG).
11131 * Change BN_mod_sqrt() so that it verifies that the input value
11132 is really the square of the return value. (Previously,
11133 BN_mod_sqrt would show GIGO behaviour.)
11137 * Add named elliptic curves over binary fields from X9.62, SECG,
11138 and WAP/WTLS; add OIDs that were still missing.
11140 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11142 * Extend the EC library for elliptic curves over binary fields
11143 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11146 EC_GF2m_simple_method
11150 EC_GROUP_new_curve_GF2m
11151 EC_GROUP_set_curve_GF2m
11152 EC_GROUP_get_curve_GF2m
11153 EC_POINT_set_affine_coordinates_GF2m
11154 EC_POINT_get_affine_coordinates_GF2m
11155 EC_POINT_set_compressed_coordinates_GF2m
11157 Point compression for binary fields is disabled by default for
11158 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11161 As binary polynomials are represented as BIGNUMs, various members
11162 of the EC_GROUP and EC_POINT data structures can be shared
11163 between the implementations for prime fields and binary fields;
11164 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11165 are essentially identical to their `..._GFp` counterparts.
11166 (For simplicity, the `..._GFp` prefix has been dropped from
11167 various internal method names.)
11169 An internal 'field_div' method (similar to 'field_mul' and
11170 'field_sqr') has been added; this is used only for binary fields.
11172 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11174 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11175 through methods ('mul', 'precompute_mult').
11177 The generic implementations (now internally called 'ec_wNAF_mul'
11178 and 'ec_wNAF_precomputed_mult') remain the default if these
11179 methods are undefined.
11181 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11183 * New function EC_GROUP_get_degree, which is defined through
11184 EC_METHOD. For curves over prime fields, this returns the bit
11185 length of the modulus.
11187 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11189 * New functions EC_GROUP_dup, EC_POINT_dup.
11190 (These simply call ..._new and ..._copy).
11192 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11194 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11195 Polynomials are represented as BIGNUMs (where the sign bit is not
11196 used) in the following functions [macros]:
11199 BN_GF2m_sub [= BN_GF2m_add]
11200 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11201 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11202 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11204 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11205 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11206 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11207 BN_GF2m_cmp [= BN_ucmp]
11209 (Note that only the 'mod' functions are actually for fields GF(2^m).
11210 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11212 For some functions, an the irreducible polynomial defining a
11213 field can be given as an 'unsigned int[]' with strictly
11214 decreasing elements giving the indices of those bits that are set;
11215 i.e., p[] represents the polynomial
11216 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11218 p[0] > p[1] > ... > p[k] = 0.
11219 This applies to the following functions:
11222 BN_GF2m_mod_mul_arr
11223 BN_GF2m_mod_sqr_arr
11224 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11225 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11226 BN_GF2m_mod_exp_arr
11227 BN_GF2m_mod_sqrt_arr
11228 BN_GF2m_mod_solve_quad_arr
11232 Conversion can be performed by the following functions:
11237 bntest.c has additional tests for binary polynomial arithmetic.
11239 Two implementations for BN_GF2m_mod_div() are available.
11240 The default algorithm simply uses BN_GF2m_mod_inv() and
11241 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11242 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11243 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11245 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11247 * Add new error code 'ERR_R_DISABLED' that can be used when some
11248 functionality is disabled at compile-time.
11250 *Douglas Stebila <douglas.stebila@sun.com>*
11252 * Change default behaviour of 'openssl asn1parse' so that more
11253 information is visible when viewing, e.g., a certificate:
11255 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11256 mode the content of non-printable OCTET STRINGs is output in a
11257 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11258 avoid the appearance of a printable string.
11260 *Nils Larsch <nla@trustcenter.de>*
11262 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11264 EC_GROUP_set_asn1_flag()
11265 EC_GROUP_get_asn1_flag()
11266 EC_GROUP_set_point_conversion_form()
11267 EC_GROUP_get_point_conversion_form()
11268 These control ASN1 encoding details:
11269 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11270 has been set to OPENSSL_EC_NAMED_CURVE.
11271 - Points are encoded in uncompressed form by default; options for
11272 asn1_for are as for point2oct, namely
11273 POINT_CONVERSION_COMPRESSED
11274 POINT_CONVERSION_UNCOMPRESSED
11275 POINT_CONVERSION_HYBRID
11277 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11279 EC_GROUP_set_seed()
11280 EC_GROUP_get0_seed()
11281 EC_GROUP_get_seed_len()
11282 This is used only for ASN1 purposes (so far).
11284 *Nils Larsch <nla@trustcenter.de>*
11286 * Add 'field_type' member to EC_METHOD, which holds the NID
11287 of the appropriate field type OID. The new function
11288 EC_METHOD_get_field_type() returns this value.
11290 *Nils Larsch <nla@trustcenter.de>*
11293 EC_POINT_point2bn()
11294 EC_POINT_bn2point()
11295 EC_POINT_point2hex()
11296 EC_POINT_hex2point()
11297 providing useful interfaces to EC_POINT_point2oct() and
11298 EC_POINT_oct2point().
11300 *Nils Larsch <nla@trustcenter.de>*
11302 * Change internals of the EC library so that the functions
11303 EC_GROUP_set_generator()
11304 EC_GROUP_get_generator()
11305 EC_GROUP_get_order()
11306 EC_GROUP_get_cofactor()
11307 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11308 to methods, which would lead to unnecessary code duplication when
11309 adding different types of curves.
11311 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11313 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11314 arithmetic, and such that modified wNAFs are generated
11315 (which avoid length expansion in many cases).
11319 * Add a function EC_GROUP_check_discriminant() (defined via
11320 EC_METHOD) that verifies that the curve discriminant is non-zero.
11322 Add a function EC_GROUP_check() that makes some sanity tests
11323 on a EC_GROUP, its generator and order. This includes
11324 EC_GROUP_check_discriminant().
11326 *Nils Larsch <nla@trustcenter.de>*
11328 * Add ECDSA in new directory crypto/ecdsa/.
11330 Add applications 'openssl ecparam' and 'openssl ecdsa'
11331 (these are based on 'openssl dsaparam' and 'openssl dsa').
11333 ECDSA support is also included in various other files across the
11334 library. Most notably,
11335 - 'openssl req' now has a '-newkey ecdsa:file' option;
11336 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11337 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11338 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11339 them suitable for ECDSA where domain parameters must be
11340 extracted before the specific public key;
11341 - ECDSA engine support has been added.
11343 *Nils Larsch <nla@trustcenter.de>*
11345 * Include some named elliptic curves, and add OIDs from X9.62,
11346 SECG, and WAP/WTLS. Each curve can be obtained from the new
11348 EC_GROUP_new_by_curve_name(),
11349 and the list of available named curves can be obtained with
11350 EC_get_builtin_curves().
11351 Also add a 'curve_name' member to EC_GROUP objects, which can be
11353 EC_GROUP_set_curve_name()
11354 EC_GROUP_get_curve_name()
11356 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11358 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11359 was actually never needed) and in BN_mul(). The removal in BN_mul()
11360 required a small change in bn_mul_part_recursive() and the addition
11361 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11362 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11363 bn_sub_words() and bn_add_words() except they take arrays with
11368 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11370 * Cleanse PEM buffers before freeing them since they may contain
11373 *Benjamin Bennett <ben@psc.edu>*
11375 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11376 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11377 authentication-only ciphersuites.
11381 * Since AES128 and AES256 share a single mask bit in the logic of
11382 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11383 kludge to work properly if AES128 is available and AES256 isn't.
11387 * Expand security boundary to match 1.1.1 module.
11391 * Remove redundant features: hash file source, editing of test vectors
11392 modify fipsld to use external fips_premain.c signature.
11396 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11397 run algorithm test programs.
11401 * Make algorithm test programs more tolerant of whitespace.
11405 * Have SSL/TLS server implementation tolerate "mismatched" record
11406 protocol version while receiving ClientHello even if the
11407 ClientHello is fragmented. (The server can't insist on the
11408 particular protocol version it has chosen before the ServerHello
11409 message has informed the client about his choice.)
11413 * Load error codes if they are not already present instead of using a
11414 static variable. This allows them to be cleanly unloaded and reloaded.
11418 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11420 * Introduce limits to prevent malicious keys being able to
11421 cause a denial of service. ([CVE-2006-2940])
11423 *Steve Henson, Bodo Moeller*
11425 * Fix ASN.1 parsing of certain invalid structures that can result
11426 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11428 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11429 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11431 * Fix SSL client code which could crash if connecting to a
11432 malicious SSLv2 server. ([CVE-2006-4343])
11434 *Tavis Ormandy and Will Drewry, Google Security Team*
11436 * Change ciphersuite string processing so that an explicit
11437 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11438 will no longer include "AES128-SHA"), and any other similar
11439 ciphersuite (same bitmap) from *other* protocol versions (so that
11440 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11441 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11442 changes from 0.9.8b and 0.9.8d.
11446 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11448 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11449 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11451 * Change the Unix randomness entropy gathering to use poll() when
11452 possible instead of select(), since the latter has some
11453 undesirable limitations.
11455 *Darryl Miles via Richard Levitte and Bodo Moeller*
11457 * Disable rogue ciphersuites:
11459 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11460 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11461 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11463 The latter two were purportedly from
11464 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11467 Also deactivate the remaining ciphersuites from
11468 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11469 unofficial, and the ID has long expired.
11473 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11474 dual-core machines) and other potential thread-safety issues.
11478 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11480 * Adapt fipsld and the build system to link against the validated FIPS
11481 module in FIPS mode.
11485 * Fixes for VC++ 2005 build under Windows.
11489 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11490 from a Windows bash shell such as MSYS. It is autodetected from the
11491 "config" script when run from a VC++ environment. Modify standard VC++
11492 build to use fipscanister.o from the GNU make build.
11496 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11498 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11499 The value now differs depending on if you build for FIPS or not.
11500 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11501 safely run with a non-FIPSed libcrypto, as it may crash because of
11502 the difference induced by this change.
11506 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11508 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11509 (part of SSL_OP_ALL). This option used to disable the
11510 countermeasure against man-in-the-middle protocol-version
11511 rollback in the SSL 2.0 server implementation, which is a bad
11512 idea. ([CVE-2005-2969])
11514 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11515 for Information Security, National Institute of Advanced Industrial
11516 Science and Technology [AIST, Japan)]*
11518 * Minimal support for X9.31 signatures and PSS padding modes. This is
11519 mainly for FIPS compliance and not fully integrated at this stage.
11523 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11524 the exponentiation using a fixed-length exponent. (Otherwise,
11525 the information leaked through timing could expose the secret key
11526 after many signatures; cf. Bleichenbacher's attack on DSA with
11531 * Make a new fixed-window mod_exp implementation the default for
11532 RSA, DSA, and DH private-key operations so that the sequence of
11533 squares and multiplies and the memory access pattern are
11534 independent of the particular secret key. This will mitigate
11535 cache-timing and potential related attacks.
11537 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11538 and this is automatically used by BN_mod_exp_mont() if the new flag
11539 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11540 will use this BN flag for private exponents unless the flag
11541 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11542 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11544 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11546 * Change the client implementation for SSLv23_method() and
11547 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11548 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11549 (Previously, the SSL 2.0 backwards compatible Client Hello
11550 message format would be used even with SSL_OP_NO_SSLv2.)
11554 * Add support for smime-type MIME parameter in S/MIME messages which some
11559 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11560 a threadsafe manner. Modify rsa code to use new function and add calls
11561 to dsa and dh code (which had race conditions before).
11565 * Include the fixed error library code in the C error file definitions
11566 instead of fixing them up at runtime. This keeps the error code
11567 structures constant.
11571 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11573 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11576 * Fixes for newer kerberos headers. NB: the casts are needed because
11577 the 'length' field is signed on one version and unsigned on another
11578 with no (?) obvious way to tell the difference, without these VC++
11579 complains. Also the "definition" of FAR (blank) is no longer included
11580 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11581 some needed definitions.
11585 * Undo Cygwin change.
11589 * Added support for proxy certificates according to RFC 3820.
11590 Because they may be a security thread to unaware applications,
11591 they must be explicitly allowed in run-time. See
11592 docs/HOWTO/proxy_certificates.txt for further information.
11596 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11598 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11599 server and client random values. Previously
11600 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11601 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11603 This change has negligible security impact because:
11605 1. Server and client random values still have 24 bytes of pseudo random
11608 2. Server and client random values are sent in the clear in the initial
11611 3. The master secret is derived using the premaster secret (48 bytes in
11612 size for static RSA ciphersuites) as well as client server and random
11615 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11618 *Stephen Henson, reported by UK NISCC*
11620 * Use Windows randomness collection on Cygwin.
11624 * Fix hang in EGD/PRNGD query when communication socket is closed
11625 prematurely by EGD/PRNGD.
11627 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11629 * Prompt for pass phrases when appropriate for PKCS12 input format.
11633 * Back-port of selected performance improvements from development
11634 branch, as well as improved support for PowerPC platforms.
11638 * Add lots of checks for memory allocation failure, error codes to indicate
11639 failure and freeing up memory if a failure occurs.
11641 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11643 * Add new -passin argument to dgst.
11647 * Perform some character comparisons of different types in X509_NAME_cmp:
11648 this is needed for some certificates that re-encode DNs into UTF8Strings
11649 (in violation of RFC3280) and can't or won't issue name rollover
11654 * Make an explicit check during certificate validation to see that
11655 the CA setting in each certificate on the chain is correct. As a
11656 side effect always do the following basic checks on extensions,
11657 not just when there's an associated purpose to the check:
11659 - if there is an unhandled critical extension (unless the user
11660 has chosen to ignore this fault)
11661 - if the path length has been exceeded (if one is set at all)
11662 - that certain extensions fit the associated purpose (if one has
11667 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11669 * Avoid a race condition when CRLs are checked in a multi threaded
11670 environment. This would happen due to the reordering of the revoked
11671 entries during signature checking and serial number lookup. Now the
11672 encoding is cached and the serial number sort performed under a lock.
11673 Add new STACK function sk_is_sorted().
11677 * Add Delta CRL to the extension code.
11681 * Various fixes to s3_pkt.c so alerts are sent properly.
11683 *David Holmes <d.holmes@f5.com>*
11685 * Reduce the chances of duplicate issuer name and serial numbers (in
11686 violation of RFC3280) using the OpenSSL certificate creation utilities.
11687 This is done by creating a random 64 bit value for the initial serial
11688 number when a serial number file is created or when a self signed
11689 certificate is created using 'openssl req -x509'. The initial serial
11690 number file is created using 'openssl x509 -next_serial' in CA.pl
11691 rather than being initialized to 1.
11695 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11697 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11698 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11700 *Joe Orton, Steve Henson*
11702 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11705 *Joe Orton, Steve Henson*
11707 * Make it possible to have multiple active certificates with the same
11708 subject in the CA index file. This is done only if the keyword
11709 'unique_subject' is set to 'no' in the main CA section (default
11710 if 'CA_default') of the configuration file. The value is saved
11711 with the database itself in a separate index attribute file,
11712 named like the index file with '.attr' appended to the name.
11716 * X509 verify fixes. Disable broken certificate workarounds when
11717 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11718 keyUsage extension present. Don't accept CRLs with unhandled critical
11719 extensions: since verify currently doesn't process CRL extensions this
11720 rejects a CRL with *any* critical extensions. Add new verify error codes
11725 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11726 A clarification of RFC2560 will require the use of OCTET STRINGs and
11727 some implementations cannot handle the current raw format. Since OpenSSL
11728 copies and compares OCSP nonces as opaque blobs without any attempt at
11729 parsing them this should not create any compatibility issues.
11733 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11734 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11735 this HMAC (and other) operations are several times slower than OpenSSL
11740 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11742 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11744 * Use the correct content when signing type "other".
11748 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11750 * Fix various bugs revealed by running the NISCC test suite:
11752 Stop out of bounds reads in the ASN1 code when presented with
11753 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11755 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11757 If verify callback ignores invalid public key errors don't try to check
11758 certificate signature with the NULL public key.
11762 * New -ignore_err option in ocsp application to stop the server
11763 exiting on the first error in a request.
11767 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11768 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11773 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11774 extra data after the compression methods not only for TLS 1.0
11775 but also for SSL 3.0 (as required by the specification).
11777 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11779 * Change X509_certificate_type() to mark the key as exported/exportable
11780 when it's 512 *bits* long, not 512 bytes.
11784 * Change AES_cbc_encrypt() so it outputs exact multiple of
11785 blocks during encryption.
11789 * Various fixes to base64 BIO and non blocking I/O. On write
11790 flushes were not handled properly if the BIO retried. On read
11791 data was not being buffered properly and had various logic bugs.
11792 This also affects blocking I/O when the data being decoded is a
11797 * Various S/MIME bugfixes and compatibility changes:
11798 output correct application/pkcs7 MIME type if
11799 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11800 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11801 of files as .eml work). Correctly handle very long lines in MIME
11806 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11808 * Countermeasure against the Klima-Pokorny-Rosa extension of
11809 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11810 a protocol version number mismatch like a decryption error
11811 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11815 * Turn on RSA blinding by default in the default implementation
11816 to avoid a timing attack. Applications that don't want it can call
11817 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11818 They would be ill-advised to do so in most cases.
11820 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11822 * Change RSA blinding code so that it works when the PRNG is not
11823 seeded (in this case, the secret RSA exponent is abused as
11824 an unpredictable seed -- if it is not unpredictable, there
11825 is no point in blinding anyway). Make RSA blinding thread-safe
11826 by remembering the creator's thread ID in rsa->blinding and
11827 having all other threads use local one-time blinding factors
11828 (this requires more computation than sharing rsa->blinding, but
11829 avoids excessive locking; and if an RSA object is not shared
11830 between threads, blinding will still be very fast).
11834 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11835 ENGINE as defaults for all supported algorithms irrespective of
11836 the 'flags' parameter. 'flags' is now honoured, so applications
11837 should make sure they are passing it correctly.
11841 * Target "mingw" now allows native Windows code to be generated in
11842 the Cygwin environment as well as with the MinGW compiler.
11846 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11848 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11849 via timing by performing a MAC computation even if incorrect
11850 block cipher padding has been found. This is a countermeasure
11851 against active attacks where the attacker has to distinguish
11852 between bad padding and a MAC verification error. ([CVE-2003-0078])
11854 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11855 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11856 Martin Vuagnoux (EPFL, Ilion)*
11858 * Make the no-err option work as intended. The intention with no-err
11859 is not to have the whole error stack handling routines removed from
11860 libcrypto, it's only intended to remove all the function name and
11861 reason texts, thereby removing some of the footprint that may not
11862 be interesting if those errors aren't displayed anyway.
11864 NOTE: it's still possible for any application or module to have its
11865 own set of error texts inserted. The routines are there, just not
11866 used by default when no-err is given.
11870 * Add support for FreeBSD on IA64.
11872 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11874 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11875 Kerberos function mit_des_cbc_cksum(). Before this change,
11876 the value returned by DES_cbc_cksum() was like the one from
11877 mit_des_cbc_cksum(), except the bytes were swapped.
11879 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11881 * Allow an application to disable the automatic SSL chain building.
11882 Before this a rather primitive chain build was always performed in
11883 ssl3_output_cert_chain(): an application had no way to send the
11884 correct chain if the automatic operation produced an incorrect result.
11886 Now the chain builder is disabled if either:
11888 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11890 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11892 The reasoning behind this is that an application would not want the
11893 auto chain building to take place if extra chain certificates are
11894 present and it might also want a means of sending no additional
11895 certificates (for example the chain has two certificates and the
11900 * Add the possibility to build without the ENGINE framework.
11902 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11904 * Under Win32 gmtime() can return NULL: check return value in
11905 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11909 * DSA routines: under certain error conditions uninitialized BN objects
11910 could be freed. Solution: make sure initialization is performed early
11911 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11912 Nils Larsch <nla@trustcenter.de> via PR#459)
11916 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11917 checked on reconnect on the client side, therefore session resumption
11918 could still fail with a "ssl session id is different" error. This
11919 behaviour is masked when SSL_OP_ALL is used due to
11920 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11921 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11922 followup to PR #377.
11926 * IA-32 assembler support enhancements: unified ELF targets, support
11927 for SCO/Caldera platforms, fix for Cygwin shared build.
11931 * Add support for FreeBSD on sparc64. As a consequence, support for
11932 FreeBSD on non-x86 processors is separate from x86 processors on
11933 the config script, much like the NetBSD support.
11935 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11937 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11939 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11942 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11943 code (06) was taken as the first octet of the session ID and the last
11944 octet was ignored consequently. As a result SSLv2 client side session
11945 caching could not have worked due to the session ID mismatch between
11947 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11952 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11953 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11958 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11959 seems that in spite of existing for more than a year, many application
11960 author have done nothing to provide the necessary callbacks, which
11961 means that this particular engine will not work properly anywhere.
11962 This is a very unfortunate situation which forces us, in the name
11963 of usability, to give the hw_ncipher.c a static lock, which is part
11965 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11966 appear in 0.9.8 or later. We EXPECT application authors to have
11967 dealt properly with this when 0.9.8 is released (unless we actually
11968 make such changes in the libcrypto locking code that changes will
11969 have to be made anyway).
11973 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11974 octets have been read, EOF or an error occurs. Without this change
11975 some truncated ASN1 structures will not produce an error.
11979 * Disable Heimdal support, since it hasn't been fully implemented.
11980 Still give the possibility to force the use of Heimdal, but with
11981 warnings and a request that patches get sent to openssl-dev.
11985 * Add the VC-CE target, introduce the WINCE sysname, and add
11986 INSTALL.WCE and appropriate conditionals to make it build.
11988 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11990 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11991 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11992 edit numbers of the version.
11994 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11996 * Introduce safe string copy and catenation functions
11997 (BUF_strlcpy() and BUF_strlcat()).
11999 *Ben Laurie (CHATS) and Richard Levitte*
12001 * Avoid using fixed-size buffers for one-line DNs.
12003 *Ben Laurie (CHATS)*
12005 * Add BUF_MEM_grow_clean() to avoid information leakage when
12006 resizing buffers containing secrets, and use where appropriate.
12008 *Ben Laurie (CHATS)*
12010 * Avoid using fixed size buffers for configuration file location.
12012 *Ben Laurie (CHATS)*
12014 * Avoid filename truncation for various CA files.
12016 *Ben Laurie (CHATS)*
12018 * Use sizeof in preference to magic numbers.
12020 *Ben Laurie (CHATS)*
12022 * Avoid filename truncation in cert requests.
12024 *Ben Laurie (CHATS)*
12026 * Add assertions to check for (supposedly impossible) buffer
12029 *Ben Laurie (CHATS)*
12031 * Don't cache truncated DNS entries in the local cache (this could
12032 potentially lead to a spoofing attack).
12034 *Ben Laurie (CHATS)*
12036 * Fix various buffers to be large enough for hex/decimal
12037 representations in a platform independent manner.
12039 *Ben Laurie (CHATS)*
12041 * Add CRYPTO_realloc_clean() to avoid information leakage when
12042 resizing buffers containing secrets, and use where appropriate.
12044 *Ben Laurie (CHATS)*
12046 * Add BIO_indent() to avoid much slightly worrying code to do
12049 *Ben Laurie (CHATS)*
12051 * Convert sprintf()/BIO_puts() to BIO_printf().
12053 *Ben Laurie (CHATS)*
12055 * buffer_gets() could terminate with the buffer only half
12058 *Ben Laurie (CHATS)*
12060 * Add assertions to prevent user-supplied crypto functions from
12061 overflowing internal buffers by having large block sizes, etc.
12063 *Ben Laurie (CHATS)*
12065 * New OPENSSL_assert() macro (similar to assert(), but enabled
12068 *Ben Laurie (CHATS)*
12070 * Eliminate unused copy of key in RC4.
12072 *Ben Laurie (CHATS)*
12074 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12076 *Ben Laurie (CHATS)*
12078 * Fix off-by-one error in EGD path.
12080 *Ben Laurie (CHATS)*
12082 * If RANDFILE path is too long, ignore instead of truncating.
12084 *Ben Laurie (CHATS)*
12086 * Eliminate unused and incorrectly sized X.509 structure
12089 *Ben Laurie (CHATS)*
12091 * Eliminate unused and dangerous function knumber().
12093 *Ben Laurie (CHATS)*
12095 * Eliminate unused and dangerous structure, KSSL_ERR.
12097 *Ben Laurie (CHATS)*
12099 * Protect against overlong session ID context length in an encoded
12100 session object. Since these are local, this does not appear to be
12103 *Ben Laurie (CHATS)*
12105 * Change from security patch (see 0.9.6e below) that did not affect
12106 the 0.9.6 release series:
12108 Remote buffer overflow in SSL3 protocol - an attacker could
12109 supply an oversized master key in Kerberos-enabled versions.
12112 *Ben Laurie (CHATS)*
12114 * Change the SSL kerb5 codes to match RFC 2712.
12118 * Make -nameopt work fully for req and add -reqopt switch.
12120 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12122 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12124 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12126 * Make sure tests can be performed even if the corresponding algorithms
12127 have been removed entirely. This was also the last step to make
12128 OpenSSL compilable with DJGPP under all reasonable conditions.
12130 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12132 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12133 to allow version independent disabling of normally unselected ciphers,
12134 which may be activated as a side-effect of selecting a single cipher.
12136 (E.g., cipher list string "RSA" enables ciphersuites that are left
12137 out of "ALL" because they do not provide symmetric encryption.
