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 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
27 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
28 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
29 will need to load the legacy crypto provider.
33 * The various OBJ_* functions have been made thread safe.
37 * CCM8 cipher suites in TLS have been downgraded to security level zero
38 because they use a short authentication tag which lowers their strength.
42 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
50 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
51 listed here are only a brief description.
52 The migration guide contains more detailed information related to new features,
53 breaking changes, and mappings for the large list of deprecated functions.
55 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
57 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
59 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
64 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
65 S390X capability vector to zero. This simplifies testing of different code
66 paths on S390X architecture.
70 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
71 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
72 SP 800-38D". The communication will fail at this point.
76 * The EC_GROUP_clear_free() function is deprecated as there is nothing
77 confidential in EC_GROUP data.
81 * The byte order mark (BOM) character is ignored if encountered at the
82 beginning of a PEM-formatted file.
86 * Added CMS support for the Russian GOST algorithms.
90 * Due to move of the implementation of cryptographic operations
91 to the providers, validation of various operation parameters can
92 be postponed until the actual operation is executed where previously
93 it happened immediately when an operation parameter was set.
95 For example when setting an unsupported curve with
96 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
97 fail but later keygen operations with the EVP_PKEY_CTX will fail.
99 *OpenSSL team members and many third party contributors*
101 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
102 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
103 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
104 instead to retrieve these algorithms from a provider.
108 * On build targets where the multilib postfix is set in the build
109 configuration the libdir directory was changing based on whether
110 the lib directory with the multilib postfix exists on the system
111 or not. This unpredictable behavior was removed and eventual
112 multilib postfix is now always added to the default libdir. Use
113 `--libdir=lib` to override the libdir if adding the postfix is
118 * The triple DES key wrap functionality now conforms to RFC 3217 but is
119 no longer interoperable with OpenSSL 1.1.1.
123 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
124 function codes, this function can only cause problems for calling
129 * Add a configurable flag to output date formats as ISO 8601. Does not
130 change the default date format.
134 * Version of MSVC earlier than 1300 could get link warnings, which could
135 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
136 Support for this flag has been removed.
140 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
141 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
142 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
143 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
144 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
148 * The signatures of the functions to get and set options on SSL and
149 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
150 Some source code changes may be required.
154 * The public definitions of conf_method_st and conf_st have been
155 deprecated. They will be made opaque in a future release.
157 *Rich Salz and Tomáš Mráz*
159 * Client-initiated renegotiation is disabled by default. To allow it, use
160 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
161 flag, or the "ClientRenegotiation" config parameter as appropriate.
165 * Add "abspath" and "includedir" pragma's to config files, to prevent,
166 or modify relative pathname inclusion.
170 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
171 validated. Please consult the README-FIPS and
172 README-PROVIDERS files, as well as the migration guide.
174 *OpenSSL team members and many third party contributors*
176 * For the key types DH and DHX the allowed settable parameters are now different.
180 * The openssl commands that read keys, certificates, and CRLs now
181 automatically detect the PEM or DER format of the input files.
183 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
185 * Added enhanced PKCS#12 APIs which accept a library context.
189 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
193 * Added support for Kernel TLS (KTLS).
195 *Boris Pismenny, John Baldwin and Andrew Gallatin*
197 * Support for RFC 5746 secure renegotiation is now required by default for
198 SSL or TLS connections to succeed.
202 * The signature of the `copy` functional parameter of the
203 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
204 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
205 the signature of the `pub_decode` functional parameter of the
206 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
207 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
211 * The error return values from some control calls (ctrl) have changed.
215 * A public key check is now performed during EVP_PKEY_derive_set_peer().
219 * Many functions in the EVP_ namespace that are getters of values from
220 implementations or contexts were renamed to include get or get0 in their
221 names. Old names are provided as macro aliases for compatibility and
226 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
227 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
228 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
233 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
236 * The output from the command line applications may have minor
241 * The output from numerous "printing" may have minor changes.
245 * Windows thread synchronization uses read/write primitives (SRWLock) when
246 supported by the OS, otherwise CriticalSection continues to be used.
250 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
251 work on read only BIO source/sinks that do not support these functions.
252 This allows piping or redirection of a file BIO using stdin to be buffered
253 into memory. This is used internally in OSSL_DECODER_from_bio().
257 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
258 this function would return one of the values OSSL_STORE_INFO_NAME,
259 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
260 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
261 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
262 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
263 using this function should be amended to handle the changed return value.
267 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
268 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
269 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
270 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
271 contains more than one certificate identifier: This means that all
272 certificates referenced there MUST be part of the validation chain.
276 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
277 RC5, DESX and DES have been moved to the legacy provider.
281 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
282 RIPEMD-160 have been moved to the legacy provider.
286 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
291 * The deprecated functions EVP_PKEY_get0_RSA(),
292 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
293 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
294 well as the similarly named "get1" functions behave differently in
299 * A number of functions handling low-level keys or engines were deprecated
300 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
301 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
302 EVP_PKEY_get0_siphash().
306 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
307 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
308 will need to load the legacy crypto provider. This includes these PBE
309 algorithms which use this KDF:
310 - NID_pbeWithMD2AndDES_CBC
311 - NID_pbeWithMD5AndDES_CBC
312 - NID_pbeWithSHA1AndRC2_CBC
313 - NID_pbeWithMD2AndRC2_CBC
314 - NID_pbeWithMD5AndRC2_CBC
315 - NID_pbeWithSHA1AndDES_CBC
319 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
320 BIO_debug_callback() functions.
324 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
325 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
329 * The RAND_METHOD APIs have been deprecated.
333 * The SRP APIs have been deprecated.
337 * Add a compile time option to prevent the caching of provider fetched
338 algorithms. This is enabled by including the no-cached-fetch option
339 at configuration time.
343 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
344 count of PKCS12_DEFAULT_ITER.
346 *Tomáš Mráz and Sahana Prasad*
348 * The openssl speed command does not use low-level API calls anymore.
352 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
355 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
357 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
361 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
362 providers may supply their own group implementations (using either the "key
363 exchange" or the "key encapsulation" methods) which will automatically be
364 detected and used by libssl.
366 *Matt Caswell, Nicola Tuveri*
368 * The undocumented function X509_certificate_type() has been deprecated;
372 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
376 * Removed RSA padding mode for SSLv23 (which was only used for
377 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
378 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
383 * Deprecated the obsolete X9.31 RSA key generation related functions.
385 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
386 is not allowed to return a value > 1, this is no more taken as failure.
388 *Viktor Dukhovni and David von Oheimb*
390 * Deprecated the obsolete X9.31 RSA key generation related functions
391 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
392 BN_X931_generate_prime_ex().
396 * The default key generation method for the regular 2-prime RSA keys was
397 changed to the FIPS 186-4 B.3.6 method.
401 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
405 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
409 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
410 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
412 *Rich Salz, Richard Levitte, and David von Oheimb*
414 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
418 * Deprecated `OCSP_parse_url()`.
422 * Validation of SM2 keys has been separated from the validation of regular EC
427 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
428 switches: a validation failure triggers an early exit, returning a failure
429 exit status to the parent process.
433 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
434 to ignore unknown ciphers.
438 * The `-cipher-commands` and `-digest-commands` options
439 of the command line utility `list` have been deprecated.
440 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
444 * Added convenience functions for generating asymmetric key pairs:
445 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
446 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
450 * All of the low level EC_KEY functions have been deprecated.
452 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
454 * Deprecated all the libcrypto and libssl error string loading
459 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
460 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
465 * The `-crypt` option to the `passwd` command line tool has been removed.
469 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
474 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
478 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
479 EVP_PKEY_get1_tls_encodedpoint().
483 * The security callback, which can be customised by application code, supports
484 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
485 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
489 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
490 interface. Their functionality remains unchanged.
494 * Added new option for 'openssl list', '-providers', which will display the
495 list of loaded providers, their names, version and status. It optionally
496 displays their gettable parameters.
500 * Removed EVP_PKEY_set_alias_type().
504 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
505 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
509 * Changed all "STACK" functions to be macros instead of inline functions. Macro
510 parameters are still checked for type safety at compile time via helper
515 * Remove the RAND_DRBG API
517 *Paul Dale and Matthias St. Pierre*
519 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
520 as well as actual hostnames.
524 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
525 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
526 conversely, silently ignore DTLS protocol version bounds when configuring
527 TLS-based contexts. The commands can be repeated to set bounds of both
528 types. The same applies with the corresponding "min_protocol" and
529 "max_protocol" command-line switches, in case some application uses both TLS
532 SSL_CTX instances that are created for a fixed protocol version (e.g.
533 `TLSv1_server_method()`) also silently ignore version bounds. Previously
534 attempts to apply bounds to these protocol versions would result in an
535 error. Now only the "version-flexible" SSL_CTX instances are subject to
536 limits in configuration files in command-line options.
540 * Deprecated the `ENGINE` API. Engines should be replaced with providers
545 * Reworked the recorded ERR codes to make better space for system errors.
546 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
547 given code is a system error (true) or an OpenSSL error (false).
551 * Reworked the test perl framework to better allow parallel testing.
553 *Nicola Tuveri and David von Oheimb*
555 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
556 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
560 * 'Configure' has been changed to figure out the configuration target if
561 none is given on the command line. Consequently, the 'config' script is
562 now only a mere wrapper. All documentation is changed to only mention
565 *Rich Salz and Richard Levitte*
567 * Added a library context `OSSL_LIB_CTX` that applications as well as
568 other libraries can use to form a separate context within which
569 libcrypto operations are performed.
573 * Added various `_ex` functions to the OpenSSL API that support using
574 a non-default `OSSL_LIB_CTX`.
578 * Handshake now fails if Extended Master Secret extension is dropped
583 * Dropped interactive mode from the `openssl` program.
587 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
589 *David von Oheimb and Shane Lontis*
591 * Deprecated `EC_METHOD_get_field_type()`.
595 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
596 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
597 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
601 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
605 * Add CAdES-BES signature verification support, mostly derived
606 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
608 *Filipe Raimundo da Silva*
610 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
614 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
615 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
619 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
623 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
624 EC_KEY_precompute_mult().
628 * Deprecated EC_POINTs_mul().
632 * Removed FIPS_mode() and FIPS_mode_set().
636 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
640 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
641 EC_POINT_get_Jprojective_coordinates_GFp().
645 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
646 arrays to be more easily constructed via a series of utility functions.
647 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
648 the various push functions and finally convert to a passable OSSL_PARAM
649 array using OSSL_PARAM_BLD_to_param().
653 * The security strength of SHA1 and MD5 based signatures in TLS has been
658 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
659 contain a provider side internal key.
663 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
667 * Project text documents not yet having a proper file name extension
668 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
669 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
673 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
674 have been converted to Markdown with the goal to produce documents
675 which not only look pretty when viewed online in the browser, but
676 remain well readable inside a plain text editor.
678 To achieve this goal, a 'minimalistic' Markdown style has been applied
679 which avoids formatting elements that interfere too much with the
680 reading flow in the text file. For example, it
682 * avoids [ATX headings][] and uses [setext headings][] instead
683 (which works for `<h1>` and `<h2>` headings only).
684 * avoids [inline links][] and uses [reference links][] instead.
685 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
687 [ATX headings]: https://github.github.com/gfm/#atx-headings
688 [setext headings]: https://github.github.com/gfm/#setext-headings
689 [inline links]: https://github.github.com/gfm/#inline-link
690 [reference links]: https://github.github.com/gfm/#reference-link
691 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
692 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
694 *Matthias St. Pierre*
696 * The test suite is changed to preserve results of each test recipe.
697 A new directory test-runs/ with subdirectories named like the
698 test recipes are created in the build tree for this purpose.
702 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
703 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
704 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
706 *David von Oheimb, Martin Peylo*
708 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
709 It supports arbitrary request and response content types, GET redirection,
710 TLS, connections via HTTP(S) proxies, connections and exchange via
711 user-defined BIOs (allowing implicit connections), persistent connections,
712 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
713 The legacy OCSP-focused (and only partly documented) API
714 is retained for backward compatibility, while most of it is deprecated.
718 * Added `util/check-format.pl`, a tool for checking adherence to the
719 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
720 The checks performed are incomplete and yield some false positives.
721 Still the tool should be useful for detecting most typical glitches.
725 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
726 If domain name resolution yields multiple IP addresses all of them are tried
727 after `connect()` failures.
731 * All of the low level RSA functions have been deprecated.
735 * X509 certificates signed using SHA1 are no longer allowed at security
740 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
741 modified to use PKEY APIs. These commands are now in maintenance mode
742 and no new features will be added to them.
746 * The command line utility rsautl has been deprecated.
750 * The command line utilities genrsa and rsa have been modified to use PKEY
751 APIs. They now write PKCS#8 keys by default. These commands are now in
752 maintenance mode and no new features will be added to them.
756 * All of the low level DH functions have been deprecated.
758 *Paul Dale and Matt Caswell*
760 * All of the low level DSA functions have been deprecated.
764 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
765 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
769 * Deprecated low level ECDH and ECDSA functions.
773 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
777 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
778 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
779 a new formulation to include all the things it can be used for,
780 as well as words of caution.
784 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
788 * All of the low level HMAC functions have been deprecated.
790 *Paul Dale and David von Oheimb*
792 * Over two thousand fixes were made to the documentation, including:
793 - Common options (such as -rand/-writerand, TLS version control, etc)
794 were refactored and point to newly-enhanced descriptions in openssl.pod.
795 - Added style conformance for all options (with help from Richard Levitte),
796 documented all reported missing options, added a CI build to check
797 that all options are documented and that no unimplemented options
799 - Documented some internals, such as all use of environment variables.
800 - Addressed all internal broken L<> references.
804 * All of the low level CMAC functions have been deprecated.
808 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
809 functions have been deprecated.
811 *Paul Dale and David von Oheimb*
813 * Corrected the documentation of the return values from the `EVP_DigestSign*`
814 set of functions. The documentation mentioned negative values for some
815 errors, but this was never the case, so the mention of negative values
818 Code that followed the documentation and thereby check with something
819 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
823 * All of the low level cipher functions have been deprecated.
825 *Matt Caswell and Paul Dale*
827 * Removed include/openssl/opensslconf.h.in and replaced it with
828 include/openssl/configuration.h.in, which differs in not including
829 <openssl/macros.h>. A short header include/openssl/opensslconf.h
830 was added to include both.
832 This allows internal hacks where one might need to modify the set
833 of configured macros, for example this if deprecated symbols are
834 still supposed to be available internally:
836 #include <openssl/configuration.h>
838 #undef OPENSSL_NO_DEPRECATED
839 #define OPENSSL_SUPPRESS_DEPRECATED
841 #include <openssl/macros.h>
843 This should not be used by applications that use the exported
844 symbols, as that will lead to linking errors.