12138 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12140 *Lutz Jaenicke, Bodo Moeller*
12142 * Add appropriate support for separate platform-dependent build
12143 directories. The recommended way to make a platform-dependent
12144 build directory is the following (tested on Linux), maybe with
12147 # Place yourself outside of the OpenSSL source tree. In
12148 # this example, the environment variable OPENSSL_SOURCE
12149 # is assumed to contain the absolute OpenSSL source directory.
12150 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12151 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12152 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12153 mkdir -p `dirname $F`
12154 ln -s $OPENSSL_SOURCE/$F $F
12157 To be absolutely sure not to disturb the source tree, a "make clean"
12158 is a good thing. If it isn't successful, don't worry about it,
12159 it probably means the source directory is very clean.
12163 * Make sure any ENGINE control commands make local copies of string
12164 pointers passed to them whenever necessary. Otherwise it is possible
12165 the caller may have overwritten (or deallocated) the original string
12166 data when a later ENGINE operation tries to use the stored values.
12168 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12170 * Improve diagnostics in file reading and command-line digests.
12172 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12174 * Add AES modes CFB and OFB to the object database. Correct an
12175 error in AES-CFB decryption.
12179 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12180 allows existing EVP_CIPHER_CTX structures to be reused after
12181 calling `EVP_*Final()`. This behaviour is used by encryption
12182 BIOs and some applications. This has the side effect that
12183 applications must explicitly clean up cipher contexts with
12184 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12188 * Check the values of dna and dnb in bn_mul_recursive before calling
12189 bn_mul_comba (a non zero value means the a or b arrays do not contain
12190 n2 elements) and fallback to bn_mul_normal if either is not zero.
12194 * Fix escaping of non-ASCII characters when using the -subj option
12195 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12199 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12200 form for "surname", serialNumber has no short form.
12201 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12202 therefore remove "mail" short name for "internet 7".
12203 The OID for unique identifiers in X509 certificates is
12204 x500UniqueIdentifier, not uniqueIdentifier.
12205 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12209 * Add an "init" command to the ENGINE config module and auto initialize
12210 ENGINEs. Without any "init" command the ENGINE will be initialized
12211 after all ctrl commands have been executed on it. If init=1 the
12212 ENGINE is initialized at that point (ctrls before that point are run
12213 on the uninitialized ENGINE and after on the initialized one). If
12214 init=0 then the ENGINE will not be initialized at all.
12218 * Fix the 'app_verify_callback' interface so that the user-defined
12219 argument is actually passed to the callback: In the
12220 SSL_CTX_set_cert_verify_callback() prototype, the callback
12221 declaration has been changed from
12224 int (*cb)(X509_STORE_CTX *,void *);
12225 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12226 i=s->ctx->app_verify_callback(&ctx)
12227 has been changed into
12228 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12230 To update applications using SSL_CTX_set_cert_verify_callback(),
12231 a dummy argument can be added to their callback functions.
12233 *D. K. Smetters <smetters@parc.xerox.com>*
12235 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12237 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12239 * Add and OPENSSL_LOAD_CONF define which will cause
12240 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12241 This allows older applications to transparently support certain
12242 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12243 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12244 load the config file and OPENSSL_add_all_algorithms_conf() which will
12245 always load it have also been added.
12249 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12250 Adjust NIDs and EVP layer.
12252 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12254 * Config modules support in openssl utility.
12256 Most commands now load modules from the config file,
12257 though in a few (such as version) this isn't done
12258 because it couldn't be used for anything.
12260 In the case of ca and req the config file used is
12261 the same as the utility itself: that is the -config
12262 command line option can be used to specify an
12267 * Move default behaviour from OPENSSL_config(). If appname is NULL
12268 use "openssl_conf" if filename is NULL use default openssl config file.
12272 * Add an argument to OPENSSL_config() to allow the use of an alternative
12273 config section name. Add a new flag to tolerate a missing config file
12274 and move code to CONF_modules_load_file().
12278 * Support for crypto accelerator cards from Accelerated Encryption
12279 Processing, www.aep.ie. (Use engine 'aep')
12280 The support was copied from 0.9.6c [engine] and adapted/corrected
12281 to work with the new engine framework.
12283 *AEP Inc. and Richard Levitte*
12285 * Support for SureWare crypto accelerator cards from Baltimore
12286 Technologies. (Use engine 'sureware')
12287 The support was copied from 0.9.6c [engine] and adapted
12288 to work with the new engine framework.
12292 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12293 make the newer ENGINE framework commands for the CHIL engine work.
12295 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12297 * Make it possible to produce shared libraries on ReliantUNIX.
12299 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12301 * Add the configuration target debug-linux-ppro.
12302 Make 'openssl rsa' use the general key loading routines
12303 implemented in `apps.c`, and make those routines able to
12304 handle the key format FORMAT_NETSCAPE and the variant
12307 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12309 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12311 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12313 * Add -keyform to rsautl, and document -engine.
12315 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12317 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12318 BIO_R_NO_SUCH_FILE error code rather than the generic
12319 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12323 * Add new functions
12324 ERR_peek_last_error
12325 ERR_peek_last_error_line
12326 ERR_peek_last_error_line_data.
12327 These are similar to
12329 ERR_peek_error_line
12330 ERR_peek_error_line_data,
12331 but report on the latest error recorded rather than the first one
12332 still in the error queue.
12334 *Ben Laurie, Bodo Moeller*
12336 * default_algorithms option in ENGINE config module. This allows things
12338 default_algorithms = ALL
12339 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12343 * Preliminary ENGINE config module.
12347 * New experimental application configuration code.
12351 * Change the AES code to follow the same name structure as all other
12352 symmetric ciphers, and behave the same way. Move everything to
12353 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12355 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12357 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12359 *Ben Laurie and Theo de Raadt*
12361 * Add option to output public keys in req command.
12363 *Massimiliano Pala madwolf@openca.org*
12365 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12366 (up to about 10% better than before for P-192 and P-224).
12370 * New functions/macros
12372 SSL_CTX_set_msg_callback(ctx, cb)
12373 SSL_CTX_set_msg_callback_arg(ctx, arg)
12374 SSL_set_msg_callback(ssl, cb)
12375 SSL_set_msg_callback_arg(ssl, arg)
12377 to request calling a callback function
12379 void cb(int write_p, int version, int content_type,
12380 const void *buf, size_t len, SSL *ssl, void *arg)
12382 whenever a protocol message has been completely received
12383 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12384 protocol version according to which the SSL library interprets
12385 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12386 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12387 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12388 specification (change_cipher_spec(20), alert(21), handshake(22)).
12389 'buf' and 'len' point to the actual message, 'ssl' to the
12390 SSL object, and 'arg' is the application-defined value set by
12391 SSL[_CTX]_set_msg_callback_arg().
12393 'openssl s_client' and 'openssl s_server' have new '-msg' options
12394 to enable a callback that displays all protocol messages.
12398 * Change the shared library support so shared libraries are built as
12399 soon as the corresponding static library is finished, and thereby get
12400 openssl and the test programs linked against the shared library.
12401 This still only happens when the keyword "shard" has been given to
12402 the configuration scripts.
12404 NOTE: shared library support is still an experimental thing, and
12405 backward binary compatibility is still not guaranteed.
12407 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12409 * Add support for Subject Information Access extension.
12411 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12413 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12414 additional bytes when new memory had to be allocated, not just
12415 when reusing an existing buffer.
12419 * New command line and configuration option 'utf8' for the req command.
12420 This allows field values to be specified as UTF8 strings.
12424 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12425 runs for the former and machine-readable output for the latter.
12429 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12430 of the e-mail address in the DN (i.e., it will go into a certificate
12431 extension only). The new configuration file option 'email_in_dn = no'
12432 has the same effect.
12434 *Massimiliano Pala madwolf@openca.org*
12436 * Change all functions with names starting with `des_` to be starting
12437 with `DES_` instead. Add wrappers that are compatible with libdes,
12438 but are named `_ossl_old_des_*`. Finally, add macros that map the
12439 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12440 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12441 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12444 Since we provide two compatibility mappings, the user needs to
12445 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12446 compatibility is desired. The default (i.e., when that macro
12447 isn't defined) is OpenSSL 0.9.6c compatibility.
12449 There are also macros that enable and disable the support of old
12450 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12451 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12452 are defined, the default will apply: to support the old des routines.
12454 In either case, one must include openssl/des.h to get the correct
12455 definitions. Do not try to just include openssl/des_old.h, that
12458 NOTE: This is a major break of an old API into a new one. Software
12459 authors are encouraged to switch to the `DES_` style functions. Some
12460 time in the future, des_old.h and the libdes compatibility functions
12461 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12462 default), and then completely removed.
12466 * Test for certificates which contain unsupported critical extensions.
12467 If such a certificate is found during a verify operation it is
12468 rejected by default: this behaviour can be overridden by either
12469 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12470 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12471 X509_supported_extension() has also been added which returns 1 if a
12472 particular extension is supported.
12476 * Modify the behaviour of EVP cipher functions in similar way to digests
12477 to retain compatibility with existing code.
12481 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12482 compatibility with existing code. In particular the 'ctx' parameter does
12483 not have to be to be initialized before the call to EVP_DigestInit() and
12484 it is tidied up after a call to EVP_DigestFinal(). New function
12485 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12486 EVP_MD_CTX_copy() changed to not require the destination to be
12487 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12488 requires the destination to be valid.
12490 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12491 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12495 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12496 so that complete 'Handshake' protocol structures are kept in memory
12497 instead of overwriting 'msg_type' and 'length' with 'body' data.
12501 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12503 *Massimo Santin via Richard Levitte*
12505 * Major restructuring to the underlying ENGINE code. This includes
12506 reduction of linker bloat, separation of pure "ENGINE" manipulation
12507 (initialisation, etc) from functionality dealing with implementations
12508 of specific crypto interfaces. This change also introduces integrated
12509 support for symmetric ciphers and digest implementations - so ENGINEs
12510 can now accelerate these by providing EVP_CIPHER and EVP_MD
12511 implementations of their own. This is detailed in
12512 [crypto/engine/README.md](crypto/engine/README.md)
12513 as it couldn't be adequately described here. However, there are a few
12514 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12515 were changed in the original introduction of ENGINE code have now
12516 reverted back - the hooking from this code to ENGINE is now a good
12517 deal more passive and at run-time, operations deal directly with
12518 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12519 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12520 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12521 they were not being used by the framework as there is no concept of a
12522 BIGNUM_METHOD and they could not be generalised to the new
12523 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12524 ENGINE_cpy() has been removed as it cannot be consistently defined in
12529 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12533 * Change mkdef.pl to sort symbols that get the same entry number,
12534 and make sure the automatically generated functions `ERR_load_*`
12535 become part of libeay.num as well.
12539 * New function SSL_renegotiate_pending(). This returns true once
12540 renegotiation has been requested (either SSL_renegotiate() call
12541 or HelloRequest/ClientHello received from the peer) and becomes
12542 false once a handshake has been completed.
12543 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12544 sends a HelloRequest, but does not ensure that a handshake takes
12545 place. SSL_renegotiate_pending() is useful for checking if the
12546 client has followed the request.)
12550 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12551 By default, clients may request session resumption even during
12552 renegotiation (if session ID contexts permit); with this option,
12553 session resumption is possible only in the first handshake.
12555 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12556 more bits available for options that should not be part of
12557 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12561 * Add some demos for certificate and certificate request creation.
12565 * Make maximum certificate chain size accepted from the peer application
12566 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12567 "Douglas E. Engert" <deengert@anl.gov>.
12571 * Add support for shared libraries for Unixware-7
12572 (Boyd Lynn Gerber <gerberb@zenez.com>).
12576 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12577 be done prior to destruction. Use this to unload error strings from
12578 ENGINEs that load their own error strings. NB: This adds two new API
12579 functions to "get" and "set" this destroy handler in an ENGINE.
12583 * Alter all existing ENGINE implementations (except "openssl" and
12584 "openbsd") to dynamically instantiate their own error strings. This
12585 makes them more flexible to be built both as statically-linked ENGINEs
12586 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12587 Also, add stub code to each that makes building them as self-contained
12588 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12592 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12593 implementations into applications that are completely implemented in
12594 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12595 commands that can be used to configure what shared-library to load and
12596 to control aspects of the way it is handled. Also, made an update to
12597 the [README-Engine.md](README-Engine.md) file
12598 that brings its information up-to-date and
12599 provides some information and instructions on the "dynamic" ENGINE
12600 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12604 * Make it possible to unload ranges of ERR strings with a new
12605 "ERR_unload_strings" function.
12609 * Add a copy() function to EVP_MD.
12613 * Make EVP_MD routines take a context pointer instead of just the
12614 md_data void pointer.
12618 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12619 that the digest can only process a single chunk of data
12620 (typically because it is provided by a piece of
12621 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12622 is only going to provide a single chunk of data, and hence the
12623 framework needn't accumulate the data for oneshot drivers.
12627 * As with "ERR", make it possible to replace the underlying "ex_data"
12628 functions. This change also alters the storage and management of global
12629 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12630 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12631 index counters. The API functions that use this state have been changed
12632 to take a "class_index" rather than pointers to the class's local STACK
12633 and counter, and there is now an API function to dynamically create new
12634 classes. This centralisation allows us to (a) plug a lot of the
12635 thread-safety problems that existed, and (b) makes it possible to clean
12636 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12637 such data would previously have always leaked in application code and
12638 workarounds were in place to make the memory debugging turn a blind eye
12639 to it. Application code that doesn't use this new function will still
12640 leak as before, but their memory debugging output will announce it now
12641 rather than letting it slide.
12643 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12644 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12645 has a return value to indicate success or failure.
12649 * Make it possible to replace the underlying "ERR" functions such that the
12650 global state (2 LHASH tables and 2 locks) is only used by the "default"
12651 implementation. This change also adds two functions to "get" and "set"
12652 the implementation prior to it being automatically set the first time
12653 any other ERR function takes place. Ie. an application can call "get",
12654 pass the return value to a module it has just loaded, and that module
12655 can call its own "set" function using that value. This means the
12656 module's "ERR" operations will use (and modify) the error state in the
12657 application and not in its own statically linked copy of OpenSSL code.
12661 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12662 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12663 the operation, and provides a more encapsulated way for external code
12664 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12665 to use these functions rather than manually incrementing the counts.
12667 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12671 * Add EVP test program.
12675 * Add symmetric cipher support to ENGINE. Expect the API to change!
12679 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12680 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12681 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12682 These allow a CRL to be built without having to access X509_CRL fields
12683 directly. Modify 'ca' application to use new functions.
12687 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12688 bug workarounds. Rollback attack detection is a security feature.
12689 The problem will only arise on OpenSSL servers when TLSv1 is not
12690 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12691 Software authors not wanting to support TLSv1 will have special reasons
12692 for their choice and can explicitly enable this option.
12694 *Bodo Moeller, Lutz Jaenicke*
12696 * Rationalise EVP so it can be extended: don't include a union of
12697 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12698 (similar to those existing for EVP_CIPHER_CTX).
12703 EVP_MD_CTX_init(&md); /* new function call */
12704 EVP_DigestInit(&md, EVP_sha1());
12705 EVP_DigestUpdate(&md, in, len);
12706 EVP_DigestFinal(&md, out, NULL);
12707 EVP_MD_CTX_cleanup(&md); /* new function call */
12711 * Make DES key schedule conform to the usual scheme, as well as
12712 correcting its structure. This means that calls to DES functions
12713 now have to pass a pointer to a des_key_schedule instead of a
12714 plain des_key_schedule (which was actually always a pointer
12717 des_key_schedule ks;
12719 des_set_key_checked(..., &ks);
12720 des_ncbc_encrypt(..., &ks, ...);
12722 (Note that a later change renames 'des_...' into 'DES_...'.)
12726 * Initial reduction of linker bloat: the use of some functions, such as
12727 PEM causes large amounts of unused functions to be linked in due to
12728 poor organisation. For example pem_all.c contains every PEM function
12729 which has a knock on effect of linking in large amounts of (unused)
12730 ASN1 code. Grouping together similar functions and splitting unrelated
12731 functions prevents this.
12735 * Cleanup of EVP macros.
12739 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12740 correct `_ecb suffix`.
12744 * Add initial OCSP responder support to ocsp application. The
12745 revocation information is handled using the text based index
12746 use by the ca application. The responder can either handle
12747 requests generated internally, supplied in files (for example
12748 via a CGI script) or using an internal minimal server.
12752 * Add configuration choices to get zlib compression for TLS.
12756 * Changes to Kerberos SSL for RFC 2712 compliance:
12757 1. Implemented real KerberosWrapper, instead of just using
12758 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12759 2. Implemented optional authenticator field of KerberosWrapper.
12761 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12762 and authenticator structs; see crypto/krb5/.
12764 Generalized Kerberos calls to support multiple Kerberos libraries.
12765 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12766 via Richard Levitte*
12768 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12769 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12770 values for each of the key sizes rather than having just
12771 parameters (and 'speed' generating keys each time).
12775 * Speed up EVP routines.
12778 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12779 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12780 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12781 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12783 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12784 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12785 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12788 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12790 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12794 * Added the OS2-EMX target.
12796 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12798 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12799 New functions to support `NCONF `routines in extension code.
12800 New function `CONF_set_nconf()`
12801 to allow functions which take an `NCONF` to also handle the old `LHASH`
12802 structure: this means that the old `CONF` compatible routines can be
12803 retained (in particular w.rt. extensions) without having to duplicate the
12804 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12808 * Enhance the general user interface with mechanisms for inner control
12809 and with possibilities to have yes/no kind of prompts.
12813 * Change all calls to low-level digest routines in the library and
12814 applications to use EVP. Add missing calls to HMAC_cleanup() and
12815 don't assume HMAC_CTX can be copied using memcpy().
12817 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12819 * Add the possibility to control engines through control names but with
12820 arbitrary arguments instead of just a string.
12821 Change the key loaders to take a UI_METHOD instead of a callback
12822 function pointer. NOTE: this breaks binary compatibility with earlier
12823 versions of OpenSSL [engine].
12824 Adapt the nCipher code for these new conditions and add a card insertion
12829 * Enhance the general user interface with mechanisms to better support
12830 dialog box interfaces, application-defined prompts, the possibility
12831 to use defaults (for example default passwords from somewhere else)
12832 and interrupts/cancellations.
12836 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12837 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12841 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12842 tidy up some unnecessarily weird code in 'sk_new()').
12844 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12846 * Change the key loading routines for ENGINEs to use the same kind
12847 callback (pem_password_cb) as all other routines that need this
12852 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12853 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12854 than this minimum value is recommended.
12858 * New random seeder for OpenVMS, using the system process statistics
12859 that are easily reachable.
12863 * Windows apparently can't transparently handle global
12864 variables defined in DLLs. Initialisations such as:
12866 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12868 won't compile. This is used by the any applications that need to
12869 declare their own ASN1 modules. This was fixed by adding the option
12870 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12871 needed for static libraries under Win32.
12875 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12876 setting of purpose and trust fields. New X509_STORE trust and
12877 purpose functions and tidy up setting in other SSL functions.
12881 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12882 structure. These are inherited by X509_STORE_CTX when it is
12883 initialised. This allows various defaults to be set in the
12884 X509_STORE structure (such as flags for CRL checking and custom
12885 purpose or trust settings) for functions which only use X509_STORE_CTX
12886 internally such as S/MIME.
12888 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12889 trust settings if they are not set in X509_STORE. This allows X509_STORE
12890 purposes and trust (in S/MIME for example) to override any set by default.
12892 Add command line options for CRL checking to smime, s_client and s_server
12897 * Initial CRL based revocation checking. If the CRL checking flag(s)
12898 are set then the CRL is looked up in the X509_STORE structure and
12899 its validity and signature checked, then if the certificate is found
12900 in the CRL the verify fails with a revoked error.
12902 Various new CRL related callbacks added to X509_STORE_CTX structure.
12904 Command line options added to 'verify' application to support this.
12906 This needs some additional work, such as being able to handle multiple
12907 CRLs with different times, extension based lookup (rather than just
12908 by subject name) and ultimately more complete V2 CRL extension
12913 * Add a general user interface API (crypto/ui/). This is designed
12914 to replace things like des_read_password and friends (backward
12915 compatibility functions using this new API are provided).
12916 The purpose is to remove prompting functions from the DES code
12917 section as well as provide for prompting through dialog boxes in
12918 a window system and the like.
12922 * Add "ex_data" support to ENGINE so implementations can add state at a
12923 per-structure level rather than having to store it globally.
12927 * Make it possible for ENGINE structures to be copied when retrieved by
12928 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12929 This causes the "original" ENGINE structure to act like a template,
12930 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12931 operational state can be localised to each ENGINE structure, despite the
12932 fact they all share the same "methods". New ENGINE structures returned in
12933 this case have no functional references and the return value is the single
12934 structural reference. This matches the single structural reference returned
12935 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12940 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12941 needs to match any other type at all we need to manually clear the
12946 * Changes to the "openssl engine" utility to include;
12947 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12948 about an ENGINE's available control commands.
12949 - executing control commands from command line arguments using the
12950 '-pre' and '-post' switches. '-post' is only used if '-t' is
12951 specified and the ENGINE is successfully initialised. The syntax for
12952 the individual commands are colon-separated, for example;
12953 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12957 * New dynamic control command support for ENGINEs. ENGINEs can now
12958 declare their own commands (numbers), names (strings), descriptions,
12959 and input types for run-time discovery by calling applications. A
12960 subset of these commands are implicitly classed as "executable"
12961 depending on their input type, and only these can be invoked through
12962 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12963 can be based on user input, config files, etc). The distinction is
12964 that "executable" commands cannot return anything other than a boolean
12965 result and can only support numeric or string input, whereas some
12966 discoverable commands may only be for direct use through
12967 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12968 pointers, or other custom uses. The "executable" commands are to
12969 support parameterisations of ENGINE behaviour that can be
12970 unambiguously defined by ENGINEs and used consistently across any
12971 OpenSSL-based application. Commands have been added to all the
12972 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12973 control over shared-library paths without source code alterations.
12977 * Changed all ENGINE implementations to dynamically allocate their
12978 ENGINEs rather than declaring them statically. Apart from this being
12979 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12980 this also allows the implementations to compile without using the
12981 internal engine_int.h header.
12985 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12986 'const' value. Any code that should be able to modify a RAND_METHOD
12987 should already have non-const pointers to it (ie. they should only
12988 modify their own ones).
12992 * Made a variety of little tweaks to the ENGINE code.
12993 - "atalla" and "ubsec" string definitions were moved from header files
12994 to C code. "nuron" string definitions were placed in variables
12995 rather than hard-coded - allowing parameterisation of these values
12996 later on via ctrl() commands.
12997 - Removed unused "#if 0"'d code.
12998 - Fixed engine list iteration code so it uses ENGINE_free() to release
12999 structural references.
13000 - Constified the RAND_METHOD element of ENGINE structures.
13001 - Constified various get/set functions as appropriate and added
13002 missing functions (including a catch-all ENGINE_cpy that duplicates
13003 all ENGINE values onto a new ENGINE except reference counts/state).
13004 - Removed NULL parameter checks in get/set functions. Setting a method
13005 or function to NULL is a way of cancelling out a previously set
13006 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13007 and doesn't justify the extra error symbols and code.
13008 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13009 flags from engine_int.h to engine.h.
13010 - Changed prototypes for ENGINE handler functions (init(), finish(),
13011 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13015 * Implement binary inversion algorithm for BN_mod_inverse in addition
13016 to the algorithm using long division. The binary algorithm can be
13017 used only if the modulus is odd. On 32-bit systems, it is faster
13018 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13019 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13020 up to 450 bits. In 64-bit environments, the binary algorithm
13021 appears to be advantageous for much longer moduli; here we use it
13022 for moduli up to 2048 bits.
13026 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13027 could not support the combine flag in choice fields.
13031 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13032 extensions from a certificate request to the certificate.
13036 * Allow multiple 'certopt' and 'nameopt' options to be separated
13037 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13038 file: this allows the display of the certificate about to be
13039 signed to be customised, to allow certain fields to be included
13040 or excluded and extension details. The old system didn't display
13041 multicharacter strings properly, omitted fields not in the policy
13042 and couldn't display additional details such as extensions.
13046 * Function EC_POINTs_mul for multiple scalar multiplication
13047 of an arbitrary number of elliptic curve points
13048 \sum scalars[i]*points[i],
13049 optionally including the generator defined for the EC_GROUP:
13050 scalar*generator + \sum scalars[i]*points[i].
13052 EC_POINT_mul is a simple wrapper function for the typical case
13053 that the point list has just one item (besides the optional
13058 * First EC_METHODs for curves over GF(p):
13060 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13061 operations and provides various method functions that can also
13062 operate with faster implementations of modular arithmetic.
13064 EC_GFp_mont_method() reuses most functions that are part of
13065 EC_GFp_simple_method, but uses Montgomery arithmetic.
13067 *Bodo Moeller; point addition and point doubling
13068 implementation directly derived from source code provided by
13069 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13071 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13072 crypto/ec/ec_lib.c):
13074 Curves are EC_GROUP objects (with an optional group generator)
13075 based on EC_METHODs that are built into the library.
13077 Points are EC_POINT objects based on EC_GROUP objects.
13079 Most of the framework would be able to handle curves over arbitrary
13080 finite fields, but as there are no obvious types for fields other
13081 than GF(p), some functions are limited to that for now.
13085 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13086 that the file contains a complete HTTP response.
13090 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13091 change the def and num file printf format specifier from "%-40sXXX"
13092 to "%-39s XXX". The latter will always guarantee a space after the
13093 field while the former will cause them to run together if the field
13094 is 40 of more characters long.
13098 * Constify the cipher and digest 'method' functions and structures
13099 and modify related functions to take constant EVP_MD and EVP_CIPHER
13104 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13105 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13109 * Modify `EVP_Digest*()` routines so they now return values. Although the
13110 internal software routines can never fail additional hardware versions
13115 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13117 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13118 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13121 ERR_R_NESTED_ASN1_ERROR
13123 ERR_R_MISSING_ASN1_EOS
13124 were 4 .. 9, conflicting with
13125 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13127 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13128 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13130 Add new error code 'ERR_R_INTERNAL_ERROR'.
13134 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13139 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13140 sets the subject name for a new request or supersedes the
13141 subject name in a given request. Formats that can be parsed are
13142 'CN=Some Name, OU=myOU, C=IT'
13144 'CN=Some Name/OU=myOU/C=IT'.
13146 Add options '-batch' and '-verbose' to 'openssl req'.
13148 *Massimiliano Pala <madwolf@hackmasters.net>*
13150 * Introduce the possibility to access global variables through
13151 functions on platform were that's the best way to handle exporting
13152 global variables in shared libraries. To enable this functionality,
13153 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13154 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13155 is normally done by Configure or something similar).
13157 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13158 in the source file (foo.c) like this:
13160 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13161 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13163 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13164 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13166 OPENSSL_DECLARE_GLOBAL(int,foo);
13167 #define foo OPENSSL_GLOBAL_REF(foo)
13168 OPENSSL_DECLARE_GLOBAL(double,bar);
13169 #define bar OPENSSL_GLOBAL_REF(bar)
13171 The #defines are very important, and therefore so is including the
13172 header file everywhere where the defined globals are used.
13174 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13175 of ASN.1 items, but that structure is a bit different.
13177 The largest change is in util/mkdef.pl which has been enhanced with
13178 better and easier to understand logic to choose which symbols should
13179 go into the Windows .def files as well as a number of fixes and code
13180 cleanup (among others, algorithm keywords are now sorted
13181 lexicographically to avoid constant rewrites).
13185 * In BN_div() keep a copy of the sign of 'num' before writing the
13186 result to 'rm' because if rm==num the value will be overwritten
13187 and produce the wrong result if 'num' is negative: this caused
13188 problems with BN_mod() and BN_nnmod().
13192 * Function OCSP_request_verify(). This checks the signature on an
13193 OCSP request and verifies the signer certificate. The signer
13194 certificate is just checked for a generic purpose and OCSP request
13199 * Add OCSP_check_validity() function to check the validity of OCSP
13200 responses. OCSP responses are prepared in real time and may only
13201 be a few seconds old. Simply checking that the current time lies
13202 between thisUpdate and nextUpdate max reject otherwise valid responses
13203 caused by either OCSP responder or client clock inaccuracy. Instead
13204 we allow thisUpdate and nextUpdate to fall within a certain period of
13205 the current time. The age of the response can also optionally be
13206 checked. Two new options -validity_period and -status_age added to
13211 * If signature or public key algorithm is unrecognized print out its
13212 OID rather that just UNKNOWN.
13216 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13217 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13218 ID to be generated from the issuer certificate alone which can then be
13219 passed to OCSP_id_issuer_cmp().
13223 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13224 ASN1 modules to export functions returning ASN1_ITEM pointers
13225 instead of the ASN1_ITEM structures themselves. This adds several
13226 new macros which allow the underlying ASN1 function/structure to
13227 be accessed transparently. As a result code should not use ASN1_ITEM
13228 references directly (such as &X509_it) but instead use the relevant
13229 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13230 use of the new ASN1 code on platforms where exporting structures
13231 is problematical (for example in shared libraries) but exporting
13232 functions returning pointers to structures is not.
13236 * Add support for overriding the generation of SSL/TLS session IDs.
13237 These callbacks can be registered either in an SSL_CTX or per SSL.
13238 The purpose of this is to allow applications to control, if they wish,
13239 the arbitrary values chosen for use as session IDs, particularly as it
13240 can be useful for session caching in multiple-server environments. A
13241 command-line switch for testing this (and any client code that wishes
13242 to use such a feature) has been added to "s_server".
13244 *Geoff Thorpe, Lutz Jaenicke*
13246 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13247 of the form `#if defined(...) || defined(...) || ...` and
13248 `#if !defined(...) && !defined(...) && ...`. This also avoids
13249 the growing number of special cases it was previously handling.
13253 * Make all configuration macros available for application by making
13254 sure they are available in opensslconf.h, by giving them names starting
13255 with `OPENSSL_` to avoid conflicts with other packages and by making
13256 sure e_os2.h will cover all platform-specific cases together with
13258 Additionally, it is now possible to define configuration/platform-
13259 specific names (called "system identities"). In the C code, these
13260 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13261 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13262 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13267 * New option -set_serial to 'req' and 'x509' this allows the serial
13268 number to use to be specified on the command line. Previously self
13269 signed certificates were hard coded with serial number 0 and the
13270 CA options of 'x509' had to use a serial number in a file which was
13275 * New options to 'ca' utility to support V2 CRL entry extensions.
13276 Currently CRL reason, invalidity date and hold instruction are
13277 supported. Add new CRL extensions to V3 code and some new objects.
13281 * New function EVP_CIPHER_CTX_set_padding() this is used to
13282 disable standard block padding (aka PKCS#5 padding) in the EVP
13283 API, which was previously mandatory. This means that the data is
13284 not padded in any way and so the total length much be a multiple
13285 of the block size, otherwise an error occurs.
13289 * Initial (incomplete) OCSP SSL support.
13293 * New function OCSP_parse_url(). This splits up a URL into its host,
13294 port and path components: primarily to parse OCSP URLs. New -url
13295 option to ocsp utility.
13299 * New nonce behavior. The return value of OCSP_check_nonce() now
13300 reflects the various checks performed. Applications can decide
13301 whether to tolerate certain situations such as an absent nonce
13302 in a response when one was present in a request: the ocsp application
13303 just prints out a warning. New function OCSP_add1_basic_nonce()
13304 this is to allow responders to include a nonce in a response even if
13305 the request is nonce-less.
13309 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13310 skipped when using openssl x509 multiple times on a single input file,
13311 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13315 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13316 set string type: to handle setting ASN1_TIME structures. Fix ca
13317 utility to correctly initialize revocation date of CRLs.
13321 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13322 the clients preferred ciphersuites and rather use its own preferences.
13323 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13324 Internet Explorer by ensuring unchanged hash method during stepup.
13325 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13329 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13330 to aes and add a new 'exist' option to print out symbols that don't
13335 * Additional options to ocsp utility to allow flags to be set and
13336 additional certificates supplied.
13340 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13341 OCSP client a number of certificate to only verify the response
13346 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13347 handle the new API. Currently only ECB, CBC modes supported. Add new
13350 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13351 Encryption Standard (AES) Ciphersuites for Transport Layer
13352 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13353 not enabled by default and were not part of the "ALL" ciphersuite
13354 alias because they were not yet official; they could be
13355 explicitly requested by specifying the "AESdraft" ciphersuite
13356 group alias. In the final release of OpenSSL 0.9.7, the group
13357 alias is called "AES" and is part of "ALL".)
13359 *Ben Laurie, Steve Henson, Bodo Moeller*
13361 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13362 request to response.
13366 * Functions for OCSP responders. OCSP_request_onereq_count(),
13367 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13368 extract information from a certificate request. OCSP_response_create()
13369 creates a response and optionally adds a basic response structure.
13370 OCSP_basic_add1_status() adds a complete single response to a basic
13371 response and returns the OCSP_SINGLERESP structure just added (to allow
13372 extensions to be included for example). OCSP_basic_add1_cert() adds a
13373 certificate to a basic response and OCSP_basic_sign() signs a basic
13374 response with various flags. New helper functions ASN1_TIME_check()
13375 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13376 (converts ASN1_TIME to GeneralizedTime).
13380 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13381 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13382 structure from a certificate. X509_pubkey_digest() digests the public_key
13383 contents: this is used in various key identifiers.
13387 * Make sk_sort() tolerate a NULL argument.
13389 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13391 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13392 passed by the function are trusted implicitly. If any of them signed the
13393 response then it is assumed to be valid and is not verified.
13397 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13398 to data. This was previously part of the PKCS7 ASN1 code. This
13399 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13400 *Steve Henson, reported by Kenneth R. Robinette
13401 <support@securenetterm.com>*
13403 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13404 routines: without these tracing memory leaks is very painful.
13405 Fix leaks in PKCS12 and PKCS7 routines.
13409 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13410 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13411 effectively meant GeneralizedTime would never be used. Now it
13412 is initialised to -1 but X509_time_adj() now has to check the value
13413 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13414 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13415 *Steve Henson, reported by Kenneth R. Robinette
13416 <support@securenetterm.com>*
13418 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13419 result in a zero length in the ASN1_INTEGER structure which was
13420 not consistent with the structure when d2i_ASN1_INTEGER() was used
13421 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13422 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13423 where it did not print out a minus for negative ASN1_INTEGER.
13427 * Add summary printout to ocsp utility. The various functions which
13428 convert status values to strings have been renamed to:
13429 OCSP_response_status_str(), OCSP_cert_status_str() and
13430 OCSP_crl_reason_str() and are no longer static. New options
13431 to verify nonce values and to disable verification. OCSP response
13432 printout format cleaned up.
13436 * Add additional OCSP certificate checks. These are those specified
13437 in RFC2560. This consists of two separate checks: the CA of the
13438 certificate being checked must either be the OCSP signer certificate
13439 or the issuer of the OCSP signer certificate. In the latter case the
13440 OCSP signer certificate must contain the OCSP signing extended key
13441 usage. This check is performed by attempting to match the OCSP
13442 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13443 in the OCSP_CERTID structures of the response.
13447 * Initial OCSP certificate verification added to OCSP_basic_verify()
13448 and related routines. This uses the standard OpenSSL certificate
13449 verify routines to perform initial checks (just CA validity) and
13450 to obtain the certificate chain. Then additional checks will be
13451 performed on the chain. Currently the root CA is checked to see
13452 if it is explicitly trusted for OCSP signing. This is used to set
13453 a root CA as a global signing root: that is any certificate that
13454 chains to that CA is an acceptable OCSP signing certificate.
13458 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13459 extensions from a separate configuration file.
13460 As when reading extensions from the main configuration file,
13461 the '-extensions ...' option may be used for specifying the
13464 *Massimiliano Pala <madwolf@comune.modena.it>*
13466 * New OCSP utility. Allows OCSP requests to be generated or
13467 read. The request can be sent to a responder and the output
13468 parsed, outputted or printed in text form. Not complete yet:
13469 still needs to check the OCSP response validity.
13473 * New subcommands for 'openssl ca':
13474 `openssl ca -status <serial>` prints the status of the cert with
13475 the given serial number (according to the index file).
13476 `openssl ca -updatedb` updates the expiry status of certificates
13479 *Massimiliano Pala <madwolf@comune.modena.it>*
13481 * New '-newreq-nodes' command option to CA.pl. This is like
13482 '-newreq', but calls 'openssl req' with the '-nodes' option
13483 so that the resulting key is not encrypted.
13485 *Damien Miller <djm@mindrot.org>*
13487 * New configuration for the GNU Hurd.
13489 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13491 * Initial code to implement OCSP basic response verify. This
13492 is currently incomplete. Currently just finds the signer's
13493 certificate and verifies the signature on the response.
13497 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13498 value of OPENSSLDIR. This is available via the new '-d' option
13499 to 'openssl version', and is also included in 'openssl version -a'.
13503 * Allowing defining memory allocation callbacks that will be given
13504 file name and line number information in additional arguments
13505 (a `const char*` and an int). The basic functionality remains, as
13506 well as the original possibility to just replace malloc(),
13507 realloc() and free() by functions that do not know about these
13508 additional arguments. To register and find out the current
13509 settings for extended allocation functions, the following
13510 functions are provided:
13512 CRYPTO_set_mem_ex_functions
13513 CRYPTO_set_locked_mem_ex_functions
13514 CRYPTO_get_mem_ex_functions
13515 CRYPTO_get_locked_mem_ex_functions
13517 These work the same way as CRYPTO_set_mem_functions and friends.
13518 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13519 extended allocation function is enabled.
13520 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13521 a conventional allocation function is enabled.
13523 *Richard Levitte, Bodo Moeller*
13525 * Finish off removing the remaining LHASH function pointer casts.
13526 There should no longer be any prototype-casting required when using
13527 the LHASH abstraction, and any casts that remain are "bugs". See
13528 the callback types and macros at the head of lhash.h for details
13529 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13533 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13534 If /dev/[u]random devices are not available or do not return enough
13535 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13537 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13538 /etc/entropy will be queried once each in this sequence, querying stops
13539 when enough entropy was collected without querying more sockets.
13543 * Change the Unix RAND_poll() variant to be able to poll several
13544 random devices, as specified by DEVRANDOM, until a sufficient amount
13545 of data has been collected. We spend at most 10 ms on each file
13546 (select timeout) and read in non-blocking mode. DEVRANDOM now
13547 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13548 (previously it was just the string "/dev/urandom"), so on typical
13549 platforms the 10 ms delay will never occur.
13550 Also separate out the Unix variant to its own file, rand_unix.c.
13551 For VMS, there's a currently-empty rand_vms.c.
13555 * Move OCSP client related routines to ocsp_cl.c. These
13556 provide utility functions which an application needing
13557 to issue a request to an OCSP responder and analyse the
13558 response will typically need: as opposed to those which an
13559 OCSP responder itself would need which will be added later.
13561 OCSP_request_sign() signs an OCSP request with an API similar
13562 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13563 response. OCSP_response_get1_basic() extracts basic response
13564 from response. OCSP_resp_find_status(): finds and extracts status
13565 information from an OCSP_CERTID structure (which will be created
13566 when the request structure is built). These are built from lower
13567 level functions which work on OCSP_SINGLERESP structures but
13568 won't normally be used unless the application wishes to examine
13569 extensions in the OCSP response for example.
13571 Replace nonce routines with a pair of functions.
13572 OCSP_request_add1_nonce() adds a nonce value and optionally
13573 generates a random value. OCSP_check_nonce() checks the
13574 validity of the nonce in an OCSP response.
13578 * Change function OCSP_request_add() to OCSP_request_add0_id().
13579 This doesn't copy the supplied OCSP_CERTID and avoids the
13580 need to free up the newly created id. Change return type
13581 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13582 This can then be used to add extensions to the request.
13583 Deleted OCSP_request_new(), since most of its functionality
13584 is now in OCSP_REQUEST_new() (and the case insensitive name
13585 clash) apart from the ability to set the request name which
13586 will be added elsewhere.
13590 * Update OCSP API. Remove obsolete extensions argument from
13591 various functions. Extensions are now handled using the new
13592 OCSP extension code. New simple OCSP HTTP function which
13593 can be used to send requests and parse the response.
13597 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13598 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13599 uses the special reorder version of SET OF to sort the attributes
13600 and reorder them to match the encoded order. This resolves a long
13601 standing problem: a verify on a PKCS7 structure just after signing
13602 it used to fail because the attribute order did not match the
13603 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13604 it uses the received order. This is necessary to tolerate some broken
13605 software that does not order SET OF. This is handled by encoding
13606 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13607 to produce the required SET OF.
13611 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13612 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13613 files to get correct declarations of the ASN.1 item variables.
13617 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13618 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13619 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13620 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13621 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13622 ASN1_ITEM and no wrapper functions.
13626 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13627 replace the old function pointer based I/O routines. Change most of
13628 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13632 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13633 lines, recognize more "algorithms" that can be deselected, and make
13634 it complain about algorithm deselection that isn't recognised.
13638 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13639 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13640 to use new functions. Add NO_ASN1_OLD which can be set to remove
13641 some old style ASN1 functions: this can be used to determine if old
13642 code will still work when these eventually go away.
13646 * New extension functions for OCSP structures, these follow the
13647 same conventions as certificates and CRLs.
13651 * New function X509V3_add1_i2d(). This automatically encodes and
13652 adds an extension. Its behaviour can be customised with various
13653 flags to append, replace or delete. Various wrappers added for
13654 certificates and CRLs.
13658 * Fix to avoid calling the underlying ASN1 print routine when
13659 an extension cannot be parsed. Correct a typo in the
13660 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13664 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13665 entries for variables.
13669 * Add functionality to `apps/openssl.c` for detecting locking
13670 problems: As the program is single-threaded, all we have
13671 to do is register a locking callback using an array for
13672 storing which locks are currently held by the program.
13676 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13677 SSL_get_ex_data_X509_STORE_idx(), which is used in
13678 ssl_verify_cert_chain() and thus can be called at any time
13679 during TLS/SSL handshakes so that thread-safety is essential.
13680 Unfortunately, the ex_data design is not at all suited
13681 for multi-threaded use, so it probably should be abolished.
13685 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13687 *Broadcom, tweaked and integrated by Geoff Thorpe*
13689 * Move common extension printing code to new function
13690 X509V3_print_extensions(). Reorganise OCSP print routines and
13691 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13695 * New function X509_signature_print() to remove duplication in some
13700 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13701 set (this was treated exactly the same as SET OF previously). This
13702 is used to reorder the STACK representing the structure to match the
13703 encoding. This will be used to get round a problem where a PKCS7
13704 structure which was signed could not be verified because the STACK
13705 order did not reflect the encoded order.
13709 * Reimplement the OCSP ASN1 module using the new code.
13713 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13714 for its ASN1 operations. The old style function pointers still exist
13715 for now but they will eventually go away.
13719 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13720 completely replaces the old ASN1 functionality with a table driven
13721 encoder and decoder which interprets an ASN1_ITEM structure describing
13722 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13723 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13724 has also been converted to the new form.
13728 * Change BN_mod_exp_recp so that negative moduli are tolerated
13729 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13730 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13731 for negative moduli.
13735 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13736 of not touching the result's sign bit.
13740 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13745 * Changed the LHASH code to use prototypes for callbacks, and created
13746 macros to declare and implement thin (optionally static) functions
13747 that provide type-safety and avoid function pointer casting for the
13748 type-specific callbacks.
13752 * Added Kerberos Cipher Suites to be used with TLS, as written in
13754 *Veers Staats <staatsvr@asc.hpc.mil>,
13755 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13757 * Reformat the FAQ so the different questions and answers can be divided
13758 in sections depending on the subject.
13762 * Have the zlib compression code load ZLIB.DLL dynamically under
13767 * New function BN_mod_sqrt for computing square roots modulo a prime
13768 (using the probabilistic Tonelli-Shanks algorithm unless
13769 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13770 be handled deterministically).
13772 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13774 * Make BN_mod_inverse faster by explicitly handling small quotients
13775 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13776 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13780 * New function BN_kronecker.
13784 * Fix BN_gcd so that it works on negative inputs; the result is
13785 positive unless both parameters are zero.
13786 Previously something reasonably close to an infinite loop was
13787 possible because numbers could be growing instead of shrinking
13788 in the implementation of Euclid's algorithm.
13792 * Fix BN_is_word() and BN_is_one() macros to take into account the
13793 sign of the number in question.
13795 Fix BN_is_word(a,w) to work correctly for w == 0.
13797 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13798 because its test if the absolute value of 'a' equals 'w'.
13799 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13800 it exists mostly for use in the implementations of BN_is_zero(),
13801 BN_is_one(), and BN_is_word().
13805 * New function BN_swap.
13809 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13810 the exponentiation functions are more likely to produce reasonable
13811 results on negative inputs.
13815 * Change BN_mod_mul so that the result is always non-negative.
13816 Previously, it could be negative if one of the factors was negative;
13817 I don't think anyone really wanted that behaviour.
13821 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13822 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13823 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13824 and add new functions:
13833 BN_mod_lshift1_quick
13835 BN_mod_lshift_quick
13837 These functions always generate non-negative results.
13839 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13840 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13842 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13843 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13844 be reduced modulo `m`.
13846 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13849 The following entry accidentally appeared in the CHANGES file
13850 distributed with OpenSSL 0.9.7. The modifications described in
13851 it do *not* apply to OpenSSL 0.9.7.
13853 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13854 was actually never needed) and in BN_mul(). The removal in BN_mul()
13855 required a small change in bn_mul_part_recursive() and the addition
13856 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13857 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13858 bn_sub_words() and bn_add_words() except they take arrays with
13864 * In 'openssl passwd', verify passwords read from the terminal
13865 unless the '-salt' option is used (which usually means that
13866 verification would just waste user's time since the resulting
13867 hash is going to be compared with some given password hash)
13868 or the new '-noverify' option is used.
13870 This is an incompatible change, but it does not affect
13871 non-interactive use of 'openssl passwd' (passwords on the command
13872 line, '-stdin' option, '-in ...' option) and thus should not
13873 cause any problems.
13877 * Remove all references to RSAref, since there's no more need for it.
13881 * Make DSO load along a path given through an environment variable
13882 (SHLIB_PATH) with shl_load().
13886 * Constify the ENGINE code as a result of BIGNUM constification.
13887 Also constify the RSA code and most things related to it. In a
13888 few places, most notable in the depth of the ASN.1 code, ugly
13889 casts back to non-const were required (to be solved at a later
13894 * Make it so the openssl application has all engines loaded by default.
13898 * Constify the BIGNUM routines a little more.
13902 * Add the following functions:
13904 ENGINE_load_cswift()
13906 ENGINE_load_atalla()
13907 ENGINE_load_nuron()
13908 ENGINE_load_builtin_engines()
13910 That way, an application can itself choose if external engines that
13911 are built-in in OpenSSL shall ever be used or not. The benefit is
13912 that applications won't have to be linked with libdl or other dso
13913 libraries unless it's really needed.
13915 Changed 'openssl engine' to load all engines on demand.
13916 Changed the engine header files to avoid the duplication of some
13917 declarations (they differed!).
13921 * 'openssl engine' can now list capabilities.
13925 * Better error reporting in 'openssl engine'.
13929 * Never call load_dh_param(NULL) in s_server.
13933 * Add engine application. It can currently list engines by name and
13934 identity, and test if they are actually available.
13938 * Improve RPM specification file by forcing symbolic linking and making
13939 sure the installed documentation is also owned by root.root.
13941 *Damien Miller <djm@mindrot.org>*
13943 * Give the OpenSSL applications more possibilities to make use of
13944 keys (public as well as private) handled by engines.
13948 * Add OCSP code that comes from CertCo.
13952 * Add VMS support for the Rijndael code.
13956 * Added untested support for Nuron crypto accelerator.
13960 * Add support for external cryptographic devices. This code was
13961 previously distributed separately as the "engine" branch.
13963 *Geoff Thorpe, Richard Levitte*
13965 * Rework the filename-translation in the DSO code. It is now possible to
13966 have far greater control over how a "name" is turned into a filename
13967 depending on the operating environment and any oddities about the
13968 different shared library filenames on each system.
13972 * Support threads on FreeBSD-elf in Configure.
13976 * Fix for SHA1 assembly problem with MASM: it produces
13977 warnings about corrupt line number information when assembling
13978 with debugging information. This is caused by the overlapping
13981 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13984 NCONF_get_number() has no error checking at all. As a replacement,
13985 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13986 promoted strongly. The old NCONF_get_number is kept around for
13987 binary backward compatibility.
13988 Make it possible for methods to load from something other than a BIO,
13989 by providing a function pointer that is given a name instead of a BIO.
13990 For example, this could be used to load configuration data from an
13995 * Fix for non blocking accept BIOs. Added new I/O special reason
13996 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13997 with non blocking I/O was not possible because no retry code was
13998 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14003 * Added the beginnings of Rijndael support.
14007 * Fix for bug in DirectoryString mask setting. Add support for
14008 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14009 to allow certificate printing to more controllable, additional
14010 'certopt' option to 'x509' to allow new printing options to be
14015 * Clean old EAY MD5 hack from e_os.h.
14019 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14021 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14022 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14024 *Joe Orton, Steve Henson*
14026 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14028 * Fix additional bug revealed by the NISCC test suite:
14030 Stop bug triggering large recursion when presented with
14031 certain ASN.1 tags ([CVE-2003-0851])
14035 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14037 * Fix various bugs revealed by running the NISCC test suite:
14039 Stop out of bounds reads in the ASN1 code when presented with
14040 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14042 If verify callback ignores invalid public key errors don't try to check
14043 certificate signature with the NULL public key.
14047 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14048 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14053 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14054 extra data after the compression methods not only for TLS 1.0
14055 but also for SSL 3.0 (as required by the specification).
14057 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14059 * Change X509_certificate_type() to mark the key as exported/exportable
14060 when it's 512 *bits* long, not 512 bytes.
14064 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14066 * Countermeasure against the Klima-Pokorny-Rosa extension of
14067 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14068 a protocol version number mismatch like a decryption error
14069 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14073 * Turn on RSA blinding by default in the default implementation
14074 to avoid a timing attack. Applications that don't want it can call
14075 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14076 They would be ill-advised to do so in most cases.