848 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
849 used in exponentiation with 512-bit moduli. No EC algorithms are
850 affected. Analysis suggests that attacks against 2-prime RSA1024,
851 3-prime RSA1536, and DSA1024 as a result of this defect would be very
852 difficult to perform and are not believed likely. Attacks against DH512
853 are considered just feasible. However, for an attack the target would
854 have to re-use the DH512 private key, which is not recommended anyway.
855 Also applications directly using the low-level API BN_mod_exp may be
856 affected if they use BN_FLG_CONSTTIME.
861 * Most memory-debug features have been deprecated, and the functionality
862 replaced with no-ops.
866 * Added documentation for the STACK API.
870 * Introduced a new method type and API, OSSL_ENCODER, to represent
871 generic encoders. These do the same sort of job that PEM writers
872 and d2i functions do, but with support for methods supplied by
873 providers, and the possibility for providers to support other
878 * Introduced a new method type and API, OSSL_DECODER, to represent
879 generic decoders. These do the same sort of job that PEM readers
880 and i2d functions do, but with support for methods supplied by
881 providers, and the possibility for providers to support other
886 * Added a .pragma directive to the syntax of configuration files, to
887 allow varying behavior in a supported and predictable manner.
888 Currently added pragma:
892 This allows dollar signs to be a keyword character unless it's
893 followed by a opening brace or parenthesis. This is useful for
894 platforms where dollar signs are commonly used in names, such as
895 volume names and system directory names on VMS.
899 * Added functionality to create an EVP_PKEY from user data.
903 * Change the interpretation of the '--api' configuration option to
904 mean that this is a desired API compatibility level with no
905 further meaning. The previous interpretation, that this would
906 also mean to remove all deprecated symbols up to and including
907 the given version, no requires that 'no-deprecated' is also used
908 in the configuration.
910 When building applications, the desired API compatibility level
911 can be set with the OPENSSL_API_COMPAT macro like before. For
912 API compatibility version below 3.0, the old style numerical
913 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
914 For version 3.0 and on, the value is expected to be the decimal
915 value calculated from the major and minor version like this:
917 MAJOR * 10000 + MINOR * 100
921 -DOPENSSL_API_COMPAT=30000 For 3.0
922 -DOPENSSL_API_COMPAT=30200 For 3.2
924 To hide declarations that are deprecated up to and including the
925 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
926 given when building the application as well.
930 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
931 access to certificate and CRL stores via URIs and OSSL_STORE
934 This adds the following functions:
936 - X509_LOOKUP_store()
937 - X509_STORE_load_file()
938 - X509_STORE_load_path()
939 - X509_STORE_load_store()
940 - SSL_add_store_cert_subjects_to_stack()
941 - SSL_CTX_set_default_verify_store()
942 - SSL_CTX_load_verify_file()
943 - SSL_CTX_load_verify_dir()
944 - SSL_CTX_load_verify_store()
948 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
949 The presence of this system service is determined at run-time.
953 * Added functionality to create an EVP_PKEY context based on data
954 for methods from providers. This takes an algorithm name and a
955 property query string and simply stores them, with the intent
956 that any operation that uses this context will use those strings
957 to fetch the needed methods implicitly, thereby making the port
958 of application written for pre-3.0 OpenSSL easier.
962 * The undocumented function NCONF_WIN32() has been deprecated; for
963 conversion details see the HISTORY section of doc/man5/config.pod
967 * Introduced the new functions EVP_DigestSignInit_ex() and
968 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
969 EVP_DigestVerifyUpdate() have been converted to functions. See the man
970 pages for further details.
974 * Over two thousand fixes were made to the documentation, including:
975 adding missing command flags, better style conformance, documentation
978 *Rich Salz, Richard Levitte*
980 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
981 X25519, X448, Ed25519 and Ed448.
985 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
990 * Deprecated the public definition of `ERR_STATE` as well as the function
991 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
996 * Added ERR functionality to give callers access to the stored function
997 names that have replaced the older function code based functions.
999 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1000 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1001 ERR_peek_error_all() and ERR_peek_last_error_all().
1003 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1004 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1005 ERR_func_error_string().
1009 * Extended testing to be verbose for failing tests only. The make variables
1010 VERBOSE_FAILURE or VF can be used to enable this:
1012 $ make VF=1 test # Unix
1013 $ mms /macro=(VF=1) test ! OpenVMS
1014 $ nmake VF=1 test # Windows
1018 * Added the `-copy_extensions` option to the `x509` command for use with
1019 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1020 all extensions in the request are copied to the certificate or vice versa.
1022 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1024 * Added the `-copy_extensions` option to the `req` command for use with
1025 `-x509`. When given with the `copy` or `copyall` argument,
1026 all extensions in the certification request are copied to the certificate.
1030 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1031 they generate are by default RFC 5280 compliant in the following sense:
1032 There is a subjectKeyIdentifier extension with a hash value of the public key
1033 and for not self-signed certs there is an authorityKeyIdentifier extension
1034 with a keyIdentifier field or issuer information identifying the signing key.
1035 This is done unless some configuration overrides the new default behavior,
1036 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1040 * Added several checks to `X509_verify_cert()` according to requirements in
1041 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1042 (which may be done by using the CLI option `-x509_strict`):
1043 * The basicConstraints of CA certificates must be marked critical.
1044 * CA certificates must explicitly include the keyUsage extension.
1045 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1046 * The issuer name of any certificate must not be empty.
1047 * The subject name of CA certs, certs with keyUsage crlSign,
1048 and certs without subjectAlternativeName must not be empty.
1049 * If a subjectAlternativeName extension is given it must not be empty.
1050 * The signatureAlgorithm field and the cert signature must be consistent.
1051 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1052 must not be marked critical.
1053 * The authorityKeyIdentifier must be given for X.509v3 certs
1054 unless they are self-signed.
1055 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1059 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1060 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1064 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1065 used even when parsing explicit parameters, when loading a encoded key
1066 or calling `EC_GROUP_new_from_ecpkparameters()`/
1067 `EC_GROUP_new_from_ecparameters()`.
1068 This prevents bypass of security hardening and performance gains,
1069 especially for curves with specialized EC_METHODs.
1070 By default, if a key encoded with explicit parameters is loaded and later
1071 encoded, the output is still encoded with explicit parameters, even if
1072 internally a "named" EC_GROUP is used for computation.
1076 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1077 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1078 NULL. After this change, only the cofactor parameter can be NULL. It also
1079 does some minimal sanity checks on the passed order.
1084 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1085 An attack is simple, if the first CMS_recipientInfo is valid but the
1086 second CMS_recipientInfo is chosen ciphertext. If the second
1087 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1088 encryption key will be replaced by garbage, and the message cannot be
1089 decoded, but if the RSA decryption fails, the correct encryption key is
1090 used and the recipient will not notice the attack.
1091 As a work around for this potential attack the length of the decrypted
1092 key must be equal to the cipher default key length, in case the
1093 certifiate is not given and all recipientInfo are tried out.
1094 The old behaviour can be re-enabled in the CMS code by setting the
1095 CMS_DEBUG_DECRYPT flag.
1099 * Early start up entropy quality from the DEVRANDOM seed source has been
1100 improved for older Linux systems. The RAND subsystem will wait for
1101 /dev/random to be producing output before seeding from /dev/urandom.
1102 The seeded state is stored for future library initialisations using
1103 a system global shared memory segment. The shared memory identifier
1104 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1105 the desired value. The default identifier is 114.
1109 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1110 when primes for RSA keys are computed.
1111 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1112 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1113 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1114 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1115 This avoids possible fingerprinting of newly generated RSA modules.
1119 * Correct the extended master secret constant on EBCDIC systems. Without this
1120 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1121 negotiate EMS will fail. Unfortunately this also means that TLS connections
1122 between EBCDIC systems with this fix, and EBCDIC systems without this
1123 fix will fail if they negotiate EMS.
1127 * Changed the library initialisation so that the config file is now loaded
1128 by default. This was already the case for libssl. It now occurs for both
1129 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1130 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1134 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1135 where the former acts as a replacement for `ERR_put_error()`, and the
1136 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1137 `ERR_raise_data()` adds more flexibility by taking a format string and
1138 an arbitrary number of arguments following it, to be processed with
1143 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1144 to check if a named provider is loaded and available. When called, it
1145 will also activate all fallback providers if such are still present.
1149 * Enforce a minimum DH modulus size of 512 bits.
1153 * Changed DH parameters to generate the order q subgroup instead of 2q.
1154 Previously generated DH parameters are still accepted by DH_check
1155 but DH_generate_key works around that by clearing bit 0 of the
1156 private key for those. This avoids leaking bit 0 of the private key.
1160 * Significantly reduce secure memory usage by the randomness pools.
1164 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1169 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1170 algorithms. An implementation of a key exchange algorithm can be obtained
1171 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1172 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1173 the older EVP_PKEY_derive_init() function. See the man pages for the new
1174 functions for further details.
1178 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1182 * Removed the function names from error messages and deprecated the
1187 * Removed NextStep support and the macro OPENSSL_UNISTD
1191 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1192 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1193 Also removed "export var as function" capability; we do not export
1194 variables, only functions.
1198 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1199 an error and 1 indicating success. In previous versions of OpenSSL this
1200 was a void type. If a key was set longer than the maximum possible this
1205 * Support SM2 signing and verification schemes with X509 certificate.
1209 * Use SHA256 as the default digest for TS query in the `ts` app.
1213 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1217 * Default cipher lists/suites are now available via a function, the
1218 #defines are deprecated.
1222 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1223 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1224 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1228 * Join the directories crypto/x509 and crypto/x509v3
1232 * Added command 'openssl kdf' that uses the EVP_KDF API.
1236 * Added command 'openssl mac' that uses the EVP_MAC API.
1240 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1241 as default directories. Also added the command 'openssl info'
1242 for scripting purposes.
1246 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1251 * Add prediction resistance to the DRBG reseeding process.
1255 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1256 mandated by IEEE Std 1619-2018.
1260 * Added newline escaping functionality to a filename when using openssl dgst.
1261 This output format is to replicate the output format found in the `*sum`
1262 checksum programs. This aims to preserve backward compatibility.
1264 *Matt Eaton, Richard Levitte, and Paul Dale*
1266 * Removed the heartbeat message in DTLS feature, as it has very
1267 little usage and doesn't seem to fulfill a valuable purpose.
1268 The configuration option is now deprecated.
1272 * Changed the output of 'openssl {digestname} < file' to display the
1273 digest name in its output.
1277 * Added a new generic trace API which provides support for enabling
1278 instrumentation through trace output.
1280 *Richard Levitte & Matthias St. Pierre*
1282 * Added build tests for C++. These are generated files that only do one
1283 thing, to include one public OpenSSL head file each. This tests that
1284 the public header files can be usefully included in a C++ application.
1286 This test isn't enabled by default. It can be enabled with the option
1287 'enable-buildtest-c++'.
1291 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1295 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1299 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1303 * Added KMAC to EVP_MAC.
1307 * Added property based algorithm implementation selection framework to
1312 * Added SCA hardening for modular field inversion in EC_GROUP through
1313 a new dedicated field_inv() pointer in EC_METHOD.
1314 This also addresses a leakage affecting conversions from projective
1315 to affine coordinates.
1317 *Billy Bob Brumley, Nicola Tuveri*
1319 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1320 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1321 those algorithms that were already supported through the EVP_PKEY API
1322 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1323 and scrypt are now wrappers that call EVP_KDF.
1327 * Build devcrypto engine as a dynamic engine.
1329 *Eneas U de Queiroz*
1331 * Add keyed BLAKE2 to EVP_MAC.
1335 * Fix a bug in the computation of the endpoint-pair shared secret used
1336 by DTLS over SCTP. This breaks interoperability with older versions
1337 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1338 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1339 interoperability with such broken implementations. However, enabling
1340 this switch breaks interoperability with correct implementations.
1342 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1343 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1347 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1351 * Changed the license to the Apache License v2.0.
1355 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1357 - Major releases (indicated by incrementing the MAJOR release number)
1358 may introduce incompatible API/ABI changes.
1359 - Minor releases (indicated by incrementing the MINOR release number)
1360 may introduce new features but retain API/ABI compatibility.
1361 - Patch releases (indicated by incrementing the PATCH number)
1362 are intended for bug fixes and other improvements of existing
1363 features only (like improving performance or adding documentation)
1364 and retain API/ABI compatibility.
1368 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1372 * Remove the 'dist' target and add a tarball building script. The
1373 'dist' target has fallen out of use, and it shouldn't be
1374 necessary to configure just to create a source distribution.
1378 * Recreate the OS390-Unix config target. It no longer relies on a
1379 special script like it did for OpenSSL pre-1.1.0.
1383 * Instead of having the source directories listed in Configure, add
1384 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1389 * Add GMAC to EVP_MAC.
1393 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1397 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1398 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1399 to facilitate the continued use of MACs through raw private keys in
1400 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1404 * Deprecate ECDH_KDF_X9_62().
1408 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1409 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1410 are retained for backwards compatibility.
1414 * AES-XTS mode now enforces that its two keys are different to mitigate
1415 the attacked described in "Efficient Instantiations of Tweakable
1416 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1417 Details of this attack can be obtained from:
1418 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1422 * Rename the object files, i.e. give them other names than in previous
1423 versions. Their names now include the name of the final product, as
1424 well as its type mnemonic (bin, lib, shlib).
1428 * Added new option for 'openssl list', '-objects', which will display the
1429 list of built in objects, i.e. OIDs with names.
1433 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1434 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1439 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1440 improves application performance by removing data copies and providing
1441 applications with zero-copy system calls such as sendfile and splice.
1445 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1449 * `PKCS12_parse` now maintains the order of the parsed certificates
1450 when outputting them via `*ca` (rather than reversing it).
1454 * Deprecated pthread fork support methods.
1458 * Added support for FFDHE key exchange in TLS 1.3.
1462 * Added a new concept for OpenSSL plugability: providers. This
1463 functionality is designed to replace the ENGINE API and ENGINE
1464 implementations, and to be much more dynamic, allowing provider
1465 authors to introduce new algorithms among other things, as long as
1466 there's an API that supports the algorithm type.
1468 With this concept comes a new core API for interaction between
1469 libcrypto and provider implementations. Public libcrypto functions
1470 that want to use providers do so through this core API.
1472 The main documentation for this core API is found in
1473 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1474 refer to other manuals describing the API specific for supported
1475 algorithm types (also called operations).
1482 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1484 * Fixed an SM2 Decryption Buffer Overflow.
1486 In order to decrypt SM2 encrypted data an application is expected to
1487 call the API function EVP_PKEY_decrypt(). Typically an application will
1488 call this function twice. The first time, on entry, the "out" parameter
1489 can be NULL and, on exit, the "outlen" parameter is populated with the
1490 buffer size required to hold the decrypted plaintext. The application
1491 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1492 again, but this time passing a non-NULL value for the "out" parameter.
1494 A bug in the implementation of the SM2 decryption code means that the
1495 calculation of the buffer size required to hold the plaintext returned
1496 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1497 size required by the second call. This can lead to a buffer overflow
1498 when EVP_PKEY_decrypt() is called by the application a second time with
1499 a buffer that is too small.
1501 A malicious attacker who is able present SM2 content for decryption to
1502 an application could cause attacker chosen data to overflow the buffer
1503 by up to a maximum of 62 bytes altering the contents of other data held
1504 after the buffer, possibly changing application behaviour or causing
1505 the application to crash. The location of the buffer is application
1506 dependent but is typically heap allocated.
1511 * Fixed various read buffer overruns processing ASN.1 strings
1513 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1514 structure which contains a buffer holding the string data and a field
1515 holding the buffer length. This contrasts with normal C strings which
1516 are repesented as a buffer for the string data which is terminated
1517 with a NUL (0) byte.
1519 Although not a strict requirement, ASN.1 strings that are parsed using
1520 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1521 well as any string whose value has been set with the ASN1_STRING_set()
1522 function will additionally NUL terminate the byte array in the
1523 ASN1_STRING structure.
1525 However, it is possible for applications to directly construct valid
1526 ASN1_STRING structures which do not NUL terminate the byte array by
1527 directly setting the "data" and "length" fields in the ASN1_STRING
1528 array. This can also happen by using the ASN1_STRING_set0() function.
1530 Numerous OpenSSL functions that print ASN.1 data have been found to
1531 assume that the ASN1_STRING byte array will be NUL terminated, even
1532 though this is not guaranteed for strings that have been directly
1533 constructed. Where an application requests an ASN.1 structure to be
1534 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1535 been directly constructed by the application without NUL terminating
1536 the "data" field, then a read buffer overrun can occur.
1538 The same thing can also occur during name constraints processing
1539 of certificates (for example if a certificate has been directly
1540 constructed by the application instead of loading it via the OpenSSL
1541 parsing functions, and the certificate contains non NUL terminated
1542 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1543 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1545 If a malicious actor can cause an application to directly construct an
1546 ASN1_STRING and then process it through one of the affected OpenSSL
1547 functions then this issue could be hit. This might result in a crash
1548 (causing a Denial of Service attack). It could also result in the
1549 disclosure of private memory contents (such as private keys, or
1550 sensitive plaintext).
1555 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1557 * Fixed a problem with verifying a certificate chain when using the
1558 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1559 the certificates present in a certificate chain. It is not set by default.
1561 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1562 the chain that have explicitly encoded elliptic curve parameters was added
1563 as an additional strict check.
1565 An error in the implementation of this check meant that the result of a
1566 previous check to confirm that certificates in the chain are valid CA
1567 certificates was overwritten. This effectively bypasses the check
1568 that non-CA certificates must not be able to issue other certificates.
1570 If a "purpose" has been configured then there is a subsequent opportunity
1571 for checks that the certificate is a valid CA. All of the named "purpose"
1572 values implemented in libcrypto perform this check. Therefore, where
1573 a purpose is set the certificate chain will still be rejected even when the
1574 strict flag has been used. A purpose is set by default in libssl client and
1575 server certificate verification routines, but it can be overridden or
1576 removed by an application.
1578 In order to be affected, an application must explicitly set the
1579 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1580 for the certificate verification or, in the case of TLS client or server
1581 applications, override the default purpose.
1586 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1587 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1588 renegotiation ClientHello omits the signature_algorithms extension (where it
1589 was present in the initial ClientHello), but includes a
1590 signature_algorithms_cert extension then a NULL pointer dereference will
1591 result, leading to a crash and a denial of service attack.
1593 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1594 (which is the default configuration). OpenSSL TLS clients are not impacted by
1598 *Peter Kästle and Samuel Sapalski*
1600 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1602 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1603 create a unique hash value based on the issuer and serial number data
1604 contained within an X509 certificate. However it was failing to correctly
1605 handle any errors that may occur while parsing the issuer field (which might
1606 occur if the issuer field is maliciously constructed). This may subsequently
1607 result in a NULL pointer deref and a crash leading to a potential denial of
1613 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1614 padding mode to correctly check for rollback attacks. This is considered a
1615 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1620 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1621 functions. Previously they could overflow the output length argument in some
1622 cases where the input length is close to the maximum permissable length for
1623 an integer on the platform. In such cases the return value from the function
1624 call would be 1 (indicating success), but the output length value would be
1625 negative. This could cause applications to behave incorrectly or crash.
1630 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1631 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1632 could be exploited in a side channel attack to recover the password. Since
1633 the attack is local host only this is outside of the current OpenSSL
1634 threat model and therefore no CVE is assigned.
1636 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1641 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1643 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1644 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1645 If an attacker can control both items being compared then this could lead
1646 to a possible denial of service attack. OpenSSL itself uses the
1647 GENERAL_NAME_cmp function for two purposes:
1648 1) Comparing CRL distribution point names between an available CRL and a
1649 CRL distribution point embedded in an X509 certificate
1650 2) When verifying that a timestamp response token signer matches the
1651 timestamp authority name (exposed via the API functions
1652 TS_RESP_verify_response and TS_RESP_verify_token)
1657 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1659 * Certificates with explicit curve parameters are now disallowed in
1660 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1664 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1665 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1666 conversely, silently ignore DTLS protocol version bounds when configuring
1667 TLS-based contexts. The commands can be repeated to set bounds of both
1668 types. The same applies with the corresponding "min_protocol" and
1669 "max_protocol" command-line switches, in case some application uses both TLS
1672 SSL_CTX instances that are created for a fixed protocol version (e.g.
1673 TLSv1_server_method()) also silently ignore version bounds. Previously
1674 attempts to apply bounds to these protocol versions would result in an
1675 error. Now only the "version-flexible" SSL_CTX instances are subject to
1676 limits in configuration files in command-line options.
1680 * Handshake now fails if Extended Master Secret extension is dropped
1685 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1687 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1689 * Fixed segmentation fault in SSL_check_chain()
1690 Server or client applications that call the SSL_check_chain() function
1691 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1692 dereference as a result of incorrect handling of the
1693 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1694 or unrecognised signature algorithm is received from the peer. This could
1695 be exploited by a malicious peer in a Denial of Service attack.
1700 * Added AES consttime code for no-asm configurations
1701 an optional constant time support for AES was added
1702 when building openssl for no-asm.
1703 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1704 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1705 At this time this feature is by default disabled.
1706 It will be enabled by default in 3.0.
1710 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1712 * Revert the change of EOF detection while reading in libssl to avoid
1713 regressions in applications depending on the current way of reporting
1714 the EOF. As the existing method is not fully accurate the change to
1715 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1716 branch and will be present in the 3.0 release.
1720 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1721 when primes for RSA keys are computed.
1722 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1723 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1724 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1725 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1726 This avoids possible fingerprinting of newly generated RSA modules.
1730 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1732 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1733 while reading in libssl then we would report an error back to the
1734 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1735 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1736 therefore give a hint as to what went wrong.
1740 * Check that ed25519 and ed448 are allowed by the security level. Previously
1741 signature algorithms not using an MD were not being checked that they were
1742 allowed by the security level.
1746 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1747 was not quite right. The behaviour was not consistent between resumption
1748 and normal handshakes, and also not quite consistent with historical
1749 behaviour. The behaviour in various scenarios has been clarified and
1750 it has been updated to make it match historical behaviour as closely as
1755 * *[VMS only]* The header files that the VMS compilers include automatically,
1756 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1757 that the C++ compiler doesn't understand. This is a shortcoming in the
1758 compiler, but can be worked around with `__cplusplus` guards.
1760 C++ applications that use OpenSSL libraries must be compiled using the
1761 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1762 functions. Otherwise, only functions with symbols of less than 31
1763 characters can be used, as the linker will not be able to successfully
1764 resolve symbols with longer names.
1768 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1769 The presence of this system service is determined at run-time.
1773 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1778 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1780 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1781 number generator (RNG). This was intended to include protection in the
1782 event of a fork() system call in order to ensure that the parent and child
1783 processes did not share the same RNG state. However this protection was not
1784 being used in the default case.
1786 A partial mitigation for this issue is that the output from a high
1787 precision timer is mixed into the RNG state so the likelihood of a parent
1788 and child process sharing state is significantly reduced.
1790 If an application already calls OPENSSL_init_crypto() explicitly using
1791 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1794 *Matthias St. Pierre*
1796 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1797 used even when parsing explicit parameters, when loading a encoded key
1798 or calling `EC_GROUP_new_from_ecpkparameters()`/
1799 `EC_GROUP_new_from_ecparameters()`.
1800 This prevents bypass of security hardening and performance gains,
1801 especially for curves with specialized EC_METHODs.
1802 By default, if a key encoded with explicit parameters is loaded and later
1803 encoded, the output is still encoded with explicit parameters, even if
1804 internally a "named" EC_GROUP is used for computation.
1808 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1809 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1810 NULL. After this change, only the cofactor parameter can be NULL. It also
1811 does some minimal sanity checks on the passed order.
1816 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1817 An attack is simple, if the first CMS_recipientInfo is valid but the
1818 second CMS_recipientInfo is chosen ciphertext. If the second
1819 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1820 encryption key will be replaced by garbage, and the message cannot be
1821 decoded, but if the RSA decryption fails, the correct encryption key is
1822 used and the recipient will not notice the attack.
1823 As a work around for this potential attack the length of the decrypted
1824 key must be equal to the cipher default key length, in case the
1825 certifiate is not given and all recipientInfo are tried out.
1826 The old behaviour can be re-enabled in the CMS code by setting the
1827 CMS_DEBUG_DECRYPT flag.
1832 * Early start up entropy quality from the DEVRANDOM seed source has been
1833 improved for older Linux systems. The RAND subsystem will wait for
1834 /dev/random to be producing output before seeding from /dev/urandom.
1835 The seeded state is stored for future library initialisations using
1836 a system global shared memory segment. The shared memory identifier
1837 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1838 the desired value. The default identifier is 114.
1842 * Correct the extended master secret constant on EBCDIC systems. Without this
1843 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1844 negotiate EMS will fail. Unfortunately this also means that TLS connections
1845 between EBCDIC systems with this fix, and EBCDIC systems without this
1846 fix will fail if they negotiate EMS.
1850 * Use Windows installation paths in the mingw builds
1852 Mingw isn't a POSIX environment per se, which means that Windows
1853 paths should be used for installation.
1858 * Changed DH_check to accept parameters with order q and 2q subgroups.
1859 With order 2q subgroups the bit 0 of the private key is not secret
1860 but DH_generate_key works around that by clearing bit 0 of the
1861 private key for those. This avoids leaking bit 0 of the private key.
1865 * Significantly reduce secure memory usage by the randomness pools.
1869 * Revert the DEVRANDOM_WAIT feature for Linux systems
1871 The DEVRANDOM_WAIT feature added a select() call to wait for the
1872 /dev/random device to become readable before reading from the
1873 /dev/urandom device.
1875 It turned out that this change had negative side effects on
1876 performance which were not acceptable. After some discussion it
1877 was decided to revert this feature and leave it up to the OS
1878 resp. the platform maintainer to ensure a proper initialization
1879 during early boot time.
1881 *Matthias St. Pierre*
1883 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1885 * Add build tests for C++. These are generated files that only do one
1886 thing, to include one public OpenSSL head file each. This tests that
1887 the public header files can be usefully included in a C++ application.
1889 This test isn't enabled by default. It can be enabled with the option
1890 'enable-buildtest-c++'.
1894 * Enable SHA3 pre-hashing for ECDSA and DSA.
1898 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1899 This changes the size when using the `genpkey` command when no size is given.
1900 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1901 generation commands to use 2048 bits by default.
1905 * Reorganize the manual pages to consistently have RETURN VALUES,
1906 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1907 util/fix-doc-nits accordingly.
1909 *Paul Yang, Joshua Lock*
1911 * Add the missing accessor EVP_PKEY_get0_engine()
1915 * Have commands like `s_client` and `s_server` output the signature scheme
1916 along with other cipher suite parameters when debugging.
1920 * Make OPENSSL_config() error agnostic again.
1924 * Do the error handling in RSA decryption constant time.
1928 * Prevent over long nonces in ChaCha20-Poly1305.
1930 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1931 for every encryption operation. RFC 7539 specifies that the nonce value
1932 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1933 and front pads the nonce with 0 bytes if it is less than 12
1934 bytes. However it also incorrectly allows a nonce to be set of up to 16
1935 bytes. In this case only the last 12 bytes are significant and any
1936 additional leading bytes are ignored.
1938 It is a requirement of using this cipher that nonce values are
1939 unique. Messages encrypted using a reused nonce value are susceptible to
1940 serious confidentiality and integrity attacks. If an application changes
1941 the default nonce length to be longer than 12 bytes and then makes a
1942 change to the leading bytes of the nonce expecting the new value to be a
1943 new unique nonce then such an application could inadvertently encrypt
1944 messages with a reused nonce.
1946 Additionally the ignored bytes in a long nonce are not covered by the
1947 integrity guarantee of this cipher. Any application that relies on the
1948 integrity of these ignored leading bytes of a long nonce may be further
1949 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1950 is safe because no such use sets such a long nonce value. However user
1951 applications that use this cipher directly and set a non-default nonce
1952 length to be longer than 12 bytes may be vulnerable.
1954 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1960 * Add DEVRANDOM_WAIT feature for Linux systems
1962 On older Linux systems where the getrandom() system call is not available,
1963 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1964 Contrary to getrandom(), the /dev/urandom device will not block during
1965 early boot when the kernel CSPRNG has not been seeded yet.
1967 To mitigate this known weakness, use select() to wait for /dev/random to
1968 become readable before reading from /dev/urandom.