14078 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14080 * Change RSA blinding code so that it works when the PRNG is not
14081 seeded (in this case, the secret RSA exponent is abused as
14082 an unpredictable seed -- if it is not unpredictable, there
14083 is no point in blinding anyway). Make RSA blinding thread-safe
14084 by remembering the creator's thread ID in rsa->blinding and
14085 having all other threads use local one-time blinding factors
14086 (this requires more computation than sharing rsa->blinding, but
14087 avoids excessive locking; and if an RSA object is not shared
14088 between threads, blinding will still be very fast).
14092 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14094 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14095 via timing by performing a MAC computation even if incorrect
14096 block cipher padding has been found. This is a countermeasure
14097 against active attacks where the attacker has to distinguish
14098 between bad padding and a MAC verification error. ([CVE-2003-0078])
14100 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14101 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14102 Martin Vuagnoux (EPFL, Ilion)*
14104 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14106 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14107 memory from its contents. This is done with a counter that will
14108 place alternating values in each byte. This can be used to solve
14109 two issues: 1) the removal of calls to memset() by highly optimizing
14110 compilers, and 2) cleansing with other values than 0, since those can
14111 be read through on certain media, for example a swap space on disk.
14115 * Bugfix: client side session caching did not work with external caching,
14116 because the session->cipher setting was not restored when reloading
14117 from the external cache. This problem was masked, when
14118 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14119 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14123 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14124 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14126 *Zeev Lieber <zeev-l@yahoo.com>*
14128 * Undo an undocumented change introduced in 0.9.6e which caused
14129 repeated calls to OpenSSL_add_all_ciphers() and
14130 OpenSSL_add_all_digests() to be ignored, even after calling
14135 * Change the default configuration reader to deal with last line not
14136 being properly terminated.
14140 * Change X509_NAME_cmp() so it applies the special rules on handling
14141 DN values that are of type PrintableString, as well as RDNs of type
14142 emailAddress where the value has the type ia5String.
14144 *stefank@valicert.com via Richard Levitte*
14146 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14147 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14148 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14149 the bitwise-OR of the two for use by the majority of applications
14150 wanting this behaviour, and update the docs. The documented
14151 behaviour and actual behaviour were inconsistent and had been
14152 changing anyway, so this is more a bug-fix than a behavioural
14155 *Geoff Thorpe, diagnosed by Nadav Har'El*
14157 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14158 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14162 * Fix initialization code race conditions in
14163 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14164 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14165 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14166 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14167 ssl2_get_cipher_by_char(),
14168 ssl3_get_cipher_by_char().
14170 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14172 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14173 the cached sessions are flushed, as the remove_cb() might use ex_data
14174 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14175 (see [openssl.org #212]).
14177 *Geoff Thorpe, Lutz Jaenicke*
14179 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14180 length, instead of the encoding length to d2i_ASN1_OBJECT.
14184 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14186 * [In 0.9.6g-engine release:]
14187 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14189 *Lynn Gazis <lgazis@rainbow.com>*
14191 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14193 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14194 and get fix the header length calculation.
14195 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14196 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14198 * Use proper error handling instead of 'assertions' in buffer
14199 overflow checks added in 0.9.6e. This prevents DoS (the
14200 assertions could call abort()).
14202 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14204 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14206 * Add various sanity checks to asn1_get_length() to reject
14207 the ASN1 length bytes if they exceed sizeof(long), will appear
14208 negative or the content length exceeds the length of the
14211 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14213 * Fix cipher selection routines: ciphers without encryption had no flags
14214 for the cipher strength set and where therefore not handled correctly
14215 by the selection routines (PR #130).
14219 * Fix EVP_dsa_sha macro.
14224 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14225 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14226 that was added in OpenSSL 0.9.6d.
14228 As the countermeasure turned out to be incompatible with some
14229 broken SSL implementations, the new option is part of SSL_OP_ALL.
14230 SSL_OP_ALL is usually employed when compatibility with weird SSL
14231 implementations is desired (e.g. '-bugs' option to 's_client' and
14232 's_server'), so the new option is automatically set in many
14237 * Changes in security patch:
14239 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14240 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14241 Air Force Materiel Command, USAF, under agreement number
14244 * Add various sanity checks to asn1_get_length() to reject
14245 the ASN1 length bytes if they exceed sizeof(long), will appear
14246 negative or the content length exceeds the length of the
14247 supplied buffer. ([CVE-2002-0659])
14249 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14251 * Assertions for various potential buffer overflows, not known to
14252 happen in practice.
14254 *Ben Laurie (CHATS)*
14256 * Various temporary buffers to hold ASCII versions of integers were
14257 too small for 64 bit platforms. ([CVE-2002-0655])
14258 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14260 * Remote buffer overflow in SSL3 protocol - an attacker could
14261 supply an oversized session ID to a client. ([CVE-2002-0656])
14263 *Ben Laurie (CHATS)*
14265 * Remote buffer overflow in SSL2 protocol - an attacker could
14266 supply an oversized client master key. ([CVE-2002-0656])
14268 *Ben Laurie (CHATS)*
14270 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14272 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14273 encoded as NULL) with id-dsa-with-sha1.
14275 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14277 * Check various `X509_...()` return values in `apps/req.c`.
14279 *Nils Larsch <nla@trustcenter.de>*
14281 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14282 an end-of-file condition would erroneously be flagged, when the CRLF
14283 was just at the end of a processed block. The bug was discovered when
14284 processing data through a buffering memory BIO handing the data to a
14285 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14286 <ptsekov@syntrex.com> and Nedelcho Stanev.
14290 * Implement a countermeasure against a vulnerability recently found
14291 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14292 before application data chunks to avoid the use of known IVs
14293 with data potentially chosen by the attacker.
14297 * Fix length checks in ssl3_get_client_hello().
14301 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14302 to prevent ssl3_read_internal() from incorrectly assuming that
14303 ssl3_read_bytes() found application data while handshake
14304 processing was enabled when in fact s->s3->in_read_app_data was
14305 merely automatically cleared during the initial handshake.
14307 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14309 * Fix object definitions for Private and Enterprise: they were not
14310 recognized in their shortname (=lowercase) representation. Extend
14311 obj_dat.pl to issue an error when using undefined keywords instead
14312 of silently ignoring the problem (Svenning Sorensen
14313 <sss@sss.dnsalias.net>).
14317 * Fix DH_generate_parameters() so that it works for 'non-standard'
14318 generators, i.e. generators other than 2 and 5. (Previously, the
14319 code did not properly initialise the 'add' and 'rem' values to
14320 BN_generate_prime().)
14322 In the new general case, we do not insist that 'generator' is
14323 actually a primitive root: This requirement is rather pointless;
14324 a generator of the order-q subgroup is just as good, if not
14329 * Map new X509 verification errors to alerts. Discovered and submitted by
14330 Tom Wu <tom@arcot.com>.
14334 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14335 returning non-zero before the data has been completely received
14336 when using non-blocking I/O.
14338 *Bodo Moeller; problem pointed out by John Hughes*
14340 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14342 *Ben Laurie, Lutz Jaenicke*
14344 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14345 Yoram Zahavi <YoramZ@gilian.com>).
14349 * Add information about CygWin 1.3 and on, and preserve proper
14350 configuration for the versions before that.
14352 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14354 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14355 check whether we deal with a copy of a session and do not delete from
14356 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14357 <izhar@checkpoint.com>.
14361 * Do not store session data into the internal session cache, if it
14362 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14363 flag is set). Proposed by Aslam <aslam@funk.com>.
14367 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14372 * [In 0.9.6d-engine release:]
14373 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14375 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14377 * Add the configuration target linux-s390x.
14379 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14381 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14382 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14383 variable as an indication that a ClientHello message has been
14384 received. As the flag value will be lost between multiple
14385 invocations of ssl3_accept when using non-blocking I/O, the
14386 function may not be aware that a handshake has actually taken
14387 place, thus preventing a new session from being added to the
14390 To avoid this problem, we now set s->new_session to 2 instead of
14391 using a local variable.
14393 *Lutz Jaenicke, Bodo Moeller*
14395 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14396 if the SSL_R_LENGTH_MISMATCH error is detected.
14398 *Geoff Thorpe, Bodo Moeller*
14400 * New 'shared_ldflag' column in Configure platform table.
14404 * Fix EVP_CIPHER_mode macro.
14406 *"Dan S. Camper" <dan@bti.net>*
14408 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14409 type, we must throw them away by setting rr->length to 0.
14411 *D P Chang <dpc@qualys.com>*
14413 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14415 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14416 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14417 worked incorrectly for those cases where range = `10..._2` and
14418 `3*range` is two bits longer than range.)
14422 * Only add signing time to PKCS7 structures if it is not already
14427 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14428 OBJ_ld_ce should be OBJ_id_ce.
14429 Also some ip-pda OIDs in crypto/objects/objects.txt were
14430 incorrect (cf. RFC 3039).
14432 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14434 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14435 returns early because it has nothing to do.
14437 *Andy Schneider <andy.schneider@bjss.co.uk>*
14439 * [In 0.9.6c-engine release:]
14440 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14442 *Andy Schneider <andy.schneider@bjss.co.uk>*
14444 * [In 0.9.6c-engine release:]
14445 Add support for Cryptographic Appliance's keyserver technology.
14446 (Use engine 'keyclient')
14448 *Cryptographic Appliances and Geoff Thorpe*
14450 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14451 is called via tools/c89.sh because arguments have to be
14452 rearranged (all '-L' options must appear before the first object
14455 *Richard Shapiro <rshapiro@abinitio.com>*
14457 * [In 0.9.6c-engine release:]
14458 Add support for Broadcom crypto accelerator cards, backported
14461 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14463 * [In 0.9.6c-engine release:]
14464 Add support for SureWare crypto accelerator cards from
14465 Baltimore Technologies. (Use engine 'sureware')
14467 *Baltimore Technologies and Mark Cox*
14469 * [In 0.9.6c-engine release:]
14470 Add support for crypto accelerator cards from Accelerated
14471 Encryption Processing, www.aep.ie. (Use engine 'aep')
14473 *AEP Inc. and Mark Cox*
14475 * Add a configuration entry for gcc on UnixWare.
14477 *Gary Benson <gbenson@redhat.com>*
14479 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14480 messages are stored in a single piece (fixed-length part and
14481 variable-length part combined) and fix various bugs found on the way.
14485 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14486 instead. BIO_gethostbyname() does not know what timeouts are
14487 appropriate, so entries would stay in cache even when they have
14489 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14491 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14492 faced with a pathologically small ClientHello fragment that does
14493 not contain client_version: Instead of aborting with an error,
14494 simply choose the highest available protocol version (i.e.,
14495 TLS 1.0 unless it is disabled). In practice, ClientHello
14496 messages are never sent like this, but this change gives us
14497 strictly correct behaviour at least for TLS.
14501 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14502 never resets s->method to s->ctx->method when called from within
14503 one of the SSL handshake functions.
14505 *Bodo Moeller; problem pointed out by Niko Baric*
14507 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14508 (sent using the client's version number) if client_version is
14509 smaller than the protocol version in use. Also change
14510 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14511 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14512 the client will at least see that alert.
14516 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14521 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14522 client receives HelloRequest while in a handshake.
14524 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14526 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14527 should end in 'break', not 'goto end' which circumvents various
14528 cleanups done in state SSL_ST_OK. But session related stuff
14529 must be disabled for SSL_ST_OK in the case that we just sent a
14532 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14533 before just sending a HelloRequest.
14535 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14537 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14538 reveal whether illegal block cipher padding was found or a MAC
14539 verification error occurred. (Neither SSLerr() codes nor alerts
14540 are directly visible to potential attackers, but the information
14541 may leak via logfiles.)
14543 Similar changes are not required for the SSL 2.0 implementation
14544 because the number of padding bytes is sent in clear for SSL 2.0,
14545 and the extra bytes are just ignored. However ssl/s2_pkt.c
14546 failed to verify that the purported number of padding bytes is in
14551 * Add OpenUNIX-8 support including shared libraries
14552 (Boyd Lynn Gerber <gerberb@zenez.com>).
14556 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14557 'wristwatch attack' using huge encoding parameters (cf.
14558 James H. Manger's CRYPTO 2001 paper). Note that the
14559 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14560 encoding parameters and hence was not vulnerable.
14564 * BN_sqr() bug fix.
14566 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14568 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14569 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14570 followed by modular reduction.
14572 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14574 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14575 equivalent based on BN_pseudo_rand() instead of BN_rand().
14579 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14580 This function was broken, as the check for a new client hello message
14581 to handle SGC did not allow these large messages.
14582 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14586 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14590 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14591 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14595 * Rework the configuration and shared library support for Tru64 Unix.
14596 The configuration part makes use of modern compiler features and
14597 still retains old compiler behavior for those that run older versions
14598 of the OS. The shared library support part includes a variant that
14599 uses the RPATH feature, and is available through the special
14600 configuration target "alpha-cc-rpath", which will never be selected
14603 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14605 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14606 with the same message size as in ssl3_get_certificate_request().
14607 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14608 messages might inadvertently be reject as too long.
14610 *Petr Lampa <lampa@fee.vutbr.cz>*
14612 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14616 * Modified SSL library such that the verify_callback that has been set
14617 specifically for an SSL object with SSL_set_verify() is actually being
14618 used. Before the change, a verify_callback set with this function was
14619 ignored and the verify_callback() set in the SSL_CTX at the time of
14620 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14621 to allow the necessary settings.
14625 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14626 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14627 done automatically (in contradiction to the requirements of the C
14628 standard). This made problems when used from OpenSSH.
14632 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14633 dh->length and always used
14635 BN_rand_range(priv_key, dh->p).
14637 BN_rand_range() is not necessary for Diffie-Hellman, and this
14638 specific range makes Diffie-Hellman unnecessarily inefficient if
14639 dh->length (recommended exponent length) is much smaller than the
14640 length of dh->p. We could use BN_rand_range() if the order of
14641 the subgroup was stored in the DH structure, but we only have
14646 BN_rand(priv_key, l, ...)
14648 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14655 RSA_eay_public_encrypt
14656 RSA_eay_private_decrypt
14657 RSA_eay_private_encrypt (signing)
14658 RSA_eay_public_decrypt (signature verification)
14660 (default implementations for RSA_public_encrypt,
14661 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14662 always reject numbers >= n.
14666 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14667 to synchronize access to 'locking_thread'. This is necessary on
14668 systems where access to 'locking_thread' (an 'unsigned long'
14669 variable) is not atomic.
14673 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14674 *before* setting the 'crypto_lock_rand' flag. The previous code had
14675 a race condition if 0 is a valid thread ID.
14677 *Travis Vitek <vitek@roguewave.com>*
14679 * Add support for shared libraries under Irix.
14681 *Albert Chin-A-Young <china@thewrittenword.com>*
14683 * Add configuration option to build on Linux on both big-endian and
14684 little-endian MIPS.
14686 *Ralf Baechle <ralf@uni-koblenz.de>*
14688 * Add the possibility to create shared libraries on HP-UX.
14692 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14694 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14695 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14696 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14697 PRNG state recovery was possible based on the output of
14698 one PRNG request appropriately sized to gain knowledge on
14699 'md' followed by enough consecutive 1-byte PRNG requests
14700 to traverse all of 'state'.
14702 1. When updating 'md_local' (the current thread's copy of 'md')
14703 during PRNG output generation, hash all of the previous
14704 'md_local' value, not just the half used for PRNG output.
14706 2. Make the number of bytes from 'state' included into the hash
14707 independent from the number of PRNG bytes requested.
14709 The first measure alone would be sufficient to avoid
14710 Markku-Juhani's attack. (Actually it had never occurred
14711 to me that the half of 'md_local' used for chaining was the
14712 half from which PRNG output bytes were taken -- I had always
14713 assumed that the secret half would be used.) The second
14714 measure makes sure that additional data from 'state' is never
14715 mixed into 'md_local' in small portions; this heuristically
14716 further strengthens the PRNG.
14720 * Fix crypto/bn/asm/mips3.s.
14724 * When only the key is given to "enc", the IV is undefined. Print out
14725 an error message in this case.
14729 * Handle special case when X509_NAME is empty in X509 printing routines.
14733 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14734 positive and less than q.
14738 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14739 used: it isn't thread safe and the add_lock_callback should handle
14742 *Paul Rose <Paul.Rose@bridge.com>*
14744 * Verify that incoming data obeys the block size in
14745 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14751 *Ulf Möller, Bodo Möller*
14753 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14754 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14755 when fixing the server behaviour for backwards-compatible 'client
14756 hello' messages. (Note that the attack is impractical against
14757 SSL 3.0 and TLS 1.0 anyway because length and version checking
14758 means that the probability of guessing a valid ciphertext is
14759 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14762 Before 0.9.5, the countermeasure (hide the error by generating a
14763 random 'decryption result') did not work properly because
14764 ERR_clear_error() was missing, meaning that SSL_get_error() would
14765 detect the supposedly ignored error.
14767 Both problems are now fixed.
14771 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14772 (previously it was 1024).
14776 * Fix for compatibility mode trust settings: ignore trust settings
14777 unless some valid trust or reject settings are present.
14781 * Fix for blowfish EVP: its a variable length cipher.
14785 * Fix various bugs related to DSA S/MIME verification. Handle missing
14786 parameters in DSA public key structures and return an error in the
14787 DSA routines if parameters are absent.
14791 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14792 in the current directory if neither $RANDFILE nor $HOME was set.
14793 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14794 caused some confusion to Windows users who haven't defined $HOME.
14795 Thus RAND_file_name() is changed again: e_os.h can define a
14796 DEFAULT_HOME, which will be used if $HOME is not set.
14797 For Windows, we use "C:"; on other platforms, we still require
14798 environment variables.
14800 * Move 'if (!initialized) RAND_poll()' into regions protected by
14801 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14802 having multiple threads call RAND_poll() concurrently.
14806 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14807 combination of a flag and a thread ID variable.
14808 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14809 flag), *other* threads can enter ssleay_add_bytes without obeying
14810 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14811 that they do not hold after the first thread unsets add_do_not_lock).
14815 * Change bctest again: '-x' expressions are not available in all
14816 versions of 'test'.
14820 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14822 * Fix a couple of memory leaks in PKCS7_dataDecode()
14824 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14826 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14827 the default extension for executables, if any. Also, make the perl
14828 scripts that use symlink() to test if it really exists and use "cp"
14829 if it doesn't. All this made OpenSSL compilable and installable in
14834 * Fix for asn1_GetSequence() for indefinite length constructed data.
14835 If SEQUENCE is length is indefinite just set c->slen to the total
14836 amount of data available.
14838 *Steve Henson, reported by shige@FreeBSD.org*
14840 *This change does not apply to 0.9.7.*
14842 * Change bctest to avoid here-documents inside command substitution
14843 (workaround for FreeBSD /bin/sh bug).
14844 For compatibility with Ultrix, avoid shell functions (introduced
14845 in the bctest version that searches along $PATH).
14849 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14850 with des_encrypt() defined on some operating systems, like Solaris
14855 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14856 On the Importance of Eliminating Errors in Cryptographic
14857 Computations, J. Cryptology 14 (2001) 2, 101-119,
14858 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14862 * MIPS assembler BIGNUM division bug fix.
14866 * Disabled incorrect Alpha assembler code.
14870 * Fix PKCS#7 decode routines so they correctly update the length
14871 after reading an EOC for the EXPLICIT tag.
14875 *This change does not apply to 0.9.7.*
14877 * Fix bug in PKCS#12 key generation routines. This was triggered
14878 if a 3DES key was generated with a 0 initial byte. Include
14879 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14880 (but broken) behaviour.
14884 * Enhance bctest to search for a working bc along $PATH and print
14887 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14889 * Fix memory leaks in err.c: free err_data string if necessary;
14890 don't write to the wrong index in ERR_set_error_data.
14894 * Implement ssl23_peek (analogous to ssl23_read), which previously
14899 * Replace rdtsc with `_emit` statements for VC++ version 5.
14901 *Jeremy Cooper <jeremy@baymoo.org>*
14903 * Make it possible to reuse SSLv2 sessions.
14907 * In copy_email() check for >= 0 as a return value for
14908 X509_NAME_get_index_by_NID() since 0 is a valid index.
14910 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14912 * Avoid coredump with unsupported or invalid public keys by checking if
14913 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14914 PKCS7_verify() fails with non detached data.
14918 * Don't use getenv in library functions when run as setuid/setgid.
14919 New function OPENSSL_issetugid().
14923 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14924 due to incorrect handling of multi-threading:
14926 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14928 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14930 3. Count how many times MemCheck_off() has been called so that
14931 nested use can be treated correctly. This also avoids
14932 inband-signalling in the previous code (which relied on the
14933 assumption that thread ID 0 is impossible).
14937 * Add "-rand" option also to s_client and s_server.
14941 * Fix CPU detection on Irix 6.x.
14942 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14943 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14945 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14950 *This change does not apply to 0.9.7.*
14952 * Use the cached encoding of an X509_NAME structure rather than
14953 copying it. This is apparently the reason for the libsafe "errors"
14954 but the code is actually correct.
14958 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14959 Bleichenbacher's DSA attack.
14960 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14961 to be set and top=0 forces the highest bit to be set; top=-1 is new
14962 and leaves the highest bit random.
14964 *Ulf Moeller, Bodo Moeller*
14966 * In the `NCONF_...`-based implementations for `CONF_...` queries
14967 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14968 a temporary CONF structure with the data component set to NULL
14969 (which gives segmentation faults in lh_retrieve).
14970 Instead, use NULL for the CONF pointer in CONF_get_string and
14971 CONF_get_number (which may use environment variables) and directly
14972 return NULL from CONF_get_section.
14976 * Fix potential buffer overrun for EBCDIC.
14980 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14981 keyUsage if basicConstraints absent for a CA.
14985 * Make SMIME_write_PKCS7() write mail header values with a format that
14986 is more generally accepted (no spaces before the semicolon), since
14987 some programs can't parse those values properly otherwise. Also make
14988 sure BIO's that break lines after each write do not create invalid
14993 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14994 macros previously used would not encode an empty SEQUENCE OF
14995 and break the signature.
14999 *This change does not apply to 0.9.7.*
15001 * Zero the premaster secret after deriving the master secret in
15006 * Add some EVP_add_digest_alias registrations (as found in
15007 OpenSSL_add_all_digests()) to SSL_library_init()
15008 aka OpenSSL_add_ssl_algorithms(). This provides improved
15009 compatibility with peers using X.509 certificates
15010 with unconventional AlgorithmIdentifier OIDs.
15014 * Fix for Irix with NO_ASM.
15016 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15018 * ./config script fixes.
15020 *Ulf Moeller, Richard Levitte*
15022 * Fix 'openssl passwd -1'.
15026 * Change PKCS12_key_gen_asc() so it can cope with non null
15027 terminated strings whose length is passed in the passlen
15028 parameter, for example from PEM callbacks. This was done
15029 by adding an extra length parameter to asc2uni().
15031 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15033 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15034 call failed, free the DSA structure.
15038 * Fix to uni2asc() to cope with zero length Unicode strings.
15039 These are present in some PKCS#12 files.
15043 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15044 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15045 when writing a 32767 byte record.
15047 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15049 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15050 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15052 (RSA objects have a reference count access to which is protected
15053 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15054 so they are meant to be shared between threads.)
15055 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15056 "Reddie, Steven" <Steven.Reddie@ca.com>*
15058 * Fix a deadlock in CRYPTO_mem_leaks().
15062 * Use better test patterns in bntest.
15066 * rand_win.c fix for Borland C.
15070 * BN_rshift bugfix for n == 0.
15074 * Add a 'bctest' script that checks for some known 'bc' bugs
15075 so that 'make test' does not abort just because 'bc' is broken.
15079 * Store verify_result within SSL_SESSION also for client side to
15080 avoid potential security hole. (Re-used sessions on the client side
15081 always resulted in verify_result==X509_V_OK, not using the original
15082 result of the server certificate verification.)
15086 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15087 SSL3_RT_APPLICATION_DATA, return 0.
15088 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15093 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15094 releases, have been re-implemented by renaming the previous
15095 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15096 and ssl3_read_internal, respectively, and adding 'peek' parameters
15097 to them. The new ssl[23]_{read,peek} functions are calls to
15098 ssl[23]_read_internal with the 'peek' flag set appropriately.
15099 A 'peek' parameter has also been added to ssl3_read_bytes, which
15100 does the actual work for ssl3_read_internal.
15104 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15105 the method-specific "init()" handler. Also clean up ex_data after
15106 calling the method-specific "finish()" handler. Previously, this was
15107 happening the other way round.
15111 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15112 The previous value, 12, was not always sufficient for BN_mod_exp().
15116 * Make sure that shared libraries get the internal name engine with
15117 the full version number and not just 0. This should mark the
15118 shared libraries as not backward compatible. Of course, this should
15119 be changed again when we can guarantee backward binary compatibility.
15123 * Fix typo in get_cert_by_subject() in by_dir.c
15125 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15127 * Rework the system to generate shared libraries:
15129 - Make note of the expected extension for the shared libraries and
15130 if there is a need for symbolic links from for example libcrypto.so.0
15131 to libcrypto.so.0.9.7. There is extended info in Configure for
15134 - Make as few rebuilds of the shared libraries as possible.
15136 - Still avoid linking the OpenSSL programs with the shared libraries.
15138 - When installing, install the shared libraries separately from the
15143 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15145 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15146 and not in SSL_clear because the latter is also used by the
15147 accept/connect functions; previously, the settings made by
15148 SSL_set_read_ahead would be lost during the handshake.
15150 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15152 * Correct util/mkdef.pl to be selective about disabled algorithms.