1970 * Ensure that SM2 only uses SM3 as digest algorithm
1974 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1976 * Change the info callback signals for the start and end of a post-handshake
1977 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1978 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1979 confused by this and assume that a TLSv1.2 renegotiation has started. This
1980 can break KeyUpdate handling. Instead we no longer signal the start and end
1981 of a post handshake message exchange (although the messages themselves are
1982 still signalled). This could break some applications that were expecting
1983 the old signals. However without this KeyUpdate is not usable for many
1988 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1990 * Timing vulnerability in DSA signature generation
1992 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1993 timing side channel attack. An attacker could use variations in the signing
1994 algorithm to recover the private key.
1996 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2001 * Timing vulnerability in ECDSA signature generation
2003 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2004 timing side channel attack. An attacker could use variations in the signing
2005 algorithm to recover the private key.
2007 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2012 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2013 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2014 of two gigabytes and the error handling improved.
2016 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2017 categorized as a normal bug, not a security issue, because the DRBG reseeds
2018 automatically and is fully functional even without additional randomness
2019 provided by the application.
2021 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2023 * Add a new ClientHello callback. Provides a callback interface that gives
2024 the application the ability to adjust the nascent SSL object at the
2025 earliest stage of ClientHello processing, immediately after extensions have
2026 been collected but before they have been processed. In particular, this
2027 callback can adjust the supported TLS versions in response to the contents
2032 * Add SM2 base algorithm support.
2036 * s390x assembly pack: add (improved) hardware-support for the following
2037 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2038 aes-cfb/cfb8, aes-ecb.
2042 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2043 parameter is no longer accepted, as it leads to a corrupt table. NULL
2044 pem_str is reserved for alias entries only.
2048 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2049 step for prime curves. The new implementation is based on formulae from
2050 differential addition-and-doubling in homogeneous projective coordinates
2051 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2052 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2053 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2054 to work in projective coordinates.
2056 *Billy Bob Brumley, Nicola Tuveri*
2058 * Change generating and checking of primes so that the error rate of not
2059 being prime depends on the intended use based on the size of the input.
2060 For larger primes this will result in more rounds of Miller-Rabin.
2061 The maximal error rate for primes with more than 1080 bits is lowered
2064 *Kurt Roeckx, Annie Yousar*
2066 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2070 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2071 moving between systems, and to avoid confusion when a Windows build is
2072 done with mingw vs with MSVC. For POSIX installs, there's still a
2073 symlink or copy named 'tsget' to avoid that confusion as well.
2077 * Revert blinding in ECDSA sign and instead make problematic addition
2078 length-invariant. Switch even to fixed-length Montgomery multiplication.
2082 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2083 step for binary curves. The new implementation is based on formulae from
2084 differential addition-and-doubling in mixed Lopez-Dahab projective
2085 coordinates, modified to independently blind the operands.
2087 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2089 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2090 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2091 EC_METHODs to implement their own specialized "ladder step", to take
2092 advantage of more favorable coordinate systems or more efficient
2093 differential addition-and-doubling algorithms.
2095 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2097 * Modified the random device based seed sources to keep the relevant
2098 file descriptors open rather than reopening them on each access.
2099 This allows such sources to operate in a chroot() jail without
2100 the associated device nodes being available. This behaviour can be
2101 controlled using RAND_keep_random_devices_open().
2105 * Numerous side-channel attack mitigations have been applied. This may have
2106 performance impacts for some algorithms for the benefit of improved
2107 security. Specific changes are noted in this change log by their respective
2112 * AIX shared library support overhaul. Switch to AIX "natural" way of
2113 handling shared libraries, which means collecting shared objects of
2114 different versions and bitnesses in one common archive. This allows to
2115 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2116 doesn't affect the way 3rd party applications are linked, only how
2117 multi-version installation is managed.
2121 * Make ec_group_do_inverse_ord() more robust and available to other
2122 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2123 mitigations are applied to the fallback BN_mod_inverse().
2124 When using this function rather than BN_mod_inverse() directly, new
2125 EC cryptosystem implementations are then safer-by-default.
2129 * Add coordinate blinding for EC_POINT and implement projective
2130 coordinate blinding for generic prime curves as a countermeasure to
2131 chosen point SCA attacks.
2133 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2135 * Add blinding to ECDSA and DSA signatures to protect against side channel
2136 attacks discovered by Keegan Ryan (NCC Group).
2140 * Enforce checking in the `pkeyutl` command to ensure that the input
2141 length does not exceed the maximum supported digest length when performing
2142 a sign, verify or verifyrecover operation.
2146 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2147 I/O in combination with something like select() or poll() will hang. This
2148 can be turned off again using SSL_CTX_clear_mode().
2149 Many applications do not properly handle non-application data records, and
2150 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2151 around the problems in those applications, but can also break some.
2152 It's recommended to read the manpages about SSL_read(), SSL_write(),
2153 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2154 SSL_CTX_set_read_ahead() again.
2158 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2159 now allow empty (zero character) pass phrases.
2163 * Apply blinding to binary field modular inversion and remove patent
2164 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2168 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2169 binary and prime elliptic curves.
2173 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2174 constant time fixed point multiplication.
2178 * Revise elliptic curve scalar multiplication with timing attack
2179 defenses: ec_wNAF_mul redirects to a constant time implementation
2180 when computing fixed point and variable point multiplication (which
2181 in OpenSSL are mostly used with secret scalars in keygen, sign,
2182 ECDH derive operations).
2183 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2186 * Updated CONTRIBUTING
2190 * Updated DRBG / RAND to request nonce and additional low entropy
2191 randomness from the system.
2193 *Matthias St. Pierre*
2195 * Updated 'openssl rehash' to use OpenSSL consistent default.
2199 * Moved the load of the ssl_conf module to libcrypto, which helps
2200 loading engines that libssl uses before libssl is initialised.
2204 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2208 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2210 *Ingo Schwarze, Rich Salz*
2212 * Added output of accepting IP address and port for 'openssl s_server'
2216 * Added a new API for TLSv1.3 ciphersuites:
2217 SSL_CTX_set_ciphersuites()
2218 SSL_set_ciphersuites()
2222 * Memory allocation failures consistently add an error to the error
2227 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2228 in libcrypto when run as setuid/setgid.
2232 * Load any config file by default when libssl is used.
2236 * Added new public header file <openssl/rand_drbg.h> and documentation
2237 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2239 *Matthias St. Pierre*
2241 * QNX support removed (cannot find contributors to get their approval
2242 for the license change).
2246 * TLSv1.3 replay protection for early data has been implemented. See the
2247 SSL_read_early_data() man page for further details.
2251 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2252 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2253 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2254 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2255 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2256 configuration has been separated out. See the ciphers man page or the
2257 SSL_CTX_set_ciphersuites() man page for more information.
2261 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2262 in responder mode now supports the new "-multi" option, which
2263 spawns the specified number of child processes to handle OCSP
2264 requests. The "-timeout" option now also limits the OCSP
2265 responder's patience to wait to receive the full client request
2266 on a newly accepted connection. Child processes are respawned
2267 as needed, and the CA index file is automatically reloaded
2268 when changed. This makes it possible to run the "ocsp" responder
2269 as a long-running service, making the OpenSSL CA somewhat more
2270 feature-complete. In this mode, most diagnostic messages logged
2271 after entering the event loop are logged via syslog(3) rather than
2276 * Added support for X448 and Ed448. Heavily based on original work by
2281 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2282 objects loaded. This adds the functions OSSL_STORE_expect() and
2283 OSSL_STORE_find() as well as needed tools to construct searches and
2284 get the search data out of them.
2288 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2289 version of OpenSSL should review their configuration settings to ensure
2290 that they are still appropriate for TLSv1.3. For further information see:
2291 <https://wiki.openssl.org/index.php/TLS1.3>
2295 * Grand redesign of the OpenSSL random generator
2297 The default RAND method now utilizes an AES-CTR DRBG according to
2298 NIST standard SP 800-90Ar1. The new random generator is essentially
2299 a port of the default random generator from the OpenSSL FIPS 2.0
2300 object module. It is a hybrid deterministic random bit generator
2301 using an AES-CTR bit stream and which seeds and reseeds itself
2302 automatically using trusted system entropy sources.
2304 Some of its new features are:
2305 - Support for multiple DRBG instances with seed chaining.
2306 - The default RAND method makes use of a DRBG.
2307 - There is a public and private DRBG instance.
2308 - The DRBG instances are fork-safe.
2309 - Keep all global DRBG instances on the secure heap if it is enabled.
2310 - The public and private DRBG instance are per thread for lock free
2313 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2315 * Changed Configure so it only says what it does and doesn't dump
2316 so much data. Instead, ./configdata.pm should be used as a script
2317 to display all sorts of configuration data.
2321 * Added processing of "make variables" to Configure.
2325 * Added SHA512/224 and SHA512/256 algorithm support.
2329 * The last traces of Netware support, first removed in 1.1.0, have
2334 * Get rid of Makefile.shared, and in the process, make the processing
2335 of certain files (rc.obj, or the .def/.map/.opt files produced from
2336 the ordinal files) more visible and hopefully easier to trace and
2337 debug (or make silent).
2341 * Make it possible to have environment variable assignments as
2342 arguments to config / Configure.
2346 * Add multi-prime RSA (RFC 8017) support.
2350 * Add SM3 implemented according to GB/T 32905-2016
2351 *Jack Lloyd <jack.lloyd@ribose.com>,*
2352 *Ronald Tse <ronald.tse@ribose.com>,*
2353 *Erick Borsboom <erick.borsboom@ribose.com>*
2355 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2356 as documented in RFC6066.
2357 Based on a patch from Tomasz Moń
2359 *Filipe Raimundo da Silva*
2361 * Add SM4 implemented according to GB/T 32907-2016.
2362 *Jack Lloyd <jack.lloyd@ribose.com>,*
2363 *Ronald Tse <ronald.tse@ribose.com>,*
2364 *Erick Borsboom <erick.borsboom@ribose.com>*
2366 * Reimplement -newreq-nodes and ERR_error_string_n; the
2367 original author does not agree with the license change.
2371 * Add ARIA AEAD TLS support.
2375 * Some macro definitions to support VS6 have been removed. Visual
2376 Studio 6 has not worked since 1.1.0
2380 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2381 without clearing the errors.
2385 * Add "atfork" functions. If building on a system that without
2386 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2387 requirements. The RAND facility now uses/requires this.
2395 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2396 not possible to disable entirely. However, it's still possible to
2397 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2400 To disable, configure with 'no-ui-console'. 'no-ui' is still
2401 possible to use as an alias. Check at compile time with the
2402 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2403 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2407 * Add a STORE module, which implements a uniform and URI based reader of
2408 stores that can contain keys, certificates, CRLs and numerous other
2409 objects. The main API is loosely based on a few stdio functions,
2410 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2411 OSSL_STORE_error and OSSL_STORE_close.
2412 The implementation uses backends called "loaders" to implement arbitrary
2413 URI schemes. There is one built in "loader" for the 'file' scheme.
2417 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2418 then adjusted to work on FreeBSD 8.4 as well.
2419 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2420 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2424 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2425 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2426 error code calls like this:
2428 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2430 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2431 that can be encoded in C. For the foreseeable future, this will only
2434 *Richard Levitte and Tim Hudson*
2436 * Removed BSD cryptodev engine.
2440 * Add a build target 'build_all_generated', to build all generated files
2441 and only that. This can be used to prepare everything that requires
2442 things like perl for a system that lacks perl and then move everything
2443 to that system and do the rest of the build there.
2447 * In the UI interface, make it possible to duplicate the user data. This
2448 can be used by engines that need to retain the data for a longer time
2449 than just the call where this user data is passed.
2453 * Ignore the '-named_curve auto' value for compatibility of applications
2456 *Tomáš Mráz <tmraz@fedoraproject.org>*
2458 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2459 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2460 alerts across multiple records (some of which could be empty). In practice
2461 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2462 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2463 support this at all. Supporting it adds significant complexity to the
2464 record layer, and its removal is unlikely to cause interoperability
2469 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2470 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2471 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2476 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2477 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2479 *Richard Levitte, Andy Polyakov*
2481 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2486 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2487 platform rather than 'mingw'.
2491 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2492 success if they are asked to add an object which already exists
2493 in the store. This change cascades to other functions which load
2494 certificates and CRLs.
2498 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2499 facilitate stack unwinding even from assembly subroutines.
2503 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2504 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2508 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2509 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2510 which is the minimum version we support.
2514 * Certificate time validation (X509_cmp_time) enforces stricter
2515 compliance with RFC 5280. Fractional seconds and timezone offsets
2516 are no longer allowed.
2520 * Add support for ARIA
2524 * s_client will now send the Server Name Indication (SNI) extension by
2525 default unless the new "-noservername" option is used. The server name is
2526 based on the host provided to the "-connect" option unless overridden by
2527 using "-servername".
2531 * Add support for SipHash
2535 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2536 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2537 prevent issues where no progress is being made and the peer continually
2538 sends unrecognised record types, using up resources processing them.
2542 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2543 using the algorithm defined in
2544 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2548 * Heartbeat support has been removed; the ABI is changed for now.
2550 *Richard Levitte, Rich Salz*
2552 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2556 * The RSA "null" method, which was partially supported to avoid patent
2557 issues, has been replaced to always returns NULL.
2564 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2566 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2567 used even when parsing explicit parameters, when loading a encoded key
2568 or calling `EC_GROUP_new_from_ecpkparameters()`/
2569 `EC_GROUP_new_from_ecparameters()`.
2570 This prevents bypass of security hardening and performance gains,
2571 especially for curves with specialized EC_METHODs.
2572 By default, if a key encoded with explicit parameters is loaded and later
2573 encoded, the output is still encoded with explicit parameters, even if
2574 internally a "named" EC_GROUP is used for computation.
2578 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2579 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2580 NULL. After this change, only the cofactor parameter can be NULL. It also
2581 does some minimal sanity checks on the passed order.
2586 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2587 An attack is simple, if the first CMS_recipientInfo is valid but the
2588 second CMS_recipientInfo is chosen ciphertext. If the second
2589 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2590 encryption key will be replaced by garbage, and the message cannot be
2591 decoded, but if the RSA decryption fails, the correct encryption key is
2592 used and the recipient will not notice the attack.
2593 As a work around for this potential attack the length of the decrypted
2594 key must be equal to the cipher default key length, in case the
2595 certifiate is not given and all recipientInfo are tried out.
2596 The old behaviour can be re-enabled in the CMS code by setting the
2597 CMS_DEBUG_DECRYPT flag.
2602 * Use Windows installation paths in the mingw builds
2604 Mingw isn't a POSIX environment per se, which means that Windows
2605 paths should be used for installation.
2610 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2612 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2613 This changes the size when using the `genpkey` command when no size is given.
2614 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2615 generation commands to use 2048 bits by default.
2619 * Prevent over long nonces in ChaCha20-Poly1305.
2621 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2622 for every encryption operation. RFC 7539 specifies that the nonce value
2623 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2624 and front pads the nonce with 0 bytes if it is less than 12
2625 bytes. However it also incorrectly allows a nonce to be set of up to 16
2626 bytes. In this case only the last 12 bytes are significant and any
2627 additional leading bytes are ignored.
2629 It is a requirement of using this cipher that nonce values are
2630 unique. Messages encrypted using a reused nonce value are susceptible to
2631 serious confidentiality and integrity attacks. If an application changes
2632 the default nonce length to be longer than 12 bytes and then makes a
2633 change to the leading bytes of the nonce expecting the new value to be a
2634 new unique nonce then such an application could inadvertently encrypt
2635 messages with a reused nonce.
2637 Additionally the ignored bytes in a long nonce are not covered by the
2638 integrity guarantee of this cipher. Any application that relies on the
2639 integrity of these ignored leading bytes of a long nonce may be further
2640 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2641 is safe because no such use sets such a long nonce value. However user
2642 applications that use this cipher directly and set a non-default nonce
2643 length to be longer than 12 bytes may be vulnerable.
2645 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2651 * Added SCA hardening for modular field inversion in EC_GROUP through
2652 a new dedicated field_inv() pointer in EC_METHOD.
2653 This also addresses a leakage affecting conversions from projective
2654 to affine coordinates.
2656 *Billy Bob Brumley, Nicola Tuveri*
2658 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2659 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2663 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2667 * Remove the 'dist' target and add a tarball building script. The
2668 'dist' target has fallen out of use, and it shouldn't be
2669 necessary to configure just to create a source distribution.
2673 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2675 * Timing vulnerability in DSA signature generation
2677 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2678 timing side channel attack. An attacker could use variations in the signing
2679 algorithm to recover the private key.
2681 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2686 * Timing vulnerability in ECDSA signature generation
2688 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2689 timing side channel attack. An attacker could use variations in the signing
2690 algorithm to recover the private key.
2692 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2697 * Add coordinate blinding for EC_POINT and implement projective
2698 coordinate blinding for generic prime curves as a countermeasure to
2699 chosen point SCA attacks.
2701 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2703 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2705 * Client DoS due to large DH parameter
2707 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2708 malicious server can send a very large prime value to the client. This will
2709 cause the client to spend an unreasonably long period of time generating a
2710 key for this prime resulting in a hang until the client has finished. This
2711 could be exploited in a Denial Of Service attack.
2713 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2718 * Cache timing vulnerability in RSA Key Generation
2720 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2721 a cache timing side channel attack. An attacker with sufficient access to
2722 mount cache timing attacks during the RSA key generation process could
2723 recover the private key.
2725 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2726 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2731 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2732 parameter is no longer accepted, as it leads to a corrupt table. NULL
2733 pem_str is reserved for alias entries only.
2737 * Revert blinding in ECDSA sign and instead make problematic addition
2738 length-invariant. Switch even to fixed-length Montgomery multiplication.
2742 * Change generating and checking of primes so that the error rate of not
2743 being prime depends on the intended use based on the size of the input.
2744 For larger primes this will result in more rounds of Miller-Rabin.
2745 The maximal error rate for primes with more than 1080 bits is lowered
2748 *Kurt Roeckx, Annie Yousar*
2750 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2754 * Add blinding to ECDSA and DSA signatures to protect against side channel
2755 attacks discovered by Keegan Ryan (NCC Group).
2759 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2760 now allow empty (zero character) pass phrases.
2764 * Certificate time validation (X509_cmp_time) enforces stricter
2765 compliance with RFC 5280. Fractional seconds and timezone offsets
2766 are no longer allowed.
2770 * Fixed a text canonicalisation bug in CMS
2772 Where a CMS detached signature is used with text content the text goes
2773 through a canonicalisation process first prior to signing or verifying a
2774 signature. This process strips trailing space at the end of lines, converts
2775 line terminators to CRLF and removes additional trailing line terminators
2776 at the end of a file. A bug in the canonicalisation process meant that
2777 some characters, such as form-feed, were incorrectly treated as whitespace
2778 and removed. This is contrary to the specification (RFC5485). This fix
2779 could mean that detached text data signed with an earlier version of
2780 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2781 signed with a fixed OpenSSL may fail to verify with an earlier version of
2782 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2783 and use the "-binary" flag (for the "cms" command line application) or set
2784 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2788 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2790 * Constructed ASN.1 types with a recursive definition could exceed the stack
2792 Constructed ASN.1 types with a recursive definition (such as can be found
2793 in PKCS7) could eventually exceed the stack given malicious input with
2794 excessive recursion. This could result in a Denial Of Service attack. There
2795 are no such structures used within SSL/TLS that come from untrusted sources
2796 so this is considered safe.
2798 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2804 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2806 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2807 effectively reduced to only comparing the least significant bit of each
2808 byte. This allows an attacker to forge messages that would be considered as
2809 authenticated in an amount of tries lower than that guaranteed by the
2810 security claims of the scheme. The module can only be compiled by the
2811 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2813 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2819 * Add a build target 'build_all_generated', to build all generated files
2820 and only that. This can be used to prepare everything that requires
2821 things like perl for a system that lacks perl and then move everything
2822 to that system and do the rest of the build there.
2826 * Backport SSL_OP_NO_RENGOTIATION
2828 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2829 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2830 changes this is no longer possible in 1.1.0. Therefore the new
2831 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2832 1.1.0 to provide equivalent functionality.
2834 Note that if an application built against 1.1.0h headers (or above) is run
2835 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2836 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2840 * Removed the OS390-Unix config target. It relied on a script that doesn't
2845 * rsaz_1024_mul_avx2 overflow bug on x86_64
2847 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2848 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2849 Analysis suggests that attacks against RSA and DSA as a result of this
2850 defect would be very difficult to perform and are not believed likely.
2851 Attacks against DH1024 are considered just feasible, because most of the
2852 work necessary to deduce information about a private key may be performed
2853 offline. The amount of resources required for such an attack would be
2854 significant. However, for an attack on TLS to be meaningful, the server
2855 would have to share the DH1024 private key among multiple clients, which is
2856 no longer an option since CVE-2016-0701.
2858 This only affects processors that support the AVX2 but not ADX extensions
2859 like Intel Haswell (4th generation).
2861 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2862 was originally found via the OSS-Fuzz project.
2867 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2869 * bn_sqrx8x_internal carry bug on x86_64
2871 There is a carry propagating bug in the x86_64 Montgomery squaring
2872 procedure. No EC algorithms are affected. Analysis suggests that attacks
2873 against RSA and DSA as a result of this defect would be very difficult to
2874 perform and are not believed likely. Attacks against DH are considered just
2875 feasible (although very difficult) because most of the work necessary to
2876 deduce information about a private key may be performed offline. The amount
2877 of resources required for such an attack would be very significant and
2878 likely only accessible to a limited number of attackers. An attacker would
2879 additionally need online access to an unpatched system using the target
2880 private key in a scenario with persistent DH parameters and a private
2881 key that is shared between multiple clients.
2883 This only affects processors that support the BMI1, BMI2 and ADX extensions
2884 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2886 This issue was reported to OpenSSL by the OSS-Fuzz project.
2891 * Malformed X.509 IPAddressFamily could cause OOB read
2893 If an X.509 certificate has a malformed IPAddressFamily extension,
2894 OpenSSL could do a one-byte buffer overread. The most likely result
2895 would be an erroneous display of the certificate in text format.
2897 This issue was reported to OpenSSL by the OSS-Fuzz project.
2902 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2904 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2905 platform rather than 'mingw'.
2909 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2910 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2911 which is the minimum version we support.
2915 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2917 * Encrypt-Then-Mac renegotiation crash
2919 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2920 negotiated where it was not in the original handshake (or vice-versa) then
2921 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2922 and servers are affected.
2924 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2929 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2931 * Truncated packet could crash via OOB read
2933 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2934 cipher is being used, then a truncated packet can cause that host to
2935 perform an out-of-bounds read, usually resulting in a crash.
2937 This issue was reported to OpenSSL by Robert Święcki of Google.
2942 * Bad (EC)DHE parameters cause a client crash
2944 If a malicious server supplies bad parameters for a DHE or ECDHE key
2945 exchange then this can result in the client attempting to dereference a
2946 NULL pointer leading to a client crash. This could be exploited in a Denial
2949 This issue was reported to OpenSSL by Guido Vranken.
2954 * BN_mod_exp may produce incorrect results on x86_64
2956 There is a carry propagating bug in the x86_64 Montgomery squaring
2957 procedure. No EC algorithms are affected. Analysis suggests that attacks
2958 against RSA and DSA as a result of this defect would be very difficult to
2959 perform and are not believed likely. Attacks against DH are considered just
2960 feasible (although very difficult) because most of the work necessary to
2961 deduce information about a private key may be performed offline. The amount
2962 of resources required for such an attack would be very significant and
2963 likely only accessible to a limited number of attackers. An attacker would
2964 additionally need online access to an unpatched system using the target
2965 private key in a scenario with persistent DH parameters and a private
2966 key that is shared between multiple clients. For example this can occur by
2967 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2968 similar to CVE-2015-3193 but must be treated as a separate problem.
2970 This issue was reported to OpenSSL by the OSS-Fuzz project.
2975 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2977 * ChaCha20/Poly1305 heap-buffer-overflow
2979 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2980 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2981 crash. This issue is not considered to be exploitable beyond a DoS.
2983 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2988 * CMS Null dereference
2990 Applications parsing invalid CMS structures can crash with a NULL pointer
2991 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2992 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2993 structure callback if an attempt is made to free certain invalid encodings.
2994 Only CHOICE structures using a callback which do not handle NULL value are
2997 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3002 * Montgomery multiplication may produce incorrect results
3004 There is a carry propagating bug in the Broadwell-specific Montgomery
3005 multiplication procedure that handles input lengths divisible by, but
3006 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3007 and DH private keys are impossible. This is because the subroutine in
3008 question is not used in operations with the private key itself and an input
3009 of the attacker's direct choice. Otherwise the bug can manifest itself as
3010 transient authentication and key negotiation failures or reproducible
3011 erroneous outcome of public-key operations with specially crafted input.
3012 Among EC algorithms only Brainpool P-512 curves are affected and one
3013 presumably can attack ECDH key negotiation. Impact was not analyzed in
3014 detail, because pre-requisites for attack are considered unlikely. Namely
3015 multiple clients have to choose the curve in question and the server has to
3016 share the private key among them, neither of which is default behaviour.
3017 Even then only clients that chose the curve will be affected.
3019 This issue was publicly reported as transient failures and was not
3020 initially recognized as a security issue. Thanks to Richard Morgan for
3021 providing reproducible case.
3026 * Removed automatic addition of RPATH in shared libraries and executables,
3027 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3031 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3033 * Fix Use After Free for large message sizes
3035 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3036 message larger than approx 16k is received then the underlying buffer to
3037 store the incoming message is reallocated and moved. Unfortunately a
3038 dangling pointer to the old location is left which results in an attempt to
3039 write to the previously freed location. This is likely to result in a
3040 crash, however it could potentially lead to execution of arbitrary code.
3042 This issue only affects OpenSSL 1.1.0a.
3044 This issue was reported to OpenSSL by Robert Święcki.
3049 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3051 * OCSP Status Request extension unbounded memory growth
3053 A malicious client can send an excessively large OCSP Status Request
3054 extension. If that client continually requests renegotiation, sending a
3055 large OCSP Status Request extension each time, then there will be unbounded
3056 memory growth on the server. This will eventually lead to a Denial Of
3057 Service attack through memory exhaustion. Servers with a default
3058 configuration are vulnerable even if they do not support OCSP. Builds using
3059 the "no-ocsp" build time option are not affected.
3061 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3066 * SSL_peek() hang on empty record
3068 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3069 sends an empty record. This could be exploited by a malicious peer in a
3070 Denial Of Service attack.
3072 This issue was reported to OpenSSL by Alex Gaynor.
3077 * Excessive allocation of memory in tls_get_message_header() and
3078 dtls1_preprocess_fragment()
3080 A (D)TLS message includes 3 bytes for its length in the header for the
3081 message. This would allow for messages up to 16Mb in length. Messages of
3082 this length are excessive and OpenSSL includes a check to ensure that a
3083 peer is sending reasonably sized messages in order to avoid too much memory
3084 being consumed to service a connection. A flaw in the logic of version
3085 1.1.0 means that memory for the message is allocated too early, prior to
3086 the excessive message length check. Due to way memory is allocated in
3087 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3088 to service a connection. This could lead to a Denial of Service through
3089 memory exhaustion. However, the excessive message length check still takes
3090 place, and this would cause the connection to immediately fail. Assuming
3091 that the application calls SSL_free() on the failed connection in a timely
3092 manner then the 21Mb of allocated memory will then be immediately freed
3093 again. Therefore the excessive memory allocation will be transitory in
3094 nature. This then means that there is only a security impact if:
3096 1) The application does not call SSL_free() in a timely manner in the event
3097 that the connection fails
3099 2) The application is working in a constrained environment where there is
3100 very little free memory
3102 3) The attacker initiates multiple connection attempts such that there are
3103 multiple connections in a state where memory has been allocated for the
3104 connection; SSL_free() has not yet been called; and there is insufficient
3105 memory to service the multiple requests.
3107 Except in the instance of (1) above any Denial Of Service is likely to be
3108 transitory because as soon as the connection fails the memory is
3109 subsequently freed again in the SSL_free() call. However there is an
3110 increased risk during this period of application crashes due to the lack of
3111 memory - which would then mean a more serious Denial of Service.
3113 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3114 (CVE-2016-6307 and CVE-2016-6308)
3118 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3119 had to be removed. Primary reason is that vendor assembler can't
3120 assemble our modules with -KPIC flag. As result it, assembly
3121 support, was not even available as option. But its lack means
3122 lack of side-channel resistant code, which is incompatible with
3123 security by todays standards. Fortunately gcc is readily available
3124 prepackaged option, which we firmly point at...
3128 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3130 * Windows command-line tool supports UTF-8 opt-in option for arguments
3131 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3132 (to any value) allows Windows user to access PKCS#12 file generated
3133 with Windows CryptoAPI and protected with non-ASCII password, as well
3134 as files generated under UTF-8 locale on Linux also protected with
3139 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3140 have been disabled by default and removed from DEFAULT, just like RC4.
3141 See the RC4 item below to re-enable both.
3145 * The method for finding the storage location for the Windows RAND seed file
3146 has changed. First we check %RANDFILE%. If that is not set then we check
3147 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3148 all else fails we fall back to C:\.