15153 Previously, it would create entries for disabled algorithms no
15158 * Added several new manual pages for SSL_* function.
15162 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15164 * In ssl23_get_client_hello, generate an error message when faced
15165 with an initial SSL 3.0/TLS record that is too small to contain the
15166 first two bytes of the ClientHello message, i.e. client_version.
15167 (Note that this is a pathologic case that probably has never happened
15168 in real life.) The previous approach was to use the version number
15169 from the record header as a substitute; but our protocol choice
15170 should not depend on that one because it is not authenticated
15171 by the Finished messages.
15175 * More robust randomness gathering functions for Windows.
15177 *Jeffrey Altman <jaltman@columbia.edu>*
15179 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15180 not set then we don't setup the error code for issuer check errors
15181 to avoid possibly overwriting other errors which the callback does
15182 handle. If an application does set the flag then we assume it knows
15183 what it is doing and can handle the new informational codes
15188 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15189 a general "ANY" type, as such it should be able to decode anything
15190 including tagged types. However it didn't check the class so it would
15191 wrongly interpret tagged types in the same way as their universal
15192 counterpart and unknown types were just rejected. Changed so that the
15193 tagged and unknown types are handled in the same way as a SEQUENCE:
15194 that is the encoding is stored intact. There is also a new type
15195 "V_ASN1_OTHER" which is used when the class is not universal, in this
15196 case we have no idea what the actual type is so we just lump them all
15201 * On VMS, stdout may very well lead to a file that is written to
15202 in a record-oriented fashion. That means that every write() will
15203 write a separate record, which will be read separately by the
15204 programs trying to read from it. This can be very confusing.
15206 The solution is to put a BIO filter in the way that will buffer
15207 text until a linefeed is reached, and then write everything a
15208 line at a time, so every record written will be an actual line,
15209 not chunks of lines and not (usually doesn't happen, but I've
15210 seen it once) several lines in one record. BIO_f_linebuffer() is
15213 Currently, it's a VMS-only method, because that's where it has
15214 been tested well enough.
15218 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15219 it can return incorrect results.
15220 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15221 but it was in 0.9.6-beta[12].)
15225 * Disable the check for content being present when verifying detached
15226 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15227 include zero length content when signing messages.
15231 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15232 BIO_ctrl (for BIO pairs).
15236 * Add DSO method for VMS.
15240 * Bug fix: Montgomery multiplication could produce results with the
15245 * Add RPM specification openssl.spec and modify it to build three
15246 packages. The default package contains applications, application
15247 documentation and run-time libraries. The devel package contains
15248 include files, static libraries and function documentation. The
15249 doc package contains the contents of the doc directory. The original
15250 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15254 * Add a large number of documentation files for many SSL routines.
15256 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15258 * Add a configuration entry for Sony News 4.
15260 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15262 * Don't set the two most significant bits to one when generating a
15263 random number < q in the DSA library.
15267 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15268 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15269 the underlying transport is blocking) if a handshake took place.
15270 (The default behaviour is needed by applications such as s_client
15271 and s_server that use select() to determine when to use SSL_read;
15272 but for applications that know in advance when to expect data, it
15273 just makes things more complicated.)
15277 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15282 * Add a few more EBCDIC conditionals that make `req` and `x509`
15283 work better on such systems.
15285 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15287 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15288 Update PKCS12_parse() so it copies the friendlyName and the
15289 keyid to the certificates aux info.
15293 * Fix bug in PKCS7_verify() which caused an infinite loop
15294 if there was more than one signature.
15296 *Sven Uszpelkat <su@celocom.de>*
15298 * Major change in util/mkdef.pl to include extra information
15299 about each symbol, as well as presenting variables as well
15300 as functions. This change means that there's n more need
15301 to rebuild the .num files when some algorithms are excluded.
15305 * Allow the verify time to be set by an application,
15306 rather than always using the current time.
15310 * Phase 2 verify code reorganisation. The certificate
15311 verify code now looks up an issuer certificate by a
15312 number of criteria: subject name, authority key id
15313 and key usage. It also verifies self signed certificates
15314 by the same criteria. The main comparison function is
15315 X509_check_issued() which performs these checks.
15317 Lot of changes were necessary in order to support this
15318 without completely rewriting the lookup code.
15320 Authority and subject key identifier are now cached.
15322 The LHASH 'certs' is X509_STORE has now been replaced
15323 by a STACK_OF(X509_OBJECT). This is mainly because an
15324 LHASH can't store or retrieve multiple objects with
15325 the same hash value.
15327 As a result various functions (which were all internal
15328 use only) have changed to handle the new X509_STORE
15329 structure. This will break anything that messed round
15330 with X509_STORE internally.
15332 The functions X509_STORE_add_cert() now checks for an
15333 exact match, rather than just subject name.
15335 The X509_STORE API doesn't directly support the retrieval
15336 of multiple certificates matching a given criteria, however
15337 this can be worked round by performing a lookup first
15338 (which will fill the cache with candidate certificates)
15339 and then examining the cache for matches. This is probably
15340 the best we can do without throwing out X509_LOOKUP
15341 entirely (maybe later...).
15343 The X509_VERIFY_CTX structure has been enhanced considerably.
15345 All certificate lookup operations now go via a get_issuer()
15346 callback. Although this currently uses an X509_STORE it
15347 can be replaced by custom lookups. This is a simple way
15348 to bypass the X509_STORE hackery necessary to make this
15349 work and makes it possible to use more efficient techniques
15350 in future. A very simple version which uses a simple
15351 STACK for its trusted certificate store is also provided
15352 using X509_STORE_CTX_trusted_stack().
15354 The verify_cb() and verify() callbacks now have equivalents
15355 in the X509_STORE_CTX structure.
15357 X509_STORE_CTX also has a 'flags' field which can be used
15358 to customise the verify behaviour.
15362 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15363 excludes S/MIME capabilities.
15367 * When a certificate request is read in keep a copy of the
15368 original encoding of the signed data and use it when outputting
15369 again. Signatures then use the original encoding rather than
15370 a decoded, encoded version which may cause problems if the
15371 request is improperly encoded.
15375 * For consistency with other BIO_puts implementations, call
15376 buffer_write(b, ...) directly in buffer_puts instead of calling
15379 In BIO_puts, increment b->num_write as in BIO_write.
15381 *Peter.Sylvester@EdelWeb.fr*
15383 * Fix BN_mul_word for the case where the word is 0. (We have to use
15384 BN_zero, we may not return a BIGNUM with an array consisting of
15385 words set to zero.)
15389 * Avoid calling abort() from within the library when problems are
15390 detected, except if preprocessor symbols have been defined
15391 (such as REF_CHECK, BN_DEBUG etc.).
15395 * New openssl application 'rsautl'. This utility can be
15396 used for low-level RSA operations. DER public key
15397 BIO/fp routines also added.
15401 * New Configure entry and patches for compiling on QNX 4.
15403 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15405 * A demo state-machine implementation was sponsored by
15406 Nuron (<http://www.nuron.com/>) and is now available in
15407 demos/state_machine.
15411 * New options added to the 'dgst' utility for signature
15412 generation and verification.
15416 * Unrecognized PKCS#7 content types are now handled via a
15417 catch all ASN1_TYPE structure. This allows unsupported
15418 types to be stored as a "blob" and an application can
15419 encode and decode it manually.
15423 * Fix various signed/unsigned issues to make a_strex.c
15424 compile under VC++.
15426 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15428 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15429 length if passed a buffer. ASN1_INTEGER_to_BN failed
15430 if passed a NULL BN and its argument was negative.
15432 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15434 * Modification to PKCS#7 encoding routines to output definite
15435 length encoding. Since currently the whole structures are in
15436 memory there's not real point in using indefinite length
15437 constructed encoding. However if OpenSSL is compiled with
15438 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15442 * Added BIO_vprintf() and BIO_vsnprintf().
15446 * Added more prefixes to parse for in the strings written
15447 through a logging bio, to cover all the levels that are available
15448 through syslog. The prefixes are now:
15450 PANIC, EMERG, EMR => LOG_EMERG
15451 ALERT, ALR => LOG_ALERT
15452 CRIT, CRI => LOG_CRIT
15453 ERROR, ERR => LOG_ERR
15454 WARNING, WARN, WAR => LOG_WARNING
15455 NOTICE, NOTE, NOT => LOG_NOTICE
15456 INFO, INF => LOG_INFO
15457 DEBUG, DBG => LOG_DEBUG
15459 and as before, if none of those prefixes are present at the
15460 beginning of the string, LOG_ERR is chosen.
15462 On Win32, the `LOG_*` levels are mapped according to this:
15464 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15465 LOG_WARNING => EVENTLOG_WARNING_TYPE
15466 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15470 * Made it possible to reconfigure with just the configuration
15471 argument "reconf" or "reconfigure". The command line arguments
15472 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15473 and are retrieved from there when reconfiguring.
15479 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15481 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15485 * The obj_dat.pl script was messing up the sorting of object
15486 names. The reason was that it compared the quoted version
15487 of strings as a result "OCSP" > "OCSP Signing" because
15488 " > SPACE. Changed script to store unquoted versions of
15489 names and add quotes on output. It was also omitting some
15490 names from the lookup table if they were given a default
15491 value (that is if SN is missing it is given the same
15492 value as LN and vice versa), these are now added on the
15493 grounds that if an object has a name we should be able to
15494 look it up. Finally added warning output when duplicate
15495 short or long names are found.
15499 * Changes needed for Tandem NSK.
15501 *Scott Uroff <scott@xypro.com>*
15503 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15504 RSA_padding_check_SSLv23(), special padding was never detected
15505 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15506 version rollback attacks was not effective.
15508 In s23_clnt.c, don't use special rollback-attack detection padding
15509 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15510 client; similarly, in s23_srvr.c, don't do the rollback check if
15511 SSL 2.0 is the only protocol enabled in the server.
15515 * Make it possible to get hexdumps of unprintable data with 'openssl
15516 asn1parse'. By implication, the functions ASN1_parse_dump() and
15517 BIO_dump_indent() are added.
15521 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15522 these print out strings and name structures based on various
15523 flags including RFC2253 support and proper handling of
15524 multibyte characters. Added options to the 'x509' utility
15525 to allow the various flags to be set.
15529 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15530 Also change the functions X509_cmp_current_time() and
15531 X509_gmtime_adj() work with an ASN1_TIME structure,
15532 this will enable certificates using GeneralizedTime in validity
15533 dates to be checked.
15537 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15538 negative public key encodings) on by default,
15539 NO_NEG_PUBKEY_BUG can be set to disable it.
15543 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15544 content octets. An i2c_ASN1_OBJECT is unnecessary because
15545 the encoding can be trivially obtained from the structure.
15549 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15550 not read locks (`CRYPTO_r_[un]lock`).
15554 * A first attempt at creating official support for shared
15555 libraries through configuration. I've kept it so the
15556 default is static libraries only, and the OpenSSL programs
15557 are always statically linked for now, but there are
15558 preparations for dynamic linking in place.
15559 This has been tested on Linux and Tru64.
15563 * Randomness polling function for Win9x, as described in:
15564 Peter Gutmann, Software Generation of Practically Strong
15569 * Fix so PRNG is seeded in req if using an already existing
15574 * New options to smime application. -inform and -outform
15575 allow alternative formats for the S/MIME message including
15576 PEM and DER. The -content option allows the content to be
15577 specified separately. This should allow things like Netscape
15578 form signing output easier to verify.
15582 * Fix the ASN1 encoding of tags using the 'long form'.
15586 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15587 STRING types. These convert content octets to and from the
15588 underlying type. The actual tag and length octets are
15589 already assumed to have been read in and checked. These
15590 are needed because all other string types have virtually
15591 identical handling apart from the tag. By having versions
15592 of the ASN1 functions that just operate on content octets
15593 IMPLICIT tagging can be handled properly. It also allows
15594 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15595 and ASN1_INTEGER are identical apart from the tag.
15599 * Change the handling of OID objects as follows:
15601 - New object identifiers are inserted in objects.txt, following
15602 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15603 - objects.pl is used to process obj_mac.num and create a new
15605 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15608 This is currently kind of a hack, and the perl code in objects.pl
15609 isn't very elegant, but it works as I intended. The simplest way
15610 to check that it worked correctly is to look in obj_dat.h and
15611 check the array nid_objs and make sure the objects haven't moved
15612 around (this is important!). Additions are OK, as well as
15613 consistent name changes.
15617 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15621 * Addition of the command line parameter '-rand file' to 'openssl req'.
15622 The given file adds to whatever has already been seeded into the
15623 random pool through the RANDFILE configuration file option or
15624 environment variable, or the default random state file.
15628 * mkstack.pl now sorts each macro group into lexical order.
15629 Previously the output order depended on the order the files
15630 appeared in the directory, resulting in needless rewriting
15635 * Patches to make OpenSSL compile under Win32 again. Mostly
15636 work arounds for the VC++ problem that it treats func() as
15637 func(void). Also stripped out the parts of mkdef.pl that
15638 added extra typesafe functions: these no longer exist.
15642 * Reorganisation of the stack code. The macros are now all
15643 collected in safestack.h . Each macro is defined in terms of
15644 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15645 DEBUG_SAFESTACK is now handled in terms of function casts,
15646 this has the advantage of retaining type safety without the
15647 use of additional functions. If DEBUG_SAFESTACK is not defined
15648 then the non typesafe macros are used instead. Also modified the
15649 mkstack.pl script to handle the new form. Needs testing to see
15650 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15651 the default if no major problems. Similar behaviour for ASN1_SET_OF
15652 and PKCS12_STACK_OF.
15656 * When some versions of IIS use the 'NET' form of private key the
15657 key derivation algorithm is different. Normally MD5(password) is
15658 used as a 128 bit RC4 key. In the modified case
15659 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15660 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15661 as the old Netscape_RSA functions except they have an additional
15662 'sgckey' parameter which uses the modified algorithm. Also added
15663 an -sgckey command line option to the rsa utility. Thanks to
15664 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15665 algorithm to openssl-dev.
15669 * The evp_local.h macros were using 'c.##kname' which resulted in
15670 invalid expansion on some systems (SCO 5.0.5 for example).
15671 Corrected to 'c.kname'.
15673 *Phillip Porch <root@theporch.com>*
15675 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15676 a STACK of email addresses from a certificate or request, these look
15677 in the subject name and the subject alternative name extensions and
15678 omit any duplicate addresses.
15682 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15683 This makes DSA verification about 2 % faster.
15687 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15688 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15689 plus overhead for 1024 bit moduli).
15690 This makes exponentiations about 0.5 % faster for 1024 bit
15691 exponents (as measured by "openssl speed rsa2048").
15695 * Rename memory handling macros to avoid conflicts with other
15697 Malloc => OPENSSL_malloc
15698 Malloc_locked => OPENSSL_malloc_locked
15699 Realloc => OPENSSL_realloc
15700 Free => OPENSSL_free
15704 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15705 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15709 * CygWin32 support.
15711 *John Jarvie <jjarvie@newsguy.com>*
15713 * The type-safe stack code has been rejigged. It is now only compiled
15714 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15715 by default all type-specific stack functions are "#define"d back to
15716 standard stack functions. This results in more streamlined output
15717 but retains the type-safety checking possibilities of the original
15722 * The STACK code has been cleaned up, and certain type declarations
15723 that didn't make a lot of sense have been brought in line. This has
15724 also involved a cleanup of sorts in safestack.h to more correctly
15725 map type-safe stack functions onto their plain stack counterparts.
15726 This work has also resulted in a variety of "const"ifications of
15727 lots of the code, especially `_cmp` operations which should normally
15728 be prototyped with "const" parameters anyway.
15732 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15733 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15734 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15735 where all of 'md' is used each time the PRNG is used, but 'state'
15736 is used only indexed by a cyclic counter. As entropy may not be
15737 well distributed from the beginning, 'md' is important as a
15738 chaining variable. However, the output function chains only half
15739 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15740 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15741 in all of 'state' being rewritten, with the new values depending
15742 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15746 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15747 the handshake is continued after ssl_verify_cert_chain();
15748 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15749 can lead to 'unexplainable' connection aborts later.
15751 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15753 * Major EVP API cipher revision.
15754 Add hooks for extra EVP features. This allows various cipher
15755 parameters to be set in the EVP interface. Support added for variable
15756 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15757 setting of RC2 and RC5 parameters.
15759 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15762 Remove lots of duplicated code from the EVP library. For example *every*
15763 cipher init() function handles the 'iv' in the same way according to the
15764 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15765 for CFB and OFB modes they zero ctx->num.
15767 New functionality allows removal of S/MIME code RC2 hack.
15769 Most of the routines have the same form and so can be declared in terms
15772 By shifting this to the top level EVP_CipherInit() it can be removed from
15773 all individual ciphers. If the cipher wants to handle IVs or keys
15774 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15777 Change lots of functions like EVP_EncryptUpdate() to now return a
15778 value: although software versions of the algorithms cannot fail
15779 any installed hardware versions can.
15783 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15784 this option is set, tolerate broken clients that send the negotiated
15785 protocol version number instead of the requested protocol version
15790 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15791 i.e. non-zero for export ciphersuites, zero otherwise.
15792 Previous versions had this flag inverted, inconsistent with
15793 rsa_tmp_cb (..._TMP_RSA_CB).
15795 *Bodo Moeller; problem reported by Amit Chopra*
15797 * Add missing DSA library text string. Work around for some IIS
15798 key files with invalid SEQUENCE encoding.
15802 * Add a document (doc/standards.txt) that list all kinds of standards
15803 and so on that are implemented in OpenSSL.
15807 * Enhance c_rehash script. Old version would mishandle certificates
15808 with the same subject name hash and wouldn't handle CRLs at all.
15809 Added -fingerprint option to crl utility, to support new c_rehash
15814 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15818 * Fix for SSL server purpose checking. Server checking was
15819 rejecting certificates which had extended key usage present
15820 but no ssl client purpose.
15822 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15824 * Make PKCS#12 code work with no password. The PKCS#12 spec
15825 is a little unclear about how a blank password is handled.
15826 Since the password in encoded as a BMPString with terminating
15827 double NULL a zero length password would end up as just the
15828 double NULL. However no password at all is different and is
15829 handled differently in the PKCS#12 key generation code. NS
15830 treats a blank password as zero length. MSIE treats it as no
15831 password on export: but it will try both on import. We now do
15832 the same: PKCS12_parse() tries zero length and no password if
15833 the password is set to "" or NULL (NULL is now a valid password:
15834 it wasn't before) as does the pkcs12 application.
15838 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15839 perror when PEM_read_bio_X509_REQ fails, the error message must
15840 be obtained from the error queue.
15844 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15845 it in ERR_remove_state if appropriate, and change ERR_get_state
15846 accordingly to avoid race conditions (this is necessary because
15847 thread_hash is no longer constant once set).
15851 * Bugfix for linux-elf makefile.one.
15855 * RSA_get_default_method() will now cause a default
15856 RSA_METHOD to be chosen if one doesn't exist already.
15857 Previously this was only set during a call to RSA_new()
15858 or RSA_new_method(NULL) meaning it was possible for
15859 RSA_get_default_method() to return NULL.
15863 * Added native name translation to the existing DSO code
15864 that will convert (if the flag to do so is set) filenames
15865 that are sufficiently small and have no path information
15866 into a canonical native form. Eg. "blah" converted to
15867 "libblah.so" or "blah.dll" etc.
15871 * New function ERR_error_string_n(e, buf, len) which is like
15872 ERR_error_string(e, buf), but writes at most 'len' bytes
15873 including the 0 terminator. For ERR_error_string_n, 'buf'
15876 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15878 * CONF library reworked to become more general. A new CONF
15879 configuration file reader "class" is implemented as well as a
15880 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15881 old `CONF_*` functions are still there, but are reimplemented to
15882 work in terms of the new functions. Also, a set of functions
15883 to handle the internal storage of the configuration data is
15884 provided to make it easier to write new configuration file
15885 reader "classes" (I can definitely see something reading a
15886 configuration file in XML format, for example), called `_CONF_*`,
15887 or "the configuration storage API"...
15889 The new configuration file reading functions are:
15891 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15892 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15894 NCONF_default, NCONF_WIN32
15896 NCONF_dump_fp, NCONF_dump_bio
15898 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15899 NCONF_new creates a new CONF object. This works in the same way
15900 as other interfaces in OpenSSL, like the BIO interface.
15901 `NCONF_dump_*` dump the internal storage of the configuration file,
15902 which is useful for debugging. All other functions take the same
15903 arguments as the old `CONF_*` functions with the exception of the
15904 first that must be a `CONF *` instead of a `LHASH *`.
15906 To make it easier to use the new classes with the old `CONF_*` functions,
15907 the function CONF_set_default_method is provided.
15911 * Add '-tls1' option to 'openssl ciphers', which was already
15912 mentioned in the documentation but had not been implemented.
15913 (This option is not yet really useful because even the additional
15914 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15918 * Initial DSO code added into libcrypto for letting OpenSSL (and
15919 OpenSSL-based applications) load shared libraries and bind to
15920 them in a portable way.
15922 *Geoff Thorpe, with contributions from Richard Levitte*
15924 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15926 * Make sure _lrotl and _lrotr are only used with MSVC.
15928 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15929 (the default implementation of RAND_status).
15931 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15932 to '-clrext' (= clear extensions), as intended and documented.
15933 *Bodo Moeller; inconsistency pointed out by Michael Attili
15934 <attili@amaxo.com>*
15936 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15937 was larger than the MD block size.
15939 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15941 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15942 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15943 using the passed key: if the passed key was a private key the result
15944 of X509_print(), for example, would be to print out all the private key
15949 * des_quad_cksum() byte order bug fix.
15950 *Ulf Möller, using the problem description in krb4-0.9.7, where
15951 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15953 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15956 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15958 * For easily testing in shell scripts whether some command
15959 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15960 returns with exit code 0 iff no command of the given name is available.
15961 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15962 the output goes to stdout and nothing is printed to stderr.
15963 Additional arguments are always ignored.
15965 Since for each cipher there is a command of the same name,
15966 the 'no-cipher' compilation switches can be tested this way.
15968 ('openssl no-XXX' is not able to detect pseudo-commands such
15969 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15973 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15977 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15978 is set; it will be thrown away anyway because each handshake creates
15980 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15981 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15982 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15983 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15987 * New s_client option -ign_eof: EOF at stdin is ignored, and
15988 'Q' and 'R' lose their special meanings (quit/renegotiate).
15989 This is part of what -quiet does; unlike -quiet, -ign_eof
15990 does not suppress any output.
15994 * Add compatibility options to the purpose and trust code. The
15995 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15996 accepts a certificate or CA, this was the previous behaviour,
15997 with all the associated security issues.
15999 X509_TRUST_COMPAT is the old trust behaviour: only and
16000 automatically trust self signed roots in certificate store. A
16001 new trust setting X509_TRUST_DEFAULT is used to specify that
16002 a purpose has no associated trust setting and it should instead
16003 use the value in the default purpose.
16007 * Fix the PKCS#8 DSA private key code so it decodes keys again
16008 and fix a memory leak.
16012 * In util/mkerr.pl (which implements 'make errors'), preserve
16013 reason strings from the previous version of the .c file, as
16014 the default to have only downcase letters (and digits) in
16015 automatically generated reasons codes is not always appropriate.
16019 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16020 using strerror. Previously, ERR_reason_error_string() returned
16021 library names as reason strings for SYSerr; but SYSerr is a special
16022 case where small numbers are errno values, not library numbers.
16026 * Add '-dsaparam' option to 'openssl dhparam' application. This
16027 converts DSA parameters into DH parameters. (When creating parameters,
16028 DSA_generate_parameters is used.)
16032 * Include 'length' (recommended exponent length) in C code generated
16033 by 'openssl dhparam -C'.
16037 * The second argument to set_label in perlasm was already being used
16038 so couldn't be used as a "file scope" flag. Moved to third argument
16043 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16044 instead of RAND_bytes for encryption IVs and salts.
16048 * Include RAND_status() into RAND_METHOD instead of implementing
16049 it only for md_rand.c Otherwise replacing the PRNG by calling
16050 RAND_set_rand_method would be impossible.
16054 * Don't let DSA_generate_key() enter an infinite loop if the random
16055 number generation fails.
16059 * New 'rand' application for creating pseudo-random output.
16063 * Added configuration support for Linux/IA64
16065 *Rolf Haberrecker <rolf@suse.de>*
16067 * Assembler module support for Mingw32.
16071 * Shared library support for HPUX (in shlib/).
16073 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16075 * Shared library support for Solaris gcc.
16077 *Lutz Behnke <behnke@trustcenter.de>*
16079 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16081 * PKCS7_encrypt() was adding text MIME headers twice because they
16082 were added manually and by SMIME_crlf_copy().
16086 * In bntest.c don't call BN_rand with zero bits argument.
16088 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16090 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16091 case was implemented. This caused BN_div_recp() to fail occasionally.
16095 * Add an optional second argument to the set_label() in the perl
16096 assembly language builder. If this argument exists and is set
16097 to 1 it signals that the assembler should use a symbol whose
16098 scope is the entire file, not just the current function. This
16099 is needed with MASM which uses the format label:: for this scope.