3152 * The EVP_EncryptUpdate() function has had its return type changed from void
3153 to int. A return of 0 indicates and error while a return of 1 indicates
3158 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3159 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3160 off the constant time implementation for RSA, DSA and DH have been made
3161 no-ops and deprecated.
3165 * Windows RAND implementation was simplified to only get entropy by
3166 calling CryptGenRandom(). Various other RAND-related tickets
3169 *Joseph Wylie Yandle, Rich Salz*
3171 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3172 and `OPENSSL_LH_`, respectively. The old names are available
3173 with API compatibility. They new names are now completely documented.
3177 * Unify TYPE_up_ref(obj) methods signature.
3178 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3179 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3180 int (instead of void) like all others TYPE_up_ref() methods.
3181 So now these methods also check the return value of CRYPTO_atomic_add(),
3182 and the validity of object reference counter.
3184 *fdasilvayy@gmail.com*
3186 * With Windows Visual Studio builds, the .pdb files are installed
3187 alongside the installed libraries and executables. For a static
3188 library installation, ossl_static.pdb is the associate compiler
3189 generated .pdb file to be used when linking programs.
3193 * Remove openssl.spec. Packaging files belong with the packagers.
3197 * Automatic Darwin/OSX configuration has had a refresh, it will now
3198 recognise x86_64 architectures automatically. You can still decide
3199 to build for a different bitness with the environment variable
3200 KERNEL_BITS (can be 32 or 64), for example:
3202 KERNEL_BITS=32 ./config
3206 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3207 256 bit AES and HMAC with SHA256.
3211 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3215 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3219 * To enable users to have their own config files and build file templates,
3220 Configure looks in the directory indicated by the environment variable
3221 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3222 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3223 name and is used as is.
3227 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3228 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3229 X509_CERT_FILE_CTX was removed.
3233 * "shared" builds are now the default. To create only static libraries use
3234 the "no-shared" Configure option.
3238 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3239 All of these option have not worked for some while and are fundamental
3244 * Make various cleanup routines no-ops and mark them as deprecated. Most
3245 global cleanup functions are no longer required because they are handled
3246 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3247 Explicitly de-initing can cause problems (e.g. where a library that uses
3248 OpenSSL de-inits, but an application is still using it). The affected
3249 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3250 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3251 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3252 COMP_zlib_cleanup().
3256 * --strict-warnings no longer enables runtime debugging options
3257 such as REF_DEBUG. Instead, debug options are automatically
3258 enabled with '--debug' builds.
3260 *Andy Polyakov, Emilia Käsper*
3262 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3263 have been moved out of the public header files. New functions for managing
3264 these have been added.
3268 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3269 objects have been moved out of the public header files. New
3270 functions for managing these have been added.
3274 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3275 have been moved out of the public header files. New functions for managing
3276 these have been added.
3280 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3281 moved out of the public header files. New functions for managing these
3286 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3290 * Removed the mk1mf build scripts.
3294 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3295 it is always safe to #include a header now.
3299 * Removed the aged BC-32 config and all its supporting scripts
3303 * Removed support for Ultrix, Netware, and OS/2.
3307 * Add support for HKDF.
3309 *Alessandro Ghedini*
3311 * Add support for blake2b and blake2s
3315 * Added support for "pipelining". Ciphers that have the
3316 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3317 encryptions/decryptions simultaneously. There are currently no built-in
3318 ciphers with this property but the expectation is that engines will be able
3319 to offer it to significantly improve throughput. Support has been extended
3320 into libssl so that multiple records for a single connection can be
3321 processed in one go (for >=TLS 1.1).
3325 * Added the AFALG engine. This is an async capable engine which is able to
3326 offload work to the Linux kernel. In this initial version it only supports
3327 AES128-CBC. The kernel must be version 4.1.0 or greater.
3331 * OpenSSL now uses a new threading API. It is no longer necessary to
3332 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3333 are two supported threading models: pthreads and windows threads. It is
3334 also possible to configure OpenSSL at compile time for "no-threads". The
3335 old threading API should no longer be used. The functions have been
3336 replaced with "no-op" compatibility macros.
3338 *Alessandro Ghedini, Matt Caswell*
3340 * Modify behavior of ALPN to invoke callback after SNI/servername
3341 callback, such that updates to the SSL_CTX affect ALPN.
3345 * Add SSL_CIPHER queries for authentication and key-exchange.
3349 * Changes to the DEFAULT cipherlist:
3350 - Prefer (EC)DHE handshakes over plain RSA.
3351 - Prefer AEAD ciphers over legacy ciphers.
3352 - Prefer ECDSA over RSA when both certificates are available.
3353 - Prefer TLSv1.2 ciphers/PRF.
3354 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3359 * Change the ECC default curve list to be this, in order: x25519,
3360 secp256r1, secp521r1, secp384r1.
3364 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3365 disabled by default. They can be re-enabled using the
3366 enable-weak-ssl-ciphers option to Configure.
3370 * If the server has ALPN configured, but supports no protocols that the
3371 client advertises, send a fatal "no_application_protocol" alert.
3372 This behaviour is SHALL in RFC 7301, though it isn't universally
3373 implemented by other servers.
3377 * Add X25519 support.
3378 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3379 for public and private key encoding using the format documented in
3380 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3381 key generation and key derivation.
3383 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3388 * Deprecate SRP_VBASE_get_by_user.
3389 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3390 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3391 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3392 seed, even if the seed is configured.
3394 Users should use SRP_VBASE_get1_by_user instead. Note that in
3395 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3396 also that even though configuring the SRP seed attempts to hide
3397 invalid usernames by continuing the handshake with fake
3398 credentials, this behaviour is not constant time and no strong
3399 guarantees are made that the handshake is indistinguishable from
3400 that of a valid user.
3404 * Configuration change; it's now possible to build dynamic engines
3405 without having to build shared libraries and vice versa. This
3406 only applies to the engines in `engines/`, those in `crypto/engine/`
3407 will always be built into libcrypto (i.e. "static").
3409 Building dynamic engines is enabled by default; to disable, use
3410 the configuration option "disable-dynamic-engine".
3412 The only requirements for building dynamic engines are the
3413 presence of the DSO module and building with position independent
3414 code, so they will also automatically be disabled if configuring
3415 with "disable-dso" or "disable-pic".
3417 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3418 are also taken away from openssl/opensslconf.h, as they are
3423 * Configuration change; if there is a known flag to compile
3424 position independent code, it will always be applied on the
3425 libcrypto and libssl object files, and never on the application
3426 object files. This means other libraries that use routines from
3427 libcrypto / libssl can be made into shared libraries regardless
3428 of how OpenSSL was configured.
3430 If this isn't desirable, the configuration options "disable-pic"
3431 or "no-pic" can be used to disable the use of PIC. This will
3432 also disable building shared libraries and dynamic engines.
3436 * Removed JPAKE code. It was experimental and has no wide use.
3440 * The INSTALL_PREFIX Makefile variable has been renamed to
3441 DESTDIR. That makes for less confusion on what this variable
3442 is for. Also, the configuration option --install_prefix is
3447 * Heartbeat for TLS has been removed and is disabled by default
3448 for DTLS; configure with enable-heartbeats. Code that uses the
3449 old #define's might need to be updated.
3451 *Emilia Käsper, Rich Salz*
3453 * Rename REF_CHECK to REF_DEBUG.
3457 * New "unified" build system
3459 The "unified" build system is aimed to be a common system for all
3460 platforms we support. With it comes new support for VMS.
3462 This system builds supports building in a different directory tree
3463 than the source tree. It produces one Makefile (for unix family
3464 or lookalikes), or one descrip.mms (for VMS).
3466 The source of information to make the Makefile / descrip.mms is
3467 small files called 'build.info', holding the necessary
3468 information for each directory with source to compile, and a
3469 template in Configurations, like unix-Makefile.tmpl or
3472 With this change, the library names were also renamed on Windows
3473 and on VMS. They now have names that are closer to the standard
3474 on Unix, and include the major version number, and in certain
3475 cases, the architecture they are built for. See "Notes on shared
3476 libraries" in INSTALL.
3478 We rely heavily on the perl module Text::Template.
3482 * Added support for auto-initialisation and de-initialisation of the library.
3483 OpenSSL no longer requires explicit init or deinit routines to be called,
3484 except in certain circumstances. See the OPENSSL_init_crypto() and
3485 OPENSSL_init_ssl() man pages for further information.
3489 * The arguments to the DTLSv1_listen function have changed. Specifically the
3490 "peer" argument is now expected to be a BIO_ADDR object.
3492 * Rewrite of BIO networking library. The BIO library lacked consistent
3493 support of IPv6, and adding it required some more extensive
3494 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3495 which hold all types of addresses and chains of address information.
3496 It also introduces a new API, with functions like BIO_socket,
3497 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3498 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3499 have been adapted accordingly.
3503 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3508 * CRIME protection: disable compression by default, even if OpenSSL is
3509 compiled with zlib enabled. Applications can still enable compression
3510 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3511 using the SSL_CONF library to configure compression.
3515 * The signature of the session callback configured with
3516 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3517 was explicitly marked as `const unsigned char*` instead of
3522 * Always DPURIFY. Remove the use of uninitialized memory in the
3523 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3527 * Removed many obsolete configuration items, including
3528 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3529 MD2_CHAR, MD2_INT, MD2_LONG
3531 IDEA_SHORT, IDEA_LONG
3532 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3534 *Rich Salz, with advice from Andy Polyakov*
3536 * Many BN internals have been moved to an internal header file.
3538 *Rich Salz with help from Andy Polyakov*
3540 * Configuration and writing out the results from it has changed.
3541 Files such as Makefile include/openssl/opensslconf.h and are now
3542 produced through general templates, such as Makefile.in and
3543 crypto/opensslconf.h.in and some help from the perl module
3546 Also, the center of configuration information is no longer
3547 Makefile. Instead, Configure produces a perl module in
3548 configdata.pm which holds most of the config data (in the hash
3549 table %config), the target data that comes from the target
3550 configuration in one of the `Configurations/*.conf` files (in
3555 * To clarify their intended purposes, the Configure options
3556 --prefix and --openssldir change their semantics, and become more
3557 straightforward and less interdependent.
3559 --prefix shall be used exclusively to give the location INSTALLTOP
3560 where programs, scripts, libraries, include files and manuals are
3561 going to be installed. The default is now /usr/local.
3563 --openssldir shall be used exclusively to give the default
3564 location OPENSSLDIR where certificates, private keys, CRLs are
3565 managed. This is also where the default openssl.cnf gets
3567 If the directory given with this option is a relative path, the
3568 values of both the --prefix value and the --openssldir value will
3569 be combined to become OPENSSLDIR.
3570 The default for --openssldir is INSTALLTOP/ssl.
3572 Anyone who uses --openssldir to specify where OpenSSL is to be
3573 installed MUST change to use --prefix instead.
3577 * The GOST engine was out of date and therefore it has been removed. An up
3578 to date GOST engine is now being maintained in an external repository.
3579 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3580 support for GOST ciphersuites (these are only activated if a GOST engine
3585 * EGD is no longer supported by default; use enable-egd when
3588 *Ben Kaduk and Rich Salz*
3590 * The distribution now has Makefile.in files, which are used to
3591 create Makefile's when Configure is run. *Configure must be run
3592 before trying to build now.*
3596 * The return value for SSL_CIPHER_description() for error conditions
3601 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3603 Obtaining and performing DNSSEC validation of TLSA records is
3604 the application's responsibility. The application provides
3605 the TLSA records of its choice to OpenSSL, and these are then
3606 used to authenticate the peer.
3608 The TLSA records need not even come from DNS. They can, for
3609 example, be used to implement local end-entity certificate or
3610 trust-anchor "pinning", where the "pin" data takes the form
3611 of TLSA records, which can augment or replace verification
3612 based on the usual WebPKI public certification authorities.
3616 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3617 continues to support deprecated interfaces in default builds.
3618 However, applications are strongly advised to compile their
3619 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3620 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3621 or the 1.1.0 releases.
3623 In environments in which all applications have been ported to
3624 not use any deprecated interfaces OpenSSL's Configure script
3625 should be used with the --api=1.1.0 option to entirely remove
3626 support for the deprecated features from the library and
3627 unconditionally disable them in the installed headers.
3628 Essentially the same effect can be achieved with the "no-deprecated"
3629 argument to Configure, except that this will always restrict
3630 the build to just the latest API, rather than a fixed API
3633 As applications are ported to future revisions of the API,
3634 they should update their compile-time OPENSSL_API_COMPAT define
3635 accordingly, but in most cases should be able to continue to
3636 compile with later releases.
3638 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3639 0x10000000L and 0x00908000L, respectively. However those
3640 versions did not support the OPENSSL_API_COMPAT feature, and
3641 so applications are not typically tested for explicit support
3642 of just the undeprecated features of either release.
3646 * Add support for setting the minimum and maximum supported protocol.
3647 It can bet set via the SSL_set_min_proto_version() and
3648 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3649 MaxProtocol. It's recommended to use the new APIs to disable
3650 protocols instead of disabling individual protocols using
3651 SSL_set_options() or SSL_CONF's Protocol. This change also
3652 removes support for disabling TLS 1.2 in the OpenSSL TLS
3653 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3657 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3661 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3662 and integrates ECDSA and ECDH functionality into EC. Implementations can
3663 now redirect key generation and no longer need to convert to or from
3666 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3667 include the ec.h header file instead.
3671 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3672 ciphers who are no longer supported and drops support the ephemeral RSA key
3673 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3677 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3678 opaque. For HMAC_CTX, the following constructors and destructors
3681 HMAC_CTX *HMAC_CTX_new(void);
3682 void HMAC_CTX_free(HMAC_CTX *ctx);
3684 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3685 destroy such methods has been added. See EVP_MD_meth_new(3) and
3686 EVP_CIPHER_meth_new(3) for documentation.
3689 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3690 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3691 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3692 an already created structure.
3693 2) For consistency with the majority of our object creators and
3694 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3695 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3696 for deprecated builds.
3700 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3701 cryptographic operations to be performed asynchronously as long as an
3702 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3703 further details. Libssl has also had this capability integrated with the
3704 introduction of the new mode SSL_MODE_ASYNC and associated error
3705 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3706 pages. This work was developed in partnership with Intel Corp.
3710 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3711 always enabled now. If you want to disable the support you should
3712 exclude it using the list of supported ciphers. This also means that the
3713 "-no_ecdhe" option has been removed from s_server.
3717 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3718 SSL_{CTX_}set1_curves() which can set a list.
3722 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3723 curve you want to support using SSL_{CTX_}set1_curves().
3727 * State machine rewrite. The state machine code has been significantly
3728 refactored in order to remove much duplication of code and solve issues
3729 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3730 further details). This change does have some associated API changes.
3731 Notably the SSL_state() function has been removed and replaced by
3732 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3733 SSL_set_state() has been removed altogether. The previous handshake states
3734 defined in ssl.h and ssl3.h have also been removed.
3738 * All instances of the string "ssleay" in the public API were replaced
3739 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3740 Some error codes related to internal RSA_eay API's were renamed.
3744 * The demo files in crypto/threads were moved to demo/threads.
3748 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3751 *Matt Caswell, Rich Salz*
3753 * New ASN.1 embed macro.
3755 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3756 structure is not allocated: it is part of the parent. That is instead of
3764 This reduces memory fragmentation and make it impossible to accidentally
3765 set a mandatory field to NULL.
3767 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3768 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3769 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3774 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3778 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3779 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3780 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3781 DES and RC4 ciphersuites.
3785 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3786 This changes the decoding behaviour for some invalid messages,
3787 though the change is mostly in the more lenient direction, and
3788 legacy behaviour is preserved as much as possible.
3792 * Fix no-stdio build.
3793 *David Woodhouse <David.Woodhouse@intel.com> and also*
3794 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3796 * New testing framework
3797 The testing framework has been largely rewritten and is now using
3798 perl and the perl modules Test::Harness and an extended variant of
3799 Test::More called OpenSSL::Test to do its work. All test scripts in
3800 test/ have been rewritten into test recipes, and all direct calls to
3801 executables in test/Makefile have become individual recipes using the
3802 simplified testing OpenSSL::Test::Simple.
3804 For documentation on our testing modules, do:
3806 perldoc test/testlib/OpenSSL/Test/Simple.pm
3807 perldoc test/testlib/OpenSSL/Test.pm
3811 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3812 are used; the latter aborts on memory leaks (usually checked on exit).
3813 Some undocumented "set malloc, etc., hooks" functions were removed
3814 and others were changed. All are now documented.
3818 * In DSA_generate_parameters_ex, if the provided seed is too short,
3821 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3823 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3824 from RFC4279, RFC4785, RFC5487, RFC5489.
3826 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3827 original RSA_PSK patch.
3831 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3832 era flag was never set throughout the codebase (only read). Also removed
3833 SSL3_FLAGS_POP_BUFFER which was only used if
3834 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3838 * Changed the default name options in the "ca", "crl", "req" and "x509"
3839 to be "oneline" instead of "compat".
3843 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3844 not aware of clients that still exhibit this bug, and the workaround
3845 hasn't been working properly for a while.
3849 * The return type of BIO_number_read() and BIO_number_written() as well as
3850 the corresponding num_read and num_write members in the BIO structure has
3851 changed from unsigned long to uint64_t. On platforms where an unsigned
3852 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3857 * Given the pervasive nature of TLS extensions it is inadvisable to run
3858 OpenSSL without support for them. It also means that maintaining
3859 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3860 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3864 * Removed support for the two export grade static DH ciphersuites
3865 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3866 were newly added (along with a number of other static DH ciphersuites) to
3867 1.0.2. However the two export ones have *never* worked since they were
3868 introduced. It seems strange in any case to be adding new export
3869 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3873 * Version negotiation has been rewritten. In particular SSLv23_method(),
3874 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3875 and turned into macros which simply call the new preferred function names
3876 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3877 should use the new names instead. Also as part of this change the ssl23.h
3878 header file has been removed.
3882 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3883 code and the associated standard is no longer considered fit-for-purpose.
3887 * RT2547 was closed. When generating a private key, try to make the
3888 output file readable only by the owner. This behavior change might
3889 be noticeable when interacting with other software.
3891 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3896 * Added HTTP GET support to the ocsp command.
3900 * Changed default digest for the dgst and enc commands from MD5 to
3905 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3909 * Added support for TLS extended master secret from
3910 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3911 initial patch which was a great help during development.
3915 * All libssl internal structures have been removed from the public header
3916 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3917 now redundant). Users should not attempt to access internal structures
3918 directly. Instead they should use the provided API functions.
3922 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3923 Access to deprecated functions can be re-enabled by running config with
3924 "enable-deprecated". In addition applications wishing to use deprecated
3925 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3926 will, by default, disable some transitive includes that previously existed
3927 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3931 * Added support for OCB mode. OpenSSL has been granted a patent license
3932 compatible with the OpenSSL license for use of OCB. Details are available
3933 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3934 for OCB can be removed by calling config with no-ocb.
3938 * SSLv2 support has been removed. It still supports receiving a SSLv2
3939 compatible client hello.
3943 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3944 done while fixing the error code for the key-too-small case.
3946 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3948 * CA.sh has been removed; use CA.pl instead.
3952 * Removed old DES API.
3956 * Remove various unsupported platforms:
3962 Sinix/ReliantUNIX RM400
3967 16-bit platforms such as WIN16
3971 * Clean up OPENSSL_NO_xxx #define's
3972 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3973 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3974 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3975 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3976 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3977 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3978 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3979 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3980 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3981 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3985 * Cleaned up dead code
3986 Remove all but one '#ifdef undef' which is to be looked at.
3990 * Clean up calling of xxx_free routines.
3991 Just like free(), fix most of the xxx_free routines to accept
3992 NULL. Remove the non-null checks from callers. Save much code.
3996 * Add secure heap for storage of private keys (when possible).
3997 Add BIO_s_secmem(), CBIGNUM, etc.
3998 Contributed by Akamai Technologies under our Corporate CLA.
4002 * Experimental support for a new, fast, unbiased prime candidate generator,
4003 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4005 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4007 * New output format NSS in the sess_id command line tool. This allows
4008 exporting the session id and the master key in NSS keylog format.
4010 *Martin Kaiser <martin@kaiser.cx>*
4012 * Harmonize version and its documentation. -f flag is used to display
4015 *mancha <mancha1@zoho.com>*
4017 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4018 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4020 *mancha <mancha1@zoho.com>*
4022 * Fix some double frees. These are not thought to be exploitable.
4024 *mancha <mancha1@zoho.com>*
4026 * A missing bounds check in the handling of the TLS heartbeat extension
4027 can be used to reveal up to 64k of memory to a connected client or
4030 Thanks for Neel Mehta of Google Security for discovering this bug and to
4031 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4032 preparing the fix ([CVE-2014-0160])
4034 *Adam Langley, Bodo Moeller*
4036 * Fix for the attack described in the paper "Recovering OpenSSL
4037 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4038 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4039 <http://eprint.iacr.org/2014/140>
4041 Thanks to Yuval Yarom and Naomi Benger for discovering this
4042 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4044 *Yuval Yarom and Naomi Benger*
4046 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4047 this fixes a limitation in previous versions of OpenSSL.
4051 * Experimental encrypt-then-mac support.
4053 Experimental support for encrypt then mac from
4054 draft-gutmann-tls-encrypt-then-mac-02.txt
4056 To enable it set the appropriate extension number (0x42 for the test
4057 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4059 For non-compliant peers (i.e. just about everything) this should have no
4062 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4066 * Add EVP support for key wrapping algorithms, to avoid problems with
4067 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4068 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4069 algorithms and include tests cases.
4073 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4078 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4079 MGF1 digest and OAEP label.
4083 * Make openssl verify return errors.
4085 *Chris Palmer <palmer@google.com> and Ben Laurie*
4087 * New function ASN1_TIME_diff to calculate the difference between two
4088 ASN1_TIME structures or one structure and the current time.
4092 * Update fips_test_suite to support multiple command line options. New
4093 test to induce all self test errors in sequence and check expected
4098 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4099 sign or verify all in one operation.
4103 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4104 test programs and fips_test_suite. Includes functionality to parse
4105 the minimal script output of fipsalgest.pl directly.
4109 * Add authorisation parameter to FIPS_module_mode_set().
4113 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4117 * Use separate DRBG fields for internal and external flags. New function
4118 FIPS_drbg_health_check() to perform on demand health checking. Add
4119 generation tests to fips_test_suite with reduced health check interval to
4120 demonstrate periodic health checking. Add "nodh" option to
4121 fips_test_suite to skip very slow DH test.
4125 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4130 * More extensive health check for DRBG checking many more failure modes.
4131 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4132 combination: call this in fips_test_suite.
4136 * Add support for canonical generation of DSA parameter 'g'. See
4139 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4140 POST to handle HMAC cases.
4144 * Add functions FIPS_module_version() and FIPS_module_version_text()
4145 to return numerical and string versions of the FIPS module number.
4149 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4150 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4151 outside the validated module in the FIPS capable OpenSSL.
4155 * Minor change to DRBG entropy callback semantics. In some cases
4156 there is no multiple of the block length between min_len and
4157 max_len. Allow the callback to return more than max_len bytes
4158 of entropy but discard any extra: it is the callback's responsibility
4159 to ensure that the extra data discarded does not impact the
4160 requested amount of entropy.
4164 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4165 information in FIPS186-3, SP800-57 and SP800-131A.
4169 * CCM support via EVP. Interface is very similar to GCM case except we
4170 must supply all data in one chunk (i.e. no update, final) and the
4171 message length must be supplied if AAD is used. Add algorithm test
4176 * Initial version of POST overhaul. Add POST callback to allow the status
4177 of POST to be monitored and/or failures induced. Modify fips_test_suite
4178 to use callback. Always run all selftests even if one fails.
4182 * XTS support including algorithm test driver in the fips_gcmtest program.
4183 Note: this does increase the maximum key length from 32 to 64 bytes but
4184 there should be no binary compatibility issues as existing applications
4185 will never use XTS mode.
4189 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4190 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4191 performs algorithm blocking for unapproved PRNG types. Also do not
4192 set PRNG type in FIPS_mode_set(): leave this to the application.
4193 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4194 the standard OpenSSL PRNG: set additional data to a date time vector.
4198 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4199 This shouldn't present any incompatibility problems because applications
4200 shouldn't be using these directly and any that are will need to rethink
4201 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4205 * Extensive self tests and health checking required by SP800-90 DRBG.
4206 Remove strength parameter from FIPS_drbg_instantiate and always
4207 instantiate at maximum supported strength.
4211 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4215 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4219 * New function DH_compute_key_padded() to compute a DH key and pad with
4220 leading zeroes if needed: this complies with SP800-56A et al.
4224 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4225 anything, incomplete, subject to change and largely untested at present.
4229 * Modify fipscanisteronly build option to only build the necessary object
4230 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4234 * Add experimental option FIPSSYMS to give all symbols in
4235 fipscanister.o and FIPS or fips prefix. This will avoid
4236 conflicts with future versions of OpenSSL. Add perl script
4237 util/fipsas.pl to preprocess assembly language source files
4238 and rename any affected symbols.
4242 * Add selftest checks and algorithm block of non-fips algorithms in
4243 FIPS mode. Remove DES2 from selftests.
4247 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4248 return internal method without any ENGINE dependencies. Add new
4249 tiny fips sign and verify functions.
4253 * New build option no-ec2m to disable characteristic 2 code.
4257 * New build option "fipscanisteronly". This only builds fipscanister.o
4258 and (currently) associated fips utilities. Uses the file Makefile.fips
4259 instead of Makefile.org as the prototype.
4263 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4264 Update fips_gcmtest to use IV generator.
4268 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4269 setting output buffer to NULL. The `*Final` function must be
4270 called although it will not retrieve any additional data. The tag
4271 can be set or retrieved with a ctrl. The IV length is by default 12
4272 bytes (96 bits) but can be set to an alternative value. If the IV
4273 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4278 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4279 underlying do_cipher function handles all cipher semantics itself
4280 including padding and finalisation. This is useful if (for example)
4281 an ENGINE cipher handles block padding itself. The behaviour of
4282 do_cipher is subtly changed if this flag is set: the return value
4283 is the number of characters written to the output buffer (zero is
4284 no longer an error code) or a negative error code. Also if the
4285 input buffer is NULL and length 0 finalisation should be performed.
4289 * If a candidate issuer certificate is already part of the constructed
4290 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4294 * Improve forward-security support: add functions
4296 void SSL_CTX_set_not_resumable_session_callback(
4297 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4298 void SSL_set_not_resumable_session_callback(
4299 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4301 for use by SSL/TLS servers; the callback function will be called whenever a
4302 new session is created, and gets to decide whether the session may be
4303 cached to make it resumable (return 0) or not (return 1). (As by the
4304 SSL/TLS protocol specifications, the session_id sent by the server will be
4305 empty to indicate that the session is not resumable; also, the server will
4306 not generate RFC 4507 (RFC 5077) session tickets.)
4308 A simple reasonable callback implementation is to return is_forward_secure.
4309 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4310 by the SSL/TLS server library, indicating whether it can provide forward
4313 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4315 * New -verify_name option in command line utilities to set verification
4320 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4321 Add CMAC pkey methods.
4325 * Experimental renegotiation in s_server -www mode. If the client
4326 browses /reneg connection is renegotiated. If /renegcert it is
4327 renegotiated requesting a certificate.
4331 * Add an "external" session cache for debugging purposes to s_server. This
4332 should help trace issues which normally are only apparent in deployed
4333 multi-process servers.
4337 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4338 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4339 BIO_set_cipher() and some obscure PEM functions were changed so they
4340 can now return an error. The RAND changes required a change to the
4341 RAND_METHOD structure.
4345 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4346 a gcc attribute to warn if the result of a function is ignored. This
4347 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4348 whose return value is often ignored.
4352 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4353 These allow SCTs (signed certificate timestamps) to be requested and
4354 validated when establishing a connection.
4356 *Rob Percival <robpercival@google.com>*
4361 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4363 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4364 used even when parsing explicit parameters, when loading a encoded key
4365 or calling `EC_GROUP_new_from_ecpkparameters()`/
4366 `EC_GROUP_new_from_ecparameters()`.
4367 This prevents bypass of security hardening and performance gains,
4368 especially for curves with specialized EC_METHODs.
4369 By default, if a key encoded with explicit parameters is loaded and later
4370 encoded, the output is still encoded with explicit parameters, even if
4371 internally a "named" EC_GROUP is used for computation.
4375 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4376 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4377 NULL. After this change, only the cofactor parameter can be NULL. It also
4378 does some minimal sanity checks on the passed order.
4383 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4384 An attack is simple, if the first CMS_recipientInfo is valid but the
4385 second CMS_recipientInfo is chosen ciphertext. If the second
4386 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4387 encryption key will be replaced by garbage, and the message cannot be
4388 decoded, but if the RSA decryption fails, the correct encryption key is
4389 used and the recipient will not notice the attack.