16101 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16103 * Change the ASN1 types so they are typedefs by default. Before
16104 almost all types were #define'd to ASN1_STRING which was causing
16105 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16110 * Change names of new functions to the new get1/get0 naming
16111 convention: After 'get1', the caller owns a reference count
16112 and has to call `..._free`; 'get0' returns a pointer to some
16113 data structure without incrementing reference counters.
16114 (Some of the existing 'get' functions increment a reference
16115 counter, some don't.)
16116 Similarly, 'set1' and 'add1' functions increase reference
16117 counters or duplicate objects.
16121 * Allow for the possibility of temp RSA key generation failure:
16122 the code used to assume it always worked and crashed on failure.
16126 * Fix potential buffer overrun problem in BIO_printf().
16127 *Ulf Möller, using public domain code by Patrick Powell; problem
16128 pointed out by David Sacerdote <das33@cornell.edu>*
16130 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16131 RAND_egd() and RAND_status(). In the command line application,
16132 the EGD socket can be specified like a seed file using RANDFILE
16137 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16138 Some CAs (e.g. Verisign) distribute certificates in this form.
16142 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16143 list to exclude them. This means that no special compilation option
16144 is needed to use anonymous DH: it just needs to be included in the
16149 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16150 EVP_MD_type. The old functionality is available in a new macro called
16151 EVP_MD_md(). Change code that uses it and update docs.
16155 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16156 where the `void *` argument is replaced by a function pointer argument.
16157 Previously `void *` was abused to point to functions, which works on
16158 many platforms, but is not correct. As these functions are usually
16159 called by macros defined in OpenSSL header files, most source code
16160 should work without changes.
16164 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16165 sections with information on -D... compiler switches used for
16166 compiling the library so that applications can see them. To enable
16167 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16168 must be defined. E.g.,
16169 #define OPENSSL_ALGORITHM_DEFINES
16170 #include <openssl/opensslconf.h>
16171 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16173 *Richard Levitte, Ulf and Bodo Möller*
16175 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16180 * Change the 'other' type in certificate aux info to a STACK_OF
16181 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16182 the required ASN1 format: arbitrary types determined by an OID.
16186 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16187 argument to 'req'. This is not because the function is newer or
16188 better than others it just uses the work 'NEW' in the certificate
16189 request header lines. Some software needs this.
16193 * Reorganise password command line arguments: now passwords can be
16194 obtained from various sources. Delete the PEM_cb function and make
16195 it the default behaviour: i.e. if the callback is NULL and the
16196 usrdata argument is not NULL interpret it as a null terminated pass
16197 phrase. If usrdata and the callback are NULL then the pass phrase
16198 is prompted for as usual.
16202 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16203 the support is automatically enabled. The resulting binaries will
16204 autodetect the card and use it if present.
16206 *Ben Laurie and Compaq Inc.*
16208 * Work around for Netscape hang bug. This sends certificate request
16209 and server done in one record. Since this is perfectly legal in the
16210 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16211 the bugs/SSLv3 entry for more info.
16215 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16219 * Add -rand argument to smime and pkcs12 applications and read/write
16224 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16228 * Add command line password options to the remaining applications.
16232 * Bug fix for BN_div_recp() for numerators with an even number of
16237 * More tests in bntest.c, and changed test_bn output.
16241 * ./config recognizes MacOS X now.
16245 * Bug fix for BN_div() when the first words of num and divisor are
16246 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16250 * Add support for various broken PKCS#8 formats, and command line
16251 options to produce them.
16255 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16256 get temporary BIGNUMs from a BN_CTX.
16260 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16265 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16266 include a #define from the old name to the new. The original intent
16267 was that statically linked binaries could for example just call
16268 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16269 link with digests. This never worked because SSLeay_add_all_digests()
16270 and SSLeay_add_all_ciphers() were in the same source file so calling
16271 one would link with the other. They are now in separate source files.
16275 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16279 * Use a less unusual form of the Miller-Rabin primality test (it used
16280 a binary algorithm for exponentiation integrated into the Miller-Rabin
16281 loop, our standard modexp algorithms are faster).
16285 * Support for the EBCDIC character set completed.
16287 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16289 * Source code cleanups: use const where appropriate, eliminate casts,
16290 use `void *` instead of `char *` in lhash.
16294 * Bugfix: ssl3_send_server_key_exchange was not restartable
16295 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16296 this the server could overwrite ephemeral keys that the client
16301 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16302 using 50 iterations of the Rabin-Miller test.
16304 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16305 iterations of the Rabin-Miller test as required by the appendix
16306 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16307 As BN_is_prime_fasttest includes trial division, DSA parameter
16308 generation becomes much faster.
16310 This implies a change for the callback functions in DSA_is_prime
16311 and DSA_generate_parameters: The callback function is called once
16312 for each positive witness in the Rabin-Miller test, not just
16313 occasionally in the inner loop; and the parameters to the
16314 callback function now provide an iteration count for the outer
16315 loop rather than for the current invocation of the inner loop.
16316 DSA_generate_parameters additionally can call the callback
16317 function with an 'iteration count' of -1, meaning that a
16318 candidate has passed the trial division test (when q is generated
16319 from an application-provided seed, trial division is skipped).
16323 * New function BN_is_prime_fasttest that optionally does trial
16324 division before starting the Rabin-Miller test and has
16325 an additional BN_CTX * argument (whereas BN_is_prime always
16326 has to allocate at least one BN_CTX).
16327 'callback(1, -1, cb_arg)' is called when a number has passed the
16328 trial division stage.
16332 * Fix for bug in CRL encoding. The validity dates weren't being handled
16337 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16341 * New function BN_pseudo_rand().
16345 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16346 bignum version of BN_from_montgomery() with the working code from
16347 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16352 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16353 made it impossible to use the same SSL_SESSION data structure in
16354 SSL2 clients in multiple threads.
16358 * The return value of RAND_load_file() no longer counts bytes obtained
16359 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16360 to seed the PRNG (previously an explicit byte count was required).
16362 *Ulf Möller, Bodo Möller*
16364 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16365 used `char *` instead of `void *` and had casts all over the place.
16369 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16373 * Retain source code compatibility for BN_prime_checks macro:
16374 BN_is_prime(..., BN_prime_checks, ...) now uses
16375 BN_prime_checks_for_size to determine the appropriate number of
16376 Rabin-Miller iterations.
16380 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16381 DH_CHECK_P_NOT_SAFE_PRIME.
16382 (Check if this is true? OpenPGP calls them "strong".)
16386 * Merge the functionality of "dh" and "gendh" programs into a new program
16387 "dhparam". The old programs are retained for now but will handle DH keys
16388 (instead of parameters) in future.
16392 * Make the ciphers, s_server and s_client programs check the return values
16393 when a new cipher list is set.
16397 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16398 ciphers. Before when the 56bit ciphers were enabled the sorting was
16401 The syntax for the cipher sorting has been extended to support sorting by
16402 cipher-strength (using the strength_bits hard coded in the tables).
16403 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16405 Fix a bug in the cipher-command parser: when supplying a cipher command
16406 string with an "undefined" symbol (neither command nor alphanumeric
16407 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16408 an error is flagged.
16410 Due to the strength-sorting extension, the code of the
16411 ssl_create_cipher_list() function was completely rearranged. I hope that
16412 the readability was also increased :-)
16414 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16416 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16417 for the first serial number and places 2 in the serial number file. This
16418 avoids problems when the root CA is created with serial number zero and
16419 the first user certificate has the same issuer name and serial number
16424 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16425 the new code. Add documentation for this stuff.
16429 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16430 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16431 structures and behave in an analogous way to the X509v3 functions:
16432 they shouldn't be called directly but wrapper functions should be used
16435 So we also now have some wrapper functions that call the X509at functions
16436 when passed certificate requests. (TO DO: similar things can be done with
16437 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16438 things. Some of these need some d2i or i2d and print functionality
16439 because they handle more complex structures.)
16443 * Add missing #ifndefs that caused missing symbols when building libssl
16444 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16445 NO_RSA in `ssl/s2*.c`.
16447 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16449 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16450 has a return value which indicates the quality of the random data
16451 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16452 error queue. New function RAND_pseudo_bytes() generates output that is
16453 guaranteed to be unique but not unpredictable. RAND_add is like
16454 RAND_seed, but takes an extra argument for an entropy estimate
16455 (RAND_seed always assumes full entropy).
16459 * Do more iterations of Rabin-Miller probable prime test (specifically,
16460 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16461 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16462 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16463 false-positive rate of at most 2^-80 for random input.
16467 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16471 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16472 in the 0.9.5 release), this returns the chain
16473 from an X509_CTX structure with a dup of the stack and all
16474 the X509 reference counts upped: so the stack will exist
16475 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16478 Also make SSL_SESSION_print() print out the verify return
16483 * Add manpage for the pkcs12 command. Also change the default
16484 behaviour so MAC iteration counts are used unless the new
16485 -nomaciter option is used. This improves file security and
16486 only older versions of MSIE (4.0 for example) need it.
16490 * Honor the no-xxx Configure options when creating .DEF files.
16494 * Add PKCS#10 attributes to field table: challengePassword,
16495 unstructuredName and unstructuredAddress. These are taken from
16496 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16497 international characters are used.
16499 More changes to X509_ATTRIBUTE code: allow the setting of types
16500 based on strings. Remove the 'loc' parameter when adding
16501 attributes because these will be a SET OF encoding which is sorted
16506 * Initial changes to the 'req' utility to allow request generation
16507 automation. This will allow an application to just generate a template
16508 file containing all the field values and have req construct the
16511 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16512 used all over the place including certificate requests and PKCS#7
16513 structures. They are currently handled manually where necessary with
16514 some primitive wrappers for PKCS#7. The new functions behave in a
16515 manner analogous to the X509 extension functions: they allow
16516 attributes to be looked up by NID and added.
16518 Later something similar to the X509V3 code would be desirable to
16519 automatically handle the encoding, decoding and printing of the
16520 more complex types. The string types like challengePassword can
16521 be handled by the string table functions.
16523 Also modified the multi byte string table handling. Now there is
16524 a 'global mask' which masks out certain types. The table itself
16525 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16526 is useful when for example there is only one permissible type
16527 (as in countryName) and using the mask might result in no valid
16532 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16533 SSL_get_peer_finished to allow applications to obtain the latest
16534 Finished messages sent to the peer or expected from the peer,
16535 respectively. (SSL_get_peer_finished is usually the Finished message
16536 actually received from the peer, otherwise the protocol will be aborted.)
16538 As the Finished message are message digests of the complete handshake
16539 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16540 be used for external authentication procedures when the authentication
16541 provided by SSL/TLS is not desired or is not enough.
16545 * Enhanced support for Alpha Linux is added. Now ./config checks if
16546 the host supports BWX extension and if Compaq C is present on the
16547 $PATH. Just exploiting of the BWX extension results in 20-30%
16548 performance kick for some algorithms, e.g. DES and RC4 to mention
16549 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16554 * Add support for MS "fast SGC". This is arguably a violation of the
16555 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16556 weak crypto and after checking the certificate is SGC a second one
16557 with strong crypto. MS SGC stops the first handshake after receiving
16558 the server certificate message and sends a second client hello. Since
16559 a server will typically do all the time consuming operations before
16560 expecting any further messages from the client (server key exchange
16561 is the most expensive) there is little difference between the two.
16563 To get OpenSSL to support MS SGC we have to permit a second client
16564 hello message after we have sent server done. In addition we have to
16565 reset the MAC if we do get this second client hello.
16569 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16570 if a DER encoded private key is RSA or DSA traditional format. Changed
16571 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16572 format DER encoded private key. Newer code should use PKCS#8 format which
16573 has the key type encoded in the ASN1 structure. Added DER private key
16574 support to pkcs8 application.
16578 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16579 ciphersuites has been selected (as required by the SSL 3/TLS 1
16580 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16581 is set, we interpret this as a request to violate the specification
16582 (the worst that can happen is a handshake failure, and 'correct'
16583 behaviour would result in a handshake failure anyway).
16587 * In SSL_CTX_add_session, take into account that there might be multiple
16588 SSL_SESSION structures with the same session ID (e.g. when two threads
16589 concurrently obtain them from an external cache).
16590 The internal cache can handle only one SSL_SESSION with a given ID,
16591 so if there's a conflict, we now throw out the old one to achieve
16596 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16597 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16598 some routines that use cipher OIDs: some ciphers do not have OIDs
16599 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16604 * Simplify the trust setting structure and code. Now we just have
16605 two sequences of OIDs for trusted and rejected settings. These will
16606 typically have values the same as the extended key usage extension
16607 and any application specific purposes.
16609 The trust checking code now has a default behaviour: it will just
16610 check for an object with the same NID as the passed id. Functions can
16611 be provided to override either the default behaviour or the behaviour
16612 for a given id. SSL client, server and email already have functions
16613 in place for compatibility: they check the NID and also return "trusted"
16614 if the certificate is self signed.
16618 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16619 traditional format into an EVP_PKEY structure.
16623 * Add a password callback function PEM_cb() which either prompts for
16624 a password if usr_data is NULL or otherwise assumes it is a null
16625 terminated password. Allow passwords to be passed on command line
16626 environment or config files in a few more utilities.
16630 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16631 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16632 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16633 Update documentation.
16637 * Support for ASN1 "NULL" type. This could be handled before by using
16638 ASN1_TYPE but there wasn't any function that would try to read a NULL
16639 and produce an error if it couldn't. For compatibility we also have
16640 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16641 don't allocate anything because they don't need to.
16645 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16648 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16650 * Rebuild of the memory allocation routines used by OpenSSL code and
16651 possibly others as well. The purpose is to make an interface that
16652 provide hooks so anyone can build a separate set of allocation and
16653 deallocation routines to be used by OpenSSL, for example memory
16654 pool implementations, or something else, which was previously hard
16655 since Malloc(), Realloc() and Free() were defined as macros having
16656 the values malloc, realloc and free, respectively (except for Win32
16657 compilations). The same is provided for memory debugging code.
16658 OpenSSL already comes with functionality to find memory leaks, but
16659 this gives people a chance to debug other memory problems.
16661 With these changes, a new set of functions and macros have appeared:
16663 CRYPTO_set_mem_debug_functions() [F]
16664 CRYPTO_get_mem_debug_functions() [F]
16665 CRYPTO_dbg_set_options() [F]
16666 CRYPTO_dbg_get_options() [F]
16667 CRYPTO_malloc_debug_init() [M]
16669 The memory debug functions are NULL by default, unless the library
16670 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16671 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16672 gives the standard debugging functions that come with OpenSSL) or
16673 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16674 provided by the library user) must be used. When the standard
16675 debugging functions are used, CRYPTO_dbg_set_options can be used to
16676 request additional information:
16677 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16678 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16680 Also, things like CRYPTO_set_mem_functions will always give the
16681 expected result (the new set of functions is used for allocation
16682 and deallocation) at all times, regardless of platform and compiler
16685 To finish it up, some functions that were never use in any other
16686 way than through macros have a new API and new semantic:
16688 CRYPTO_dbg_malloc()
16689 CRYPTO_dbg_realloc()
16692 All macros of value have retained their old syntax.
16694 *Richard Levitte and Bodo Moeller*
16696 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16697 ordering of SMIMECapabilities wasn't in "strength order" and there
16698 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16703 * Some ASN1 types with illegal zero length encoding (INTEGER,
16704 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16706 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16708 * Merge in my S/MIME library for OpenSSL. This provides a simple
16709 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16710 functionality to handle multipart/signed properly) and a utility
16711 called 'smime' to call all this stuff. This is based on code I
16712 originally wrote for Celo who have kindly allowed it to be
16713 included in OpenSSL.
16717 * Add variants des_set_key_checked and des_set_key_unchecked of
16718 des_set_key (aka des_key_sched). Global variable des_check_key
16719 decides which of these is called by des_set_key; this way
16720 des_check_key behaves as it always did, but applications and
16721 the library itself, which was buggy for des_check_key == 1,
16722 have a cleaner way to pick the version they need.
16726 * New function PKCS12_newpass() which changes the password of a
16731 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16732 dynamic mix. In both cases the ids can be used as an index into the
16733 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16734 functions so they accept a list of the field values and the
16735 application doesn't need to directly manipulate the X509_TRUST
16740 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16745 * Modify the way the V3 extension code looks up extensions. This now
16746 works in a similar way to the object code: we have some "standard"
16747 extensions in a static table which is searched with OBJ_bsearch()
16748 and the application can add dynamic ones if needed. The file
16749 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16750 updated whenever a new extension is added to the core code and kept
16751 in ext_nid order. There is a simple program 'tabtest.c' which checks
16752 this. New extensions are not added too often so this file can readily
16753 be maintained manually.
16755 There are two big advantages in doing things this way. The extensions
16756 can be looked up immediately and no longer need to be "added" using
16757 X509V3_add_standard_extensions(): this function now does nothing.
16758 Side note: I get *lots* of email saying the extension code doesn't
16759 work because people forget to call this function.
16760 Also no dynamic allocation is done unless new extensions are added:
16761 so if we don't add custom extensions there is no need to call
16762 X509V3_EXT_cleanup().
16766 * Modify enc utility's salting as follows: make salting the default. Add a
16767 magic header, so unsalted files fail gracefully instead of just decrypting
16768 to garbage. This is because not salting is a big security hole, so people
16769 should be discouraged from doing it.
16773 * Fixes and enhancements to the 'x509' utility. It allowed a message
16774 digest to be passed on the command line but it only used this
16775 parameter when signing a certificate. Modified so all relevant
16776 operations are affected by the digest parameter including the
16777 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16778 DSA key was used because it didn't fix the digest.
16782 * Initial certificate chain verify code. Currently tests the untrusted
16783 certificates for consistency with the verify purpose (which is set
16784 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16786 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16787 this is because it will reject chains with invalid extensions whereas
16788 every previous version of OpenSSL and SSLeay made no checks at all.
16790 Trust code: checks the root CA for the relevant trust settings. Trust
16791 settings have an initial value consistent with the verify purpose: e.g.
16792 if the verify purpose is for SSL client use it expects the CA to be
16793 trusted for SSL client use. However the default value can be changed to
16794 permit custom trust settings: one example of this would be to only trust
16795 certificates from a specific "secure" set of CAs.
16797 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16798 which should be used for version portability: especially since the
16799 verify structure is likely to change more often now.
16801 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16802 to set them. If not set then assume SSL clients will verify SSL servers
16805 Two new options to the verify program: -untrusted allows a set of
16806 untrusted certificates to be passed in and -purpose which sets the
16807 intended purpose of the certificate. If a purpose is set then the
16808 new chain verify code is used to check extension consistency.
16812 * Support for the authority information access extension.
16816 * Modify RSA and DSA PEM read routines to transparently handle
16817 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16818 public keys in a format compatible with certificate
16819 SubjectPublicKeyInfo structures. Unfortunately there were already
16820 functions called *_PublicKey_* which used various odd formats so
16821 these are retained for compatibility: however the DSA variants were
16822 never in a public release so they have been deleted. Changed dsa/rsa
16823 utilities to handle the new format: note no releases ever handled public
16824 keys so we should be OK.
16826 The primary motivation for this change is to avoid the same fiasco
16827 that dogs private keys: there are several incompatible private key
16828 formats some of which are standard and some OpenSSL specific and
16829 require various evil hacks to allow partial transparent handling and
16830 even then it doesn't work with DER formats. Given the option anything
16831 other than PKCS#8 should be dumped: but the other formats have to
16832 stay in the name of compatibility.
16834 With public keys and the benefit of hindsight one standard format
16835 is used which works with EVP_PKEY, RSA or DSA structures: though
16836 it clearly returns an error if you try to read the wrong kind of key.
16838 Added a -pubkey option to the 'x509' utility to output the public key.
16839 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16840 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16841 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16842 that do the same as the `EVP_PKEY_assign_*()` except they up the
16843 reference count of the added key (they don't "swallow" the
16848 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16849 CRLs would fail if the file contained no certificates or no CRLs:
16850 added a new function to read in both types and return the number
16851 read: this means that if none are read it will be an error. The
16852 DER versions of the certificate and CRL reader would always fail
16853 because it isn't possible to mix certificates and CRLs in DER format
16854 without choking one or the other routine. Changed this to just read
16855 a certificate: this is the best we can do. Also modified the code
16856 in `apps/verify.c` to take notice of return codes: it was previously
16857 attempting to read in certificates from NULL pointers and ignoring
16858 any errors: this is one reason why the cert and CRL reader seemed
16859 to work. It doesn't check return codes from the default certificate
16860 routines: these may well fail if the certificates aren't installed.
16864 * Code to support otherName option in GeneralName.
16868 * First update to verify code. Change the verify utility
16869 so it warns if it is passed a self signed certificate:
16870 for consistency with the normal behaviour. X509_verify
16871 has been modified to it will now verify a self signed
16872 certificate if *exactly* the same certificate appears
16873 in the store: it was previously impossible to trust a
16874 single self signed certificate. This means that:
16875 openssl verify ss.pem
16876 now gives a warning about a self signed certificate but
16877 openssl verify -CAfile ss.pem ss.pem
16882 * For servers, store verify_result in SSL_SESSION data structure
16883 (and add it to external session representation).
16884 This is needed when client certificate verifications fails,
16885 but an application-provided verification callback (set by
16886 SSL_CTX_set_cert_verify_callback) allows accepting the session
16887 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16888 but returns 1): When the session is reused, we have to set
16889 ssl->verify_result to the appropriate error code to avoid
16892 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16894 * Fix a bug in the new PKCS#7 code: it didn't consider the
16895 case in PKCS7_dataInit() where the signed PKCS7 structure
16896 didn't contain any existing data because it was being created.
16898 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16900 * Add a salt to the key derivation routines in enc.c. This
16901 forms the first 8 bytes of the encrypted file. Also add a
16902 -S option to allow a salt to be input on the command line.
16906 * New function X509_cmp(). Oddly enough there wasn't a function
16907 to compare two certificates. We do this by working out the SHA1
16908 hash and comparing that. X509_cmp() will be needed by the trust
16913 * SSL_get1_session() is like SSL_get_session(), but increments
16914 the reference count in the SSL_SESSION returned.
16916 *Geoff Thorpe <geoff@eu.c2.net>*
16918 * Fix for 'req': it was adding a null to request attributes.
16919 Also change the X509_LOOKUP and X509_INFO code to handle
16920 certificate auxiliary information.
16924 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16929 * Add the possibility to add extra information to the memory leak
16930 detecting output, to form tracebacks, showing from where each
16931 allocation was originated: CRYPTO_push_info("constant string") adds
16932 the string plus current file name and line number to a per-thread
16933 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16934 is like calling CYRPTO_pop_info() until the stack is empty.
16935 Also updated memory leak detection code to be multi-thread-safe.
16939 * Add options -text and -noout to pkcs7 utility and delete the
16940 encryption options which never did anything. Update docs.
16944 * Add options to some of the utilities to allow the pass phrase
16945 to be included on either the command line (not recommended on
16946 OSes like Unix) or read from the environment. Update the
16947 manpages and fix a few bugs.
16951 * Add a few manpages for some of the openssl commands.
16955 * Fix the -revoke option in ca. It was freeing up memory twice,
16956 leaking and not finding already revoked certificates.
16960 * Extensive changes to support certificate auxiliary information.
16961 This involves the use of X509_CERT_AUX structure and X509_AUX
16962 functions. An X509_AUX function such as PEM_read_X509_AUX()
16963 can still read in a certificate file in the usual way but it
16964 will also read in any additional "auxiliary information". By
16965 doing things this way a fair degree of compatibility can be
16966 retained: existing certificates can have this information added
16967 using the new 'x509' options.
16969 Current auxiliary information includes an "alias" and some trust
16970 settings. The trust settings will ultimately be used in enhanced
16971 certificate chain verification routines: currently a certificate
16972 can only be trusted if it is self signed and then it is trusted
16977 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16978 The problem was that one of the replacement routines had not been working
16979 since SSLeay releases. For now the offending routine has been replaced
16980 with non-optimised assembler. Even so, this now gives around 95%
16981 performance improvement for 1024 bit RSA signs.
16985 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16986 handling. Most clients have the effective key size in bits equal to
16987 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16988 A few however don't do this and instead use the size of the decrypted key
16989 to determine the RC2 key length and the AlgorithmIdentifier to determine
16990 the effective key length. In this case the effective key length can still
16991 be 40 bits but the key length can be 168 bits for example. This is fixed
16992 by manually forcing an RC2 key into the EVP_PKEY structure because the
16993 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16994 the key length and effective key length are equal.
16998 * Add a bunch of functions that should simplify the creation of
16999 X509_NAME structures. Now you should be able to do:
17000 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17001 and have it automatically work out the correct field type and fill in
17002 the structures. The more adventurous can try:
17003 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17004 and it will (hopefully) work out the correct multibyte encoding.
17008 * Change the 'req' utility to use the new field handling and multibyte
17009 copy routines. Before the DN field creation was handled in an ad hoc
17010 way in req, ca, and x509 which was rather broken and didn't support
17011 BMPStrings or UTF8Strings. Since some software doesn't implement
17012 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17013 using the dirstring_type option. See the new comment in the default
17014 openssl.cnf for more info.
17018 * Make crypto/rand/md_rand.c more robust:
17019 - Assure unique random numbers after fork().
17020 - Make sure that concurrent threads access the global counter and
17021 md serializably so that we never lose entropy in them
17022 or use exactly the same state in multiple threads.
17023 Access to the large state is not always serializable because
17024 the additional locking could be a performance killer, and
17025 md should be large enough anyway.
17029 * New file `apps/app_rand.c` with commonly needed functionality
17030 for handling the random seed file.
17032 Use the random seed file in some applications that previously did not:
17034 dsaparam -genkey (which also ignored its '-rand' option),
17037 x509 (when signing).
17038 Except on systems with /dev/urandom, it is crucial to have a random
17039 seed file at least for key creation, DSA signing, and for DH exchanges;
17040 for RSA signatures we could do without one.
17042 gendh and gendsa (unlike genrsa) used to read only the first byte
17043 of each file listed in the '-rand' option. The function as previously
17044 found in genrsa is now in app_rand.c and is used by all programs
17045 that support '-rand'.
17049 * In RAND_write_file, use mode 0600 for creating files;
17050 don't just chmod when it may be too late.
17054 * Report an error from X509_STORE_load_locations
17055 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17059 * New function ASN1_mbstring_copy() this copies a string in either
17060 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17061 into an ASN1_STRING type. A mask of permissible types is passed
17062 and it chooses the "minimal" type to use or an error if not type
17067 * Add function equivalents to the various macros in asn1.h. The old
17068 macros are retained with an `M_` prefix. Code inside the library can
17069 use the `M_` macros. External code (including the openssl utility)
17070 should *NOT* in order to be "shared library friendly".
17074 * Add various functions that can check a certificate's extensions
17075 to see if it usable for various purposes such as SSL client,
17076 server or S/MIME and CAs of these types. This is currently
17077 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17078 verification. Also added a -purpose flag to x509 utility to
17079 print out all the purposes.
17083 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17088 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17089 for, obtain and decode and extension and obtain its critical flag.
17090 This allows all the necessary extension code to be handled in a
17091 single function call.
17095 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17096 platforms. See crypto/rc4/rc4_enc.c for further details.
17100 * New -noout option to asn1parse. This causes no output to be produced
17101 its main use is when combined with -strparse and -out to extract data
17102 from a file (which may not be in ASN.1 format).
17106 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17107 when producing the local key id.
17109 *Richard Levitte <levitte@stacken.kth.se>*
17111 * New option -dhparam in s_server. This allows a DH parameter file to be
17112 stated explicitly. If it is not stated then it tries the first server
17113 certificate file. The previous behaviour hard coded the filename
17118 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17119 a public key to be input or output. For example:
17120 openssl rsa -in key.pem -pubout -out pubkey.pem
17121 Also added necessary DSA public key functions to handle this.
17125 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17126 in the message. This was handled by allowing
17127 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17129 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17131 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17132 to the end of the strings whereas this didn't. This would cause problems
17133 if strings read with d2i_ASN1_bytes() were later modified.
17135 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17137 * Fix for base64 decode bug. When a base64 bio reads only one line of
17138 data and it contains EOF it will end up returning an error. This is
17139 caused by input 46 bytes long. The cause is due to the way base64
17140 BIOs find the start of base64 encoded data. They do this by trying a
17141 trial decode on each line until they find one that works. When they
17142 do a flag is set and it starts again knowing it can pass all the
17143 data directly through the decoder. Unfortunately it doesn't reset
17144 the context it uses. This means that if EOF is reached an attempt
17145 is made to pass two EOFs through the context and this causes the
17146 resulting error. This can also cause other problems as well. As is
17147 usual with these problems it takes *ages* to find and the fix is
17148 trivial: move one line.
17150 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17152 * Ugly workaround to get s_client and s_server working under Windows. The
17153 old code wouldn't work because it needed to select() on sockets and the
17154 tty (for keypresses and to see if data could be written). Win32 only
17155 supports select() on sockets so we select() with a 1s timeout on the
17156 sockets and then see if any characters are waiting to be read, if none
17157 are present then we retry, we also assume we can always write data to
17158 the tty. This isn't nice because the code then blocks until we've
17159 received a complete line of data and it is effectively polling the
17160 keyboard at 1s intervals: however it's quite a bit better than not
17161 working at all :-) A dedicated Windows application might handle this
17162 with an event loop for example.
17166 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17167 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17168 will be called when RSA_sign() and RSA_verify() are used. This is useful
17169 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17170 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17171 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17172 This necessitated the support of an extra signature type NID_md5_sha1
17173 for SSL signatures and modifications to the SSL library to use it instead
17174 of calling RSA_public_decrypt() and RSA_private_encrypt().
17178 * Add new -verify -CAfile and -CApath options to the crl program, these
17179 will lookup a CRL issuers certificate and verify the signature in a
17180 similar way to the verify program. Tidy up the crl program so it
17181 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17182 less strict. It will now permit CRL extensions even if it is not
17183 a V2 CRL: this will allow it to tolerate some broken CRLs.
17187 * Initialize all non-automatic variables each time one of the openssl
17188 sub-programs is started (this is necessary as they may be started
17189 multiple times from the "OpenSSL>" prompt).
17191 *Lennart Bang, Bodo Moeller*
17193 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17194 removing all other RSA functionality (this is what NO_RSA does). This
17195 is so (for example) those in the US can disable those operations covered
17196 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17201 * Non-copying interface to BIO pairs.
17202 (still largely untested)
17206 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17207 ASCII string. This was handled independently in various places before.
17211 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17212 UTF8 strings a character at a time.
17216 * Use client_version from client hello to select the protocol
17217 (s23_srvr.c) and for RSA client key exchange verification
17218 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17222 * Add various utility functions to handle SPKACs, these were previously
17223 handled by poking round in the structure internals. Added new function
17224 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17225 print, verify and generate SPKACs. Based on an original idea from
17226 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17230 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17234 * Allow the config file extension section to be overwritten on the
17235 command line. Based on an original idea from Massimiliano Pala
17236 <madwolf@comune.modena.it>. The new option is called -extensions
17237 and can be applied to ca, req and x509. Also -reqexts to override
17238 the request extensions in req and -crlexts to override the crl extensions
17243 * Add new feature to the SPKAC handling in ca. Now you can include
17244 the same field multiple times by preceding it by "XXXX." for example:
17247 this is the same syntax as used in the req config file.
17251 * Allow certificate extensions to be added to certificate requests. These
17252 are specified in a 'req_extensions' option of the req section of the
17253 config file. They can be printed out with the -text option to req but
17254 are otherwise ignored at present.
17258 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17259 data read consists of only the final block it would not decrypted because
17260 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17261 A misplaced 'break' also meant the decrypted final block might not be
17262 copied until the next read.
17266 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17267 a few extra parameters to the DH structure: these will be useful if
17268 for example we want the value of 'q' or implement X9.42 DH.
17272 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17273 provides hooks that allow the default DSA functions or functions on a
17274 "per key" basis to be replaced. This allows hardware acceleration and
17275 hardware key storage to be handled without major modification to the
17276 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17277 associated functions.
17281 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17282 as "read only": it can't be written to and the buffer it points to will
17283 not be freed. Reading from a read only BIO is much more efficient than
17284 a normal memory BIO. This was added because there are several times when
17285 an area of memory needs to be read from a BIO. The previous method was
17286 to create a memory BIO and write the data to it, this results in two
17287 copies of the data and an O(n^2) reading algorithm. There is a new
17288 function BIO_new_mem_buf() which creates a read only memory BIO from
17289 an area of memory. Also modified the PKCS#7 routines to use read only
17294 * Bugfix: ssl23_get_client_hello did not work properly when called in
17295 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17296 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17297 but a retry condition occurred while trying to read the rest.
17301 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17302 NID_pkcs7_encrypted by default: this was wrong since this should almost
17303 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17304 the encrypted data type: this is a more sensible place to put it and it
17305 allows the PKCS#12 code to be tidied up that duplicated this
17310 * Changed obj_dat.pl script so it takes its input and output files on
17311 the command line. This should avoid shell escape redirection problems
17316 * Initial support for certificate extension requests, these are included
17317 in things like Xenroll certificate requests. Included functions to allow
17318 extensions to be obtained and added.
17322 * -crlf option to s_client and s_server for sending newlines as
17323 CRLF (as required by many protocols).
17327 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17329 * Install libRSAglue.a when OpenSSL is built with RSAref.
17331 *Ralf S. Engelschall*
17333 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17335 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17337 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17342 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17343 DH parameters/keys (q is lost during that conversion, but the resulting
17344 DH parameters contain its length).
17346 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17347 much faster than DH_generate_parameters (which creates parameters
17348 where `p = 2*q + 1`), and also the smaller q makes DH computations
17349 much more efficient (160-bit exponentiation instead of 1024-bit
17350 exponentiation); so this provides a convenient way to support DHE
17351 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17352 utter importance to use
17353 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17355 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17356 when such DH parameters are used, because otherwise small subgroup
17357 attacks may become possible!
17361 * Avoid memory leak in i2d_DHparams.
17365 * Allow the -k option to be used more than once in the enc program:
17366 this allows the same encrypted message to be read by multiple recipients.
17370 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17371 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17372 it will always use the numerical form of the OID, even if it has a short
17377 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17378 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17379 otherwise bn_mod_exp was called. In the case of hardware keys for example
17380 no private key components need be present and it might store extra data
17381 in the RSA structure, which cannot be accessed from bn_mod_exp.
17382 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17383 private key operations.
17387 * Added support for SPARC Linux.
17391 * pem_password_cb function type incompatibly changed from
17392 typedef int pem_password_cb(char *buf, int size, int rwflag);
17394 ....(char *buf, int size, int rwflag, void *userdata);
17395 so that applications can pass data to their callbacks:
17396 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17397 additional void * argument, which is just handed through whenever
17398 the password callback is called.
17400 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17402 New function SSL_CTX_set_default_passwd_cb_userdata.
17404 Compatibility note: As many C implementations push function arguments
17405 onto the stack in reverse order, the new library version is likely to
17406 interoperate with programs that have been compiled with the old
17407 pem_password_cb definition (PEM_whatever takes some data that
17408 happens to be on the stack as its last argument, and the callback
17409 just ignores this garbage); but there is no guarantee whatsoever that
17412 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17413 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17414 problems not only on Windows, but also on some Unix platforms.
17415 To avoid problematic command lines, these definitions are now in an
17416 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17417 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17421 * MIPS III/IV assembler module is reimplemented.
17425 * More DES library cleanups: remove references to srand/rand and
17426 delete an unused file.
17430 * Add support for the free Netwide assembler (NASM) under Win32,
17431 since not many people have MASM (ml) and it can be hard to obtain.
17432 This is currently experimental but it seems to work OK and pass all
17433 the tests. Check out INSTALL.W32 for info.
17437 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17438 without temporary keys kept an extra copy of the server key,
17439 and connections with temporary keys did not free everything in case
17444 * New function RSA_check_key and new openssl rsa option -check
17445 for verifying the consistency of RSA keys.
17447 *Ulf Moeller, Bodo Moeller*
17449 * Various changes to make Win32 compile work:
17450 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17451 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17452 comparison" warnings.
17453 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17457 * Add a debugging option to PKCS#5 v2 key generation function: when
17458 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17459 derived keys are printed to stderr.
17463 * Copy the flags in ASN1_STRING_dup().
17465 *Roman E. Pavlov <pre@mo.msk.ru>*
17467 * The x509 application mishandled signing requests containing DSA
17468 keys when the signing key was also DSA and the parameters didn't match.
17470 It was supposed to omit the parameters when they matched the signing key:
17471 the verifying software was then supposed to automatically use the CA's
17472 parameters if they were absent from the end user certificate.
17474 Omitting parameters is no longer recommended. The test was also
17475 the wrong way round! This was probably due to unusual behaviour in
17476 EVP_cmp_parameters() which returns 1 if the parameters match.
17477 This meant that parameters were omitted when they *didn't* match and
17478 the certificate was useless. Certificates signed with 'ca' didn't have
17481 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17483 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17484 The interface is as follows:
17485 Applications can use
17486 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17487 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17488 "off" is now the default.
17489 The library internally uses
17490 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17491 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17492 to disable memory-checking temporarily.
17494 Some inconsistent states that previously were possible (and were
17495 even the default) are now avoided.
17497 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17498 with each memory chunk allocated; this is occasionally more helpful
17499 than just having a counter.
17501 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17503 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17508 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17509 which largely parallels "options", but is for changing API behaviour,
17510 whereas "options" are about protocol behaviour.
17511 Initial "mode" flags are:
17513 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17514 a single record has been written.
17515 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17516 retries use the same buffer location.
17517 (But all of the contents must be
17522 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17525 * Fix problems with no-hmac etc.
17527 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17529 * New functions RSA_get_default_method(), RSA_set_method() and
17530 RSA_get_method(). These allows replacement of RSA_METHODs without having
17531 to mess around with the internals of an RSA structure.
17535 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17536 Also really enable memory leak checks in openssl.c and in some
17539 *Chad C. Mulligan, Bodo Moeller*
17541 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17542 up the length of negative integers. This has now been simplified to just
17543 store the length when it is first determined and use it later, rather
17544 than trying to keep track of where data is copied and updating it to
17546 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17548 * Add a new function PKCS7_signatureVerify. This allows the verification
17549 of a PKCS#7 signature but with the signing certificate passed to the
17550 function itself. This contrasts with PKCS7_dataVerify which assumes the
17551 certificate is present in the PKCS#7 structure. This isn't always the
17552 case: certificates can be omitted from a PKCS#7 structure and be
17553 distributed by "out of band" means (such as a certificate database).
17557 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17558 function prototypes in pem.h, also change util/mkdef.pl to add the
17559 necessary function names.
17563 * mk1mf.pl (used by Windows builds) did not properly read the
17564 options set by Configure in the top level Makefile, and Configure
17565 was not even able to write more than one option correctly.
17566 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17570 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17571 file to be loaded from a BIO or FILE pointer. The BIO version will
17572 for example allow memory BIOs to contain config info.
17576 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17577 Whoever hopes to achieve shared-library compatibility across versions
17578 must use this, not the compile-time macro.
17579 (Exercise 0.9.4: Which is the minimum library version required by
17581 Note: All this applies only to multi-threaded programs, others don't
17586 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17587 through a BIO pair triggered the default case, i.e.
17588 SSLerr(...,SSL_R_UNKNOWN_STATE).
17592 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17593 can use the SSL library even if none of the specific BIOs is
17598 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17599 for the encoded length.
17601 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17603 * Add initial documentation of the X509V3 functions.
17607 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17608 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17609 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17610 secure PKCS#8 private key format with a high iteration count.
17614 * Fix determination of Perl interpreter: A perl or perl5
17615 *directory* in $PATH was also accepted as the interpreter.
17617 *Ralf S. Engelschall*
17619 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17620 wrong with it but it was very old and did things like calling
17621 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17622 unusual formatting.
17626 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17627 to use the new extension code.
17631 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17632 with macros. This should make it easier to change their form, add extra
17633 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17638 * Add to configuration table a new entry that can specify an alternative
17639 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17640 according to Mark Crispin <MRC@Panda.COM>.
17644 * DES CBC did not update the IV. Weird.
17648 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17649 Changing the behaviour of the former might break existing programs --
17650 where IV updating is needed, des_ncbc_encrypt can be used.
17653 * When bntest is run from "make test" it drives bc to check its
17654 calculations, as well as internally checking them. If an internal check
17655 fails, it needs to cause bc to give a non-zero result or make test carries
17656 on without noticing the failure. Fixed.
17660 * DES library cleanups.
17664 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17665 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17666 ciphers. NOTE: although the key derivation function has been verified
17667 against some published test vectors it has not been extensively tested
17668 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17673 * Instead of "mkdir -p", which is not fully portable, use new
17674 Perl script "util/mkdir-p.pl".
17678 * Rewrite the way password based encryption (PBE) is handled. It used to
17679 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17680 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17681 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17682 the 'parameter' field of the AlgorithmIdentifier is passed to the
17683 underlying key generation function so it must do its own ASN1 parsing.
17684 This has also changed the EVP_PBE_CipherInit() function which now has a
17685 'parameter' argument instead of literal salt and iteration count values
17686 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17690 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17691 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17692 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17693 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17694 value was just used as a "magic string" and not used directly its
17695 value doesn't matter.
17699 * Introduce some semblance of const correctness to BN. Shame C doesn't
17704 * "linux-sparc64" configuration (ultrapenguin).
17706 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17707 "linux-sparc" configuration.
17709 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17711 * config now generates no-xxx options for missing ciphers.
17715 * Support the EBCDIC character set (work in progress).
17716 File ebcdic.c not yet included because it has a different license.
17718 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17720 * Support BS2000/OSD-POSIX.
17722 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17724 * Make callbacks for key generation use `void *` instead of `char *`.
17728 * Make S/MIME samples compile (not yet tested).
17732 * Additional typesafe stacks.
17736 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17740 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17742 * New configuration variant "sco5-gcc".
17744 * Updated some demos.
17746 *Sean O Riordain, Wade Scholine*
17748 * Add missing BIO_free at exit of pkcs12 application.
17752 * Fix memory leak in conf.c.
17756 * Updates for Win32 to assembler version of MD5.
17760 * Set #! path to perl in `apps/der_chop` to where we found it
17761 instead of using a fixed path.
17765 * SHA library changes for irix64-mips4-cc.
17769 * Improvements for VMS support.
17773 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17775 * Bignum library bug fix. IRIX 6 passes "make test" now!
17776 This also avoids the problems with SC4.2 and unpatched SC5.
17778 *Andy Polyakov <appro@fy.chalmers.se>*
17780 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17781 These are required because of the typesafe stack would otherwise break
17782 existing code. If old code used a structure member which used to be STACK
17783 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17784 sk_num or sk_value it would produce an error because the num, data members
17785 are not present in STACK_OF. Now it just produces a warning. sk_set
17786 replaces the old method of assigning a value to sk_value
17787 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17788 that does this will no longer work (and should use sk_set instead) but
17789 this could be regarded as a "questionable" behaviour anyway.
17793 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17794 correctly handle encrypted S/MIME data.
17798 * Change type of various DES function arguments from des_cblock
17799 (which means, in function argument declarations, pointer to char)
17800 to des_cblock * (meaning pointer to array with 8 char elements),
17801 which allows the compiler to do more typechecking; it was like
17802 that back in SSLeay, but with lots of ugly casts.
17804 Introduce new type const_des_cblock.
17808 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17809 problems: find RecipientInfo structure that matches recipient certificate
17810 and initialise the ASN1 structures properly based on passed cipher.
17814 * Belatedly make the BN tests actually check the results.
17818 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17819 to and from BNs: it was completely broken. New compilation option
17820 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17821 key elements as negative integers.
17825 * Reorganize and speed up MD5.
17827 *Andy Polyakov <appro@fy.chalmers.se>*
17831 *Richard Levitte <richard@levitte.org>*
17833 * New option -out to asn1parse to allow the parsed structure to be
17834 output to a file. This is most useful when combined with the -strparse
17835 option to examine the output of things like OCTET STRINGS.
17839 * Make SSL library a little more fool-proof by not requiring any longer
17840 that `SSL_set_{accept,connect}_state` be called before
17841 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17842 in many applications because usually everything *appeared* to work as
17843 intended anyway -- now it really works as intended).
17847 * Move openssl.cnf out of lib/.
17851 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17852 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17853 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17855 *Ralf S. Engelschall*
17857 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17858 handle PKCS#7 enveloped data properly.
17860 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17862 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17863 copying pointers. The cert_st handling is changed by this in
17864 various ways (and thus what used to be known as ctx->default_cert
17865 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17866 any longer when s->cert does not give us what we need).
17867 ssl_cert_instantiate becomes obsolete by this change.
17868 As soon as we've got the new code right (possibly it already is?),
17869 we have solved a couple of bugs of the earlier code where s->cert
17870 was used as if it could not have been shared with other SSL structures.
17872 Note that using the SSL API in certain dirty ways now will result
17873 in different behaviour than observed with earlier library versions:
17874 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17875 does not influence s as it used to.
17877 In order to clean up things more thoroughly, inside SSL_SESSION
17878 we don't use CERT any longer, but a new structure SESS_CERT
17879 that holds per-session data (if available); currently, this is
17880 the peer's certificate chain and, for clients, the server's certificate
17881 and temporary key. CERT holds only those values that can have
17882 meaningful defaults in an SSL_CTX.
17886 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17887 from the internal representation. Various PKCS#7 fixes: remove some
17888 evil casts and set the enc_dig_alg field properly based on the signing
17893 * Allow PKCS#12 password to be set from the command line or the
17894 environment. Let 'ca' get its config file name from the environment
17895 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17900 * Allow certificate policies extension to use an IA5STRING for the
17901 organization field. This is contrary to the PKIX definition but
17902 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17907 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17908 without disallowing inline assembler and the like for non-pedantic builds.
17912 * Support Borland C++ builder.
17914 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17920 * SHA-1 cleanups and performance enhancements.
17922 *Andy Polyakov <appro@fy.chalmers.se>*
17924 * Sparc v8plus assembler for the bignum library.
17926 *Andy Polyakov <appro@fy.chalmers.se>*
17928 * Accept any -xxx and +xxx compiler options in Configure.
17932 * Update HPUX configuration.
17936 * Add missing `sk_<type>_unshift()` function to safestack.h
17938 *Ralf S. Engelschall*
17940 * New function SSL_CTX_use_certificate_chain_file that sets the
17941 "extra_cert"s in addition to the certificate. (This makes sense
17942 only for "PEM" format files, as chains as a whole are not
17947 * Support verify_depth from the SSL API.
17948 x509_vfy.c had what can be considered an off-by-one-error:
17949 Its depth (which was not part of the external interface)
17950 was actually counting the number of certificates in a chain;
17951 now it really counts the depth.
17955 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17956 instead of X509err, which often resulted in confusing error
17957 messages since the error codes are not globally unique
17958 (e.g. an alleged error in ssl3_accept when a certificate
17959 didn't match the private key).
17961 * New function SSL_CTX_set_session_id_context that allows to set a default
17962 value (so that you don't need SSL_set_session_id_context for each
17963 connection using the SSL_CTX).
17967 * OAEP decoding bug fix.
17971 * Support INSTALL_PREFIX for package builders, as proposed by
17976 * New Configure options "threads" and "no-threads". For systems
17977 where the proper compiler options are known (currently Solaris
17978 and Linux), "threads" is the default.
17982 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17986 * Install various scripts to $(OPENSSLDIR)/misc, not to
17987 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17988 such as /usr/local/bin.
17992 * "make linux-shared" to build shared libraries.
17994 *Niels Poppe <niels@netbox.org>*
17996 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18000 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18001 extension adding in x509 utility.
18005 * Remove NOPROTO sections and error code comments.
18009 * Partial rewrite of the DEF file generator to now parse the ANSI
18014 * New Configure options --prefix=DIR and --openssldir=DIR.
18018 * Complete rewrite of the error code script(s). It is all now handled
18019 by one script at the top level which handles error code gathering,
18020 header rewriting and C source file generation. It should be much better
18021 than the old method: it now uses a modified version of Ulf's parser to
18022 read the ANSI prototypes in all header files (thus the old K&R definitions
18023 aren't needed for error creation any more) and do a better job of
18024 translating function codes into names. The old 'ASN1 error code embedded
18025 in a comment' is no longer necessary and it doesn't use .err files which
18026 have now been deleted. Also the error code call doesn't have to appear all
18027 on one line (which resulted in some large lines...).
18031 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18035 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18036 0 (which usually indicates a closed connection), but continue reading.
18040 * Fix some race conditions.
18044 * Add support for CRL distribution points extension. Add Certificate
18045 Policies and CRL distribution points documentation.
18049 * Move the autogenerated header file parts to crypto/opensslconf.h.
18053 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18054 8 of keying material. Merlin has also confirmed interop with this fix
18055 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18057 *Merlin Hughes <merlin@baltimore.ie>*
18059 * Fix lots of warnings.
18061 *Richard Levitte <levitte@stacken.kth.se>*
18063 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18064 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18066 *Richard Levitte <levitte@stacken.kth.se>*
18068 * Fix problems with sizeof(long) == 8.
18070 *Andy Polyakov <appro@fy.chalmers.se>*
18072 * Change functions to ANSI C.
18076 * Fix typos in error codes.
18078 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18080 * Remove defunct assembler files from Configure.
18084 * SPARC v8 assembler BIGNUM implementation.
18086 *Andy Polyakov <appro@fy.chalmers.se>*
18088 * Support for Certificate Policies extension: both print and set.
18089 Various additions to support the r2i method this uses.
18093 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18094 return a const string when you are expecting an allocated buffer.
18098 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18099 types DirectoryString and DisplayText.
18103 * Add code to allow r2i extensions to access the configuration database,
18104 add an LHASH database driver and add several ctx helper functions.
18108 * Fix an evil bug in bn_expand2() which caused various BN functions to
18109 fail when they extended the size of a BIGNUM.
18113 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18114 support typesafe stack.
18118 * Fix typo in SSL_[gs]et_options().
18120 *Nils Frostberg <nils@medcom.se>*
18122 * Delete various functions and files that belonged to the (now obsolete)
18123 old X509V3 handling code.
18127 * New Configure option "rsaref".
18131 * Don't auto-generate pem.h.
18135 * Introduce type-safe ASN.1 SETs.
18139 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18141 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18143 * Introduce type-safe STACKs. This will almost certainly break lots of code
18144 that links with OpenSSL (well at least cause lots of warnings), but fear
18145 not: the conversion is trivial, and it eliminates loads of evil casts. A
18146 few STACKed things have been converted already. Feel free to convert more.
18147 In the fullness of time, I'll do away with the STACK type altogether.
18151 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18152 specified in `<certfile>` by updating the entry in the index.txt file.
18153 This way one no longer has to edit the index.txt file manually for
18154 revoking a certificate. The -revoke option does the gory details now.
18156 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18158 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18159 `-text` option at all and this way the `-noout -text` combination was
18160 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18162 *Ralf S. Engelschall*
18164 * Make sure a corresponding plain text error message exists for the
18165 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18166 verify callback function determined that a certificate was revoked.
18168 *Ralf S. Engelschall*
18170 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18171 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18172 all available ciphers including rc5, which was forgotten until now.
18173 In order to let the testing shell script know which algorithms
18174 are available, a new (up to now undocumented) command
18175 `openssl list-cipher-commands` is used.
18179 * Bugfix: s_client occasionally would sleep in select() when
18180 it should have checked SSL_pending() first.
18184 * New functions DSA_do_sign and DSA_do_verify to provide access to
18185 the raw DSA values prior to ASN.1 encoding.
18189 * Tweaks to Configure
18191 *Niels Poppe <niels@netbox.org>*
18193 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18198 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18202 * New config option to avoid instructions that are illegal on the 80386.
18203 The default code is faster, but requires at least a 486.
18207 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18208 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18209 same as SSL2_VERSION anyway.
18213 * New "-showcerts" option for s_client.
18217 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18218 application. Various cleanups and fixes.
18222 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18223 modify error routines to work internally. Add error codes and PBE init
18224 to library startup routines.
18228 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18229 packing functions to asn1 and evp. Changed function names and error
18230 codes along the way.
18234 * PKCS12 integration: and so it begins... First of several patches to
18235 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18236 objects to objects.h
18240 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18241 and display support for Thawte strong extranet extension.
18245 * Add LinuxPPC support.
18247 *Jeff Dubrule <igor@pobox.org>*
18249 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18250 bn_div_words in alpha.s.
18252 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18254 * Make sure the RSA OAEP test is skipped under -DRSAref because
18255 OAEP isn't supported when OpenSSL is built with RSAref.
18257 *Ulf Moeller <ulf@fitug.de>*
18259 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18260 so they no longer are missing under -DNOPROTO.
18262 *Soren S. Jorvang <soren@t.dk>*
18264 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18266 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18267 doesn't work when the session is reused. Coming soon!
18271 * Fix a security hole, that allows sessions to be reused in the wrong
18272 context thus bypassing client cert protection! All software that uses
18273 client certs and session caches in multiple contexts NEEDS PATCHING to
18274 allow session reuse! A fuller solution is in the works.
18276 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18278 * Some more source tree cleanups (removed obsolete files
18279 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18280 permission on "config" script to be executable) and a fix for the INSTALL
18283 *Ulf Moeller <ulf@fitug.de>*
18285 * Remove some legacy and erroneous uses of malloc, free instead of
18288 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18290 * Make rsa_oaep_test return non-zero on error.
18292 *Ulf Moeller <ulf@fitug.de>*
18294 * Add support for native Solaris shared libraries. Configure
18295 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18296 if someone would make that last step automatic.
18298 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18300 * ctx_size was not built with the right compiler during "make links". Fixed.
18304 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18305 except NULL ciphers". This means the default cipher list will no longer
18306 enable NULL ciphers. They need to be specifically enabled e.g. with
18307 the string "DEFAULT:eNULL".
18311 * Fix to RSA private encryption routines: if p < q then it would
18312 occasionally produce an invalid result. This will only happen with
18313 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18317 * Be less restrictive and allow also `perl util/perlpath.pl
18318 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18319 because this way one can also use an interpreter named `perl5` (which is
18320 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18321 installed as `perl`).
18323 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18325 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18327 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18329 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18330 advapi32.lib to Win32 build and change the pem test comparison
18331 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18332 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18333 and crypto/des/ede_cbcm_enc.c.
18337 * DES quad checksum was broken on big-endian architectures. Fixed.
18341 * Comment out two functions in bio.h that aren't implemented. Fix up the
18342 Win32 test batch file so it (might) work again. The Win32 test batch file
18343 is horrible: I feel ill....
18347 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18348 in e_os.h. Audit of header files to check ANSI and non ANSI
18349 sections: 10 functions were absent from non ANSI section and not exported
18350 from Windows DLLs. Fixed up libeay.num for new functions.
18354 * Make `openssl version` output lines consistent.
18356 *Ralf S. Engelschall*
18358 * Fix Win32 symbol export lists for BIO functions: Added
18359 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18360 to ms/libeay{16,32}.def.
18362 *Ralf S. Engelschall*
18364 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18365 fine under Unix and passes some trivial tests I've now added. But the
18366 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18367 added to make sure no one expects that this stuff really works in the
18368 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18369 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18372 *Ralf S. Engelschall*
18374 * Fix the generation of two part addresses in perl.
18376 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18378 * Add config entry for Linux on MIPS.
18380 *John Tobey <jtobey@channel1.com>*
18382 * Make links whenever Configure is run, unless we are on Windoze.
18386 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18387 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18392 * Add a useful kludge to allow package maintainers to specify compiler and
18393 other platforms details on the command line without having to patch the
18394 Configure script every time: One now can use
18395 `perl Configure <id>:<details>`,
18396 i.e. platform ids are allowed to have details appended
18397 to them (separated by colons). This is treated as there would be a static
18398 pre-configured entry in Configure's %table under key `<id>` with value
18399 `<details>` and `perl Configure <id>` is called. So, when you want to
18400 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18401 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18402 now, which overrides the FreeBSD-elf entry on-the-fly.
18404 *Ralf S. Engelschall*
18406 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18410 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18411 on the `perl Configure ...` command line. This way one can compile
18412 OpenSSL libraries with Position Independent Code (PIC) which is needed
18413 for linking it into DSOs.
18415 *Ralf S. Engelschall*
18417 * Remarkably, export ciphers were totally broken and no-one had noticed!
18422 * Cleaned up the LICENSE document: The official contact for any license
18423 questions now is the OpenSSL core team under openssl-core@openssl.org.
18424 And add a paragraph about the dual-license situation to make sure people
18425 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18426 to the OpenSSL toolkit.
18428 *Ralf S. Engelschall*
18430 * General source tree makefile cleanups: Made `making xxx in yyy...`
18431 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18432 Additionally cleaned up the `make links` target: Remove unnecessary
18433 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18434 to speed processing and no longer clutter the display with confusing
18435 stuff. Instead only the actually done links are displayed.
18437 *Ralf S. Engelschall*
18439 * Permit null encryption ciphersuites, used for authentication only. It used
18440 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18441 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18446 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18447 signed attributes when verifying signatures (this would break them),
18448 the detached data encoding was wrong and public keys obtained using
18449 X509_get_pubkey() weren't freed.
18453 * Add text documentation for the BUFFER functions. Also added a work around
18454 to a Win95 console bug. This was triggered by the password read stuff: the
18455 last character typed gets carried over to the next fread(). If you were
18456 generating a new cert request using 'req' for example then the last
18457 character of the passphrase would be CR which would then enter the first
18462 * Added the new 'Includes OpenSSL Cryptography Software' button as
18463 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18464 button and can be used by applications based on OpenSSL to show the
18465 relationship to the OpenSSL project.
18467 *Ralf S. Engelschall*
18469 * Remove confusing variables in function signatures in files
18470 ssl/ssl_lib.c and ssl/ssl.h.
18472 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18474 * Don't install bss_file.c under PREFIX/include/
18476 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18478 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18479 functions that return function pointers and has support for NT specific
18480 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18481 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18482 unsigned to signed types: this was killing the Win32 compile.
18486 * Add new certificate file to stack functions,
18487 SSL_add_dir_cert_subjects_to_stack() and
18488 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18489 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18490 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18491 This means that Apache-SSL and similar packages don't have to mess around
18492 to add as many CAs as they want to the preferred list.
18496 * Experiment with doxygen documentation. Currently only partially applied to
18498 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18499 openssl.doxy as the configuration file.
18503 * Get rid of remaining C++-style comments which strict C compilers hate.
18505 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18507 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18508 compiled in by default: it has problems with large keys.
18512 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18513 DH private keys and/or callback functions which directly correspond to
18514 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18515 is needed for applications which have to configure certificates on a
18516 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18518 For the RSA certificate situation is makes no difference, but
18519 for the DSA certificate situation this fixes the "no shared cipher"
18520 problem where the OpenSSL cipher selection procedure failed because the
18521 temporary keys were not overtaken from the context and the API provided
18522 no way to reconfigure them.
18523 The new functions now let applications reconfigure the stuff and they
18524 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18525 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18526 non-public-API function ssl_cert_instantiate() is used as a helper
18527 function and also to reduce code redundancy inside ssl_rsa.c.
18529 *Ralf S. Engelschall*
18531 * Move s_server -dcert and -dkey options out of the undocumented feature
18532 area because they are useful for the DSA situation and should be
18533 recognized by the users.
18535 *Ralf S. Engelschall*
18537 * Fix the cipher decision scheme for export ciphers: the export bits are
18538 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18539 SSL_EXP_MASK. So, the original variable has to be used instead of the
18540 already masked variable.
18542 *Richard Levitte <levitte@stacken.kth.se>*
18544 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18546 *Richard Levitte <levitte@stacken.kth.se>*
18548 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18549 from `int` to `unsigned int` because it is a length and initialized by
18550 EVP_DigestFinal() which expects an `unsigned int *`.
18552 *Richard Levitte <levitte@stacken.kth.se>*
18554 * Don't hard-code path to Perl interpreter on shebang line of Configure
18555 script. Instead use the usual Shell->Perl transition trick.
18557 *Ralf S. Engelschall*
18559 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18560 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18561 -noout -modulus` as it's already the case for `openssl rsa -noout
18562 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18563 currently the public key is printed (a decision which was already done by
18564 `openssl dsa -modulus` in the past) which serves a similar purpose.
18565 Additionally the NO_RSA no longer completely removes the whole -modulus
18566 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18569 *Ralf S. Engelschall*
18571 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18572 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18574 *Arne Ansper <arne@ats.cyber.ee>*
18576 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18577 to be added. Now both 'req' and 'ca' can use new objects defined in the
18582 * Add cool BIO that does syslog (or event log on NT).
18584 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18586 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18587 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18588 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18589 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18593 * Add preliminary config info for new extension code.
18597 * Make RSA_NO_PADDING really use no padding.
18599 *Ulf Moeller <ulf@fitug.de>*
18601 * Generate errors when private/public key check is done.
18605 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18606 for some CRL extensions and new objects added.
18610 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18611 key usage extension and fuller support for authority key id.
18615 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18616 padding method for RSA, which is recommended for new applications in PKCS
18617 #1 v2.0 (RFC 2437, October 1998).
18618 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18619 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18620 against Bleichbacher's attack on RSA.
18621 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18624 * Updates to the new SSL compression code
18626 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18628 * Fix so that the version number in the master secret, when passed
18629 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18630 (because the server will not accept higher), that the version number
18631 is 0x03,0x01, not 0x03,0x00
18633 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18635 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18636 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18637 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18641 * Support for RAW extensions where an arbitrary extension can be
18642 created by including its DER encoding. See `apps/openssl.cnf` for
18647 * Make sure latest Perl versions don't interpret some generated C array
18648 code as Perl array code in the crypto/err/err_genc.pl script.
18650 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18652 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18653 not many people have the assembler. Various Win32 compilation fixes and
18654 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18655 build instructions.
18659 * Modify configure script 'Configure' to automatically create crypto/date.h
18660 file under Win32 and also build pem.h from pem.org. New script
18661 util/mkfiles.pl to create the MINFO file on environments that can't do a
18662 'make files': perl util/mkfiles.pl >MINFO should work.
18666 * Major rework of DES function declarations, in the pursuit of correctness
18667 and purity. As a result, many evil casts evaporated, and some weirdness,
18668 too. You may find this causes warnings in your code. Zapping your evil
18669 casts will probably fix them. Mostly.
18673 * Fix for a typo in asn1.h. Bug fix to object creation script
18674 obj_dat.pl. It considered a zero in an object definition to mean
18675 "end of object": none of the objects in objects.h have any zeros
18676 so it wasn't spotted.
18678 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18680 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18681 Masking (CBCM). In the absence of test vectors, the best I have been able
18682 to do is check that the decrypt undoes the encrypt, so far. Send me test
18683 vectors if you have them.
18687 * Correct calculation of key length for export ciphers (too much space was
18688 allocated for null ciphers). This has not been tested!
18692 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18693 message is now correct (it understands "crypto" and "ssl" on its
18694 command line). There is also now an "update" option. This will update
18695 the util/ssleay.num and util/libeay.num files with any new functions.
18697 perl util/mkdef.pl crypto ssl update
18698 it will update them.
18702 * Overhauled the Perl interface:
18703 - ported BN stuff to OpenSSL's different BN library
18704 - made the perl/ source tree CVS-aware
18705 - renamed the package from SSLeay to OpenSSL (the files still contain
18706 their history because I've copied them in the repository)
18707 - removed obsolete files (the test scripts will be replaced
18708 by better Test::Harness variants in the future)
18710 *Ralf S. Engelschall*
18712 * First cut for a very conservative source tree cleanup:
18713 1. merge various obsolete readme texts into doc/ssleay.txt
18714 where we collect the old documents and readme texts.
18715 2. remove the first part of files where I'm already sure that we no
18716 longer need them because of three reasons: either they are just temporary
18717 files which were left by Eric or they are preserved original files where
18718 I've verified that the diff is also available in the CVS via "cvs diff
18719 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18720 the crypto/md/ stuff).
18722 *Ralf S. Engelschall*
18724 * More extension code. Incomplete support for subject and issuer alt
18725 name, issuer and authority key id. Change the i2v function parameters
18726 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18727 what that's for :-) Fix to ASN1 macro which messed up
18728 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18732 * Preliminary support for ENUMERATED type. This is largely copied from the
18737 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18739 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18741 * Make sure `make rehash` target really finds the `openssl` program.
18743 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18745 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18746 like to hear about it if this slows down other processors.
18750 * Add CygWin32 platform information to Configure script.
18752 *Alan Batie <batie@aahz.jf.intel.com>*
18754 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18756 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18758 * New program nseq to manipulate netscape certificate sequences
18762 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18767 * Fixes to BN code. Previously the default was to define BN_RECURSION
18768 but the BN code had some problems that would cause failures when
18769 doing certificate verification and some other functions.
18771 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18773 * Add ASN1 and PEM code to support netscape certificate sequences.
18777 * Add ASN1 and PEM code to support netscape certificate sequences.
18781 * Add several PKIX and private extended key usage OIDs.
18785 * Modify the 'ca' program to handle the new extension code. Modify
18786 openssl.cnf for new extension format, add comments.
18790 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18791 and add a sample to openssl.cnf so req -x509 now adds appropriate
18796 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18797 error code, add initial support to X509_print() and x509 application.
18801 * Takes a deep breath and start adding X509 V3 extension support code. Add
18802 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18803 stuff is currently isolated and isn't even compiled yet.
18807 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18808 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18809 Removed the versions check from X509 routines when loading extensions:
18810 this allows certain broken certificates that don't set the version
18811 properly to be processed.
18815 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18816 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18817 can still be regenerated with "make depend".
18821 * Spelling mistake in C version of CAST-128.
18823 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18825 * Changes to the error generation code. The perl script err-code.pl
18826 now reads in the old error codes and retains the old numbers, only
18827 adding new ones if necessary. It also only changes the .err files if new
18828 codes are added. The makefiles have been modified to only insert errors
18829 when needed (to avoid needlessly modifying header files). This is done
18830 by only inserting errors if the .err file is newer than the auto generated
18831 C file. To rebuild all the error codes from scratch (the old behaviour)
18832 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18833 or delete all the .err files.
18837 * CAST-128 was incorrectly implemented for short keys. The C version has
18838 been fixed, but is untested. The assembler versions are also fixed, but
18839 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18840 to regenerate it if needed.
18841 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18842 Hagino <itojun@kame.net>*
18844 * File was opened incorrectly in randfile.c.
18846 *Ulf Möller <ulf@fitug.de>*
18848 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18849 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18850 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18851 al: it's just almost always a UTCTime. Note this patch adds new error
18852 codes so do a "make errors" if there are problems.
18856 * Correct Linux 1 recognition in config.
18858 *Ulf Möller <ulf@fitug.de>*
18860 * Remove pointless MD5 hash when using DSA keys in ca.
18862 *Anonymous <nobody@replay.com>*
18864 * Generate an error if given an empty string as a cert directory. Also
18865 generate an error if handed NULL (previously returned 0 to indicate an
18866 error, but didn't set one).
18868 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18870 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18874 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18875 parameters. This was causing a warning which killed off the Win32 compile.
18879 * Remove C++ style comments from crypto/bn/bn_local.h.
18881 *Neil Costigan <neil.costigan@celocom.com>*
18883 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18884 based on a text string, looking up short and long names and finally
18885 "dot" format. The "dot" format stuff didn't work. Added new function
18886 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18887 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18888 OID is not part of the table.
18892 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18893 X509_LOOKUP_by_alias().
18897 * Sort openssl functions by name.
18901 * Get the `gendsa` command working and add it to the `list` command. Remove
18902 encryption from sample DSA keys (in case anyone is interested the password
18907 * Make *all* `*_free` functions accept a NULL pointer.
18909 *Frans Heymans <fheymans@isaserver.be>*
18911 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18914 *Anonymous <nobody@replay.com>*
18916 * s_server should send the CAfile as acceptable CAs, not its own cert.
18918 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18920 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18922 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18924 * Temp key "for export" tests were wrong in s3_srvr.c.
18926 *Anonymous <nobody@replay.com>*
18928 * Add prototype for temp key callback functions
18929 SSL_CTX_set_tmp_{rsa,dh}_callback().
18933 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18934 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18938 * X509_name_add_entry() freed the wrong thing after an error.
18940 *Arne Ansper <arne@ats.cyber.ee>*
18942 * rsa_eay.c would attempt to free a NULL context.
18944 *Arne Ansper <arne@ats.cyber.ee>*
18946 * BIO_s_socket() had a broken should_retry() on Windoze.
18948 *Arne Ansper <arne@ats.cyber.ee>*
18950 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18952 *Arne Ansper <arne@ats.cyber.ee>*
18954 * Make sure the already existing X509_STORE->depth variable is initialized
18955 in X509_STORE_new(), but document the fact that this variable is still
18956 unused in the certificate verification process.
18958 *Ralf S. Engelschall*
18960 * Fix the various library and `apps/` files to free up pkeys obtained from
18961 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18965 * Fix reference counting in X509_PUBKEY_get(). This makes
18966 demos/maurice/example2.c work, amongst others, probably.
18968 *Steve Henson and Ben Laurie*
18970 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18971 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18972 are no longer created. This way we have a single and consistent command
18973 line interface `openssl <command>`, similar to `cvs <command>`.
18975 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18977 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18978 BIT STRING wrapper always have zero unused bits.
18982 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18986 * Make the top-level INSTALL documentation easier to understand.
18990 * Makefiles updated to exit if an error occurs in a sub-directory
18991 make (including if user presses ^C) [Paul Sutton]
18993 * Make Montgomery context stuff explicit in RSA data structure.
18997 * Fix build order of pem and err to allow for generated pem.h.
19001 * Fix renumbering bug in X509_NAME_delete_entry().
19005 * Enhanced the err-ins.pl script so it makes the error library number
19006 global and can add a library name. This is needed for external ASN1 and
19007 other error libraries.
19011 * Fixed sk_insert which never worked properly.
19015 * Fix ASN1 macros so they can handle indefinite length constructed
19016 EXPLICIT tags. Some non standard certificates use these: they can now
19021 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19022 into a single doc/ssleay.txt bundle. This way the information is still
19023 preserved but no longer messes up this directory. Now it's new room for
19024 the new set of documentation files.
19026 *Ralf S. Engelschall*
19028 * SETs were incorrectly DER encoded. This was a major pain, because they
19029 shared code with SEQUENCEs, which aren't coded the same. This means that
19030 almost everything to do with SETs or SEQUENCEs has either changed name or
19031 number of arguments.
19033 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19035 * Fix test data to work with the above.
19039 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19040 was already fixed by Eric for 0.9.1 it seems.
19042 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19044 * Autodetect FreeBSD3.
19048 * Fix various bugs in Configure. This affects the following platforms:
19052 unixware-2.0-pentium
19057 * Eliminate generated files from CVS. Reorder tests to regenerate files
19058 before they are needed.
19062 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19066 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19068 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19069 changed SSLeay to OpenSSL in version strings.
19071 *Ralf S. Engelschall*
19073 * Some fixups to the top-level documents.
19077 * Fixed the nasty bug where rsaref.h was not found under compile-time
19078 because the symlink to include/ was missing.
19080 *Ralf S. Engelschall*
19082 * Incorporated the popular no-RSA/DSA-only patches
19083 which allow to compile a RSA-free SSLeay.
19085 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19087 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19088 when "ssleay" is still not found.
19090 *Ralf S. Engelschall*
19092 * Added more platforms to Configure: Cray T3E, HPUX 11,
19094 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19096 * Updated the README file.
19098 *Ralf S. Engelschall*
19100 * Added various .cvsignore files in the CVS repository subdirs
19101 to make a "cvs update" really silent.
19103 *Ralf S. Engelschall*
19105 * Recompiled the error-definition header files and added
19106 missing symbols to the Win32 linker tables.
19108 *Ralf S. Engelschall*
19110 * Cleaned up the top-level documents;
19111 o new files: CHANGES and LICENSE
19112 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19113 o merged COPYRIGHT into LICENSE
19114 o removed obsolete TODO file
19115 o renamed MICROSOFT to INSTALL.W32
19117 *Ralf S. Engelschall*
19119 * Removed dummy files from the 0.9.1b source tree:
19121 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19122 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19123 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19124 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19125 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19128 *Ralf S. Engelschall*
19130 * Added various platform portability fixes.
19134 * The Genesis of the OpenSSL rpject:
19135 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19136 Young and Tim J. Hudson created while they were working for C2Net until
19139 *The OpenSSL Project*
19141 ### Changes between 0.9.0b and 0.9.1b [not released]
19143 * Updated a few CA certificates under certs/
19147 * Changed some BIGNUM api stuff.
19151 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19152 DGUX x86, Linux Alpha, etc.
19156 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19157 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19162 * Add -strparse option to asn1pars program which parses nested
19165 *Dr Stephen Henson <shenson@bigfoot.com>*
19167 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19171 * DSA fix for "ca" program.
19175 * Added "-genkey" option to "dsaparam" program.
19179 * Added RIPE MD160 (rmd160) message digest.
19183 * Added -a (all) option to "ssleay version" command.
19187 * Added PLATFORM define which is the id given to Configure.
19191 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19195 * Extended the ASN.1 parser routines.
19199 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19203 * Added a BN_CTX to the BN library.
19207 * Fixed the weak key values in DES library
19211 * Changed API in EVP library for cipher aliases.
19215 * Added support for RC2/64bit cipher.
19219 * Converted the lhash library to the crypto/mem.c functions.
19223 * Added more recognized ASN.1 object ids.
19227 * Added more RSA padding checks for SSL/TLS.
19231 * Added BIO proxy/filter functionality.
19235 * Added extra_certs to SSL_CTX which can be used
19236 send extra CA certificates to the client in the CA cert chain sending
19237 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19241 * Now Fortezza is denied in the authentication phase because
19242 this is key exchange mechanism is not supported by SSLeay at all.
19246 * Additional PKCS1 checks.
19250 * Support the string "TLSv1" for all TLS v1 ciphers.
19254 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19255 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19259 * Fixed a few memory leaks.
19263 * Fixed various code and comment typos.
19267 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19268 bytes sent in the client random.
19270 *Edward Bishop <ebishop@spyglass.com>*
19274 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19275 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19276 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19277 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19278 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19279 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19280 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19281 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19282 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19283 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19284 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19285 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19286 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19287 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19288 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19289 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19290 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19291 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19292 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19293 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19294 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19295 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19296 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19297 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19298 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19299 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19300 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19301 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19302 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19303 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19304 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19305 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19306 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19307 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19308 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19309 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19310 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19311 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19312 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19313 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19314 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19315 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19316 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19317 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19318 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19319 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19320 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19321 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19322 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19323 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19324 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19325 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19326 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19327 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19328 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19329 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19330 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19331 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19332 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19333 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19334 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19335 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19336 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19337 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19338 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19339 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19340 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19341 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19342 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19343 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19344 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19345 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19346 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19347 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19348 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19349 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19350 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19351 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19352 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19353 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19354 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19355 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19356 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19357 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19358 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19359 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19360 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19361 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19362 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19363 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19364 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19365 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19366 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19367 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19368 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19369 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19370 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19371 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19372 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19373 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19374 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19375 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19376 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19377 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19378 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19379 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19380 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19381 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19382 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19383 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19384 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19385 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19386 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19387 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19388 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19389 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19390 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19391 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19392 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19393 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19394 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19395 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19396 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19397 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19398 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19399 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19400 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19401 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19402 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19403 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19404 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19405 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19406 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19407 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19408 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19409 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19410 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19411 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19412 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19413 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19414 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19415 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19416 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19417 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19418 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19419 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19420 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19421 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19422 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19423 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19424 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19425 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19426 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19427 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19428 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19429 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19430 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19431 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19432 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19433 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19434 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19435 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655