4390 As a work around for this potential attack the length of the decrypted
4391 key must be equal to the cipher default key length, in case the
4392 certifiate is not given and all recipientInfo are tried out.
4393 The old behaviour can be re-enabled in the CMS code by setting the
4394 CMS_DEBUG_DECRYPT flag.
4399 * Document issue with installation paths in diverse Windows builds
4401 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4402 binaries and run-time config file.
4407 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4409 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4410 This changes the size when using the `genpkey` command when no size is given.
4411 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4412 generation commands to use 2048 bits by default.
4416 * Add FIPS support for Android Arm 64-bit
4418 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4419 Module in Version 2.0.10. For some reason, the corresponding target
4420 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4421 built with FIPS support on Android Arm 64-bit. This omission has been
4424 *Matthias St. Pierre*
4426 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4428 * 0-byte record padding oracle
4430 If an application encounters a fatal protocol error and then calls
4431 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4432 then OpenSSL can respond differently to the calling application if a 0 byte
4433 record is received with invalid padding compared to if a 0 byte record is
4434 received with an invalid MAC. If the application then behaves differently
4435 based on that in a way that is detectable to the remote peer, then this
4436 amounts to a padding oracle that could be used to decrypt data.
4438 In order for this to be exploitable "non-stitched" ciphersuites must be in
4439 use. Stitched ciphersuites are optimised implementations of certain
4440 commonly used ciphersuites. Also the application must call SSL_shutdown()
4441 twice even if a protocol error has occurred (applications should not do
4442 this but some do anyway).
4444 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4445 Aviram, with additional investigation by Steven Collison and Andrew
4446 Hourselt. It was reported to OpenSSL on 10th December 2018.
4451 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4455 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4457 * Microarchitecture timing vulnerability in ECC scalar multiplication
4459 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4460 shown to be vulnerable to a microarchitecture timing side channel attack.
4461 An attacker with sufficient access to mount local timing attacks during
4462 ECDSA signature generation could recover the private key.
4464 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4465 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4471 * Timing vulnerability in DSA signature generation
4473 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4474 timing side channel attack. An attacker could use variations in the signing
4475 algorithm to recover the private key.
4477 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4482 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4483 Module, accidentally introduced while backporting security fixes from the
4484 development branch and hindering the use of ECC in FIPS mode.
4488 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4490 * Client DoS due to large DH parameter
4492 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4493 malicious server can send a very large prime value to the client. This will
4494 cause the client to spend an unreasonably long period of time generating a
4495 key for this prime resulting in a hang until the client has finished. This
4496 could be exploited in a Denial Of Service attack.
4498 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4503 * Cache timing vulnerability in RSA Key Generation
4505 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4506 a cache timing side channel attack. An attacker with sufficient access to
4507 mount cache timing attacks during the RSA key generation process could
4508 recover the private key.
4510 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4511 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4516 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4517 parameter is no longer accepted, as it leads to a corrupt table. NULL
4518 pem_str is reserved for alias entries only.
4522 * Revert blinding in ECDSA sign and instead make problematic addition
4523 length-invariant. Switch even to fixed-length Montgomery multiplication.
4527 * Change generating and checking of primes so that the error rate of not
4528 being prime depends on the intended use based on the size of the input.
4529 For larger primes this will result in more rounds of Miller-Rabin.
4530 The maximal error rate for primes with more than 1080 bits is lowered
4533 *Kurt Roeckx, Annie Yousar*
4535 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4539 * Add blinding to ECDSA and DSA signatures to protect against side channel
4540 attacks discovered by Keegan Ryan (NCC Group).
4544 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4545 now allow empty (zero character) pass phrases.
4549 * Certificate time validation (X509_cmp_time) enforces stricter
4550 compliance with RFC 5280. Fractional seconds and timezone offsets
4551 are no longer allowed.
4555 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4557 * Constructed ASN.1 types with a recursive definition could exceed the stack
4559 Constructed ASN.1 types with a recursive definition (such as can be found
4560 in PKCS7) could eventually exceed the stack given malicious input with
4561 excessive recursion. This could result in a Denial Of Service attack. There
4562 are no such structures used within SSL/TLS that come from untrusted sources
4563 so this is considered safe.
4565 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4571 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4573 * Read/write after SSL object in error state
4575 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4576 mechanism. The intent was that if a fatal error occurred during a handshake
4577 then OpenSSL would move into the error state and would immediately fail if
4578 you attempted to continue the handshake. This works as designed for the
4579 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4580 SSL_connect()), however due to a bug it does not work correctly if
4581 SSL_read() or SSL_write() is called directly. In that scenario, if the
4582 handshake fails then a fatal error will be returned in the initial function
4583 call. If SSL_read()/SSL_write() is subsequently called by the application
4584 for the same SSL object then it will succeed and the data is passed without
4585 being decrypted/encrypted directly from the SSL/TLS record layer.
4587 In order to exploit this issue an application bug would have to be present
4588 that resulted in a call to SSL_read()/SSL_write() being issued after having
4589 already received a fatal error.
4591 This issue was reported to OpenSSL by David Benjamin (Google).
4596 * rsaz_1024_mul_avx2 overflow bug on x86_64
4598 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4599 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4600 Analysis suggests that attacks against RSA and DSA as a result of this
4601 defect would be very difficult to perform and are not believed likely.
4602 Attacks against DH1024 are considered just feasible, because most of the
4603 work necessary to deduce information about a private key may be performed
4604 offline. The amount of resources required for such an attack would be
4605 significant. However, for an attack on TLS to be meaningful, the server
4606 would have to share the DH1024 private key among multiple clients, which is
4607 no longer an option since CVE-2016-0701.
4609 This only affects processors that support the AVX2 but not ADX extensions
4610 like Intel Haswell (4th generation).
4612 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4613 was originally found via the OSS-Fuzz project.
4618 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4620 * bn_sqrx8x_internal carry bug on x86_64
4622 There is a carry propagating bug in the x86_64 Montgomery squaring
4623 procedure. No EC algorithms are affected. Analysis suggests that attacks
4624 against RSA and DSA as a result of this defect would be very difficult to
4625 perform and are not believed likely. Attacks against DH are considered just
4626 feasible (although very difficult) because most of the work necessary to
4627 deduce information about a private key may be performed offline. The amount
4628 of resources required for such an attack would be very significant and
4629 likely only accessible to a limited number of attackers. An attacker would
4630 additionally need online access to an unpatched system using the target
4631 private key in a scenario with persistent DH parameters and a private
4632 key that is shared between multiple clients.
4634 This only affects processors that support the BMI1, BMI2 and ADX extensions
4635 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4637 This issue was reported to OpenSSL by the OSS-Fuzz project.
4642 * Malformed X.509 IPAddressFamily could cause OOB read
4644 If an X.509 certificate has a malformed IPAddressFamily extension,
4645 OpenSSL could do a one-byte buffer overread. The most likely result
4646 would be an erroneous display of the certificate in text format.
4648 This issue was reported to OpenSSL by the OSS-Fuzz project.
4652 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4654 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4655 platform rather than 'mingw'.
4659 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4661 * Truncated packet could crash via OOB read
4663 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4664 cipher is being used, then a truncated packet can cause that host to
4665 perform an out-of-bounds read, usually resulting in a crash.
4667 This issue was reported to OpenSSL by Robert Święcki of Google.
4672 * BN_mod_exp may produce incorrect results on x86_64
4674 There is a carry propagating bug in the x86_64 Montgomery squaring
4675 procedure. No EC algorithms are affected. Analysis suggests that attacks
4676 against RSA and DSA as a result of this defect would be very difficult to
4677 perform and are not believed likely. Attacks against DH are considered just
4678 feasible (although very difficult) because most of the work necessary to
4679 deduce information about a private key may be performed offline. The amount
4680 of resources required for such an attack would be very significant and
4681 likely only accessible to a limited number of attackers. An attacker would
4682 additionally need online access to an unpatched system using the target
4683 private key in a scenario with persistent DH parameters and a private
4684 key that is shared between multiple clients. For example this can occur by
4685 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4686 similar to CVE-2015-3193 but must be treated as a separate problem.
4688 This issue was reported to OpenSSL by the OSS-Fuzz project.
4693 * Montgomery multiplication may produce incorrect results
4695 There is a carry propagating bug in the Broadwell-specific Montgomery
4696 multiplication procedure that handles input lengths divisible by, but
4697 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4698 and DH private keys are impossible. This is because the subroutine in
4699 question is not used in operations with the private key itself and an input
4700 of the attacker's direct choice. Otherwise the bug can manifest itself as
4701 transient authentication and key negotiation failures or reproducible
4702 erroneous outcome of public-key operations with specially crafted input.
4703 Among EC algorithms only Brainpool P-512 curves are affected and one
4704 presumably can attack ECDH key negotiation. Impact was not analyzed in
4705 detail, because pre-requisites for attack are considered unlikely. Namely
4706 multiple clients have to choose the curve in question and the server has to
4707 share the private key among them, neither of which is default behaviour.
4708 Even then only clients that chose the curve will be affected.
4710 This issue was publicly reported as transient failures and was not
4711 initially recognized as a security issue. Thanks to Richard Morgan for
4712 providing reproducible case.
4717 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4718 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4719 prevent issues where no progress is being made and the peer continually
4720 sends unrecognised record types, using up resources processing them.
4724 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4726 * Missing CRL sanity check
4728 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4729 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4730 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4732 This issue only affects the OpenSSL 1.0.2i
4737 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4739 * OCSP Status Request extension unbounded memory growth
4741 A malicious client can send an excessively large OCSP Status Request
4742 extension. If that client continually requests renegotiation, sending a
4743 large OCSP Status Request extension each time, then there will be unbounded
4744 memory growth on the server. This will eventually lead to a Denial Of
4745 Service attack through memory exhaustion. Servers with a default
4746 configuration are vulnerable even if they do not support OCSP. Builds using
4747 the "no-ocsp" build time option are not affected.
4749 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4754 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4757 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4763 * OOB write in MDC2_Update()
4765 An overflow can occur in MDC2_Update() either if called directly or
4766 through the EVP_DigestUpdate() function using MDC2. If an attacker
4767 is able to supply very large amounts of input data after a previous
4768 call to EVP_EncryptUpdate() with a partial block then a length check
4769 can overflow resulting in a heap corruption.
4771 The amount of data needed is comparable to SIZE_MAX which is impractical
4774 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4779 * Malformed SHA512 ticket DoS
4781 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4782 DoS attack where a malformed ticket will result in an OOB read which will
4785 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4786 a custom server callback and ticket lookup mechanism.
4788 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4793 * OOB write in BN_bn2dec()
4795 The function BN_bn2dec() does not check the return value of BN_div_word().
4796 This can cause an OOB write if an application uses this function with an
4797 overly large BIGNUM. This could be a problem if an overly large certificate
4798 or CRL is printed out from an untrusted source. TLS is not affected because
4799 record limits will reject an oversized certificate before it is parsed.
4801 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4806 * OOB read in TS_OBJ_print_bio()
4808 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4809 the total length the OID text representation would use and not the amount
4810 of data written. This will result in OOB reads when large OIDs are
4813 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4818 * Pointer arithmetic undefined behaviour
4820 Avoid some undefined pointer arithmetic
4822 A common idiom in the codebase is to check limits in the following manner:
4825 Where "p" points to some malloc'd data of SIZE bytes and
4828 "len" here could be from some externally supplied data (e.g. from a TLS
4831 The rules of C pointer arithmetic are such that "p + len" is only well
4832 defined where len <= SIZE. Therefore the above idiom is actually
4833 undefined behaviour.
4835 For example this could cause problems if some malloc implementation
4836 provides an address for "p" such that "p + len" actually overflows for
4837 values of len that are too big and therefore p + len < limit.
4839 This issue was reported to OpenSSL by Guido Vranken
4844 * Constant time flag not preserved in DSA signing
4846 Operations in the DSA signing algorithm should run in constant time in
4847 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4848 implementation means that a non-constant time codepath is followed for
4849 certain operations. This has been demonstrated through a cache-timing
4850 attack to be sufficient for an attacker to recover the private DSA key.
4852 This issue was reported by César Pereida (Aalto University), Billy Brumley
4853 (Tampere University of Technology), and Yuval Yarom (The University of
4854 Adelaide and NICTA).
4859 * DTLS buffered message DoS
4861 In a DTLS connection where handshake messages are delivered out-of-order
4862 those messages that OpenSSL is not yet ready to process will be buffered
4863 for later use. Under certain circumstances, a flaw in the logic means that
4864 those messages do not get removed from the buffer even though the handshake
4865 has been completed. An attacker could force up to approx. 15 messages to
4866 remain in the buffer when they are no longer required. These messages will
4867 be cleared when the DTLS connection is closed. The default maximum size for
4868 a message is 100k. Therefore the attacker could force an additional 1500k
4869 to be consumed per connection. By opening many simulataneous connections an
4870 attacker could cause a DoS attack through memory exhaustion.
4872 This issue was reported to OpenSSL by Quan Luo.
4877 * DTLS replay protection DoS
4879 A flaw in the DTLS replay attack protection mechanism means that records
4880 that arrive for future epochs update the replay protection "window" before
4881 the MAC for the record has been validated. This could be exploited by an
4882 attacker by sending a record for the next epoch (which does not have to
4883 decrypt or have a valid MAC), with a very large sequence number. This means
4884 that all subsequent legitimate packets are dropped causing a denial of
4885 service for a specific DTLS connection.
4887 This issue was reported to OpenSSL by the OCAP audit team.
4892 * Certificate message OOB reads
4894 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4895 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4896 theoretical DoS risk but this has not been observed in practice on common
4899 The messages affected are client certificate, client certificate request
4900 and server certificate. As a result the attack can only be performed
4901 against a client or a server which enables client authentication.
4903 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4908 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4910 * Prevent padding oracle in AES-NI CBC MAC check
4912 A MITM attacker can use a padding oracle attack to decrypt traffic
4913 when the connection uses an AES CBC cipher and the server support
4916 This issue was introduced as part of the fix for Lucky 13 padding
4917 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4918 constant time by making sure that always the same bytes are read and
4919 compared against either the MAC or padding bytes. But it no longer
4920 checked that there was enough data to have both the MAC and padding
4923 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4927 * Fix EVP_EncodeUpdate overflow
4929 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4930 Base64 encoding of binary data. If an attacker is able to supply very large
4931 amounts of input data then a length check can overflow resulting in a heap
4934 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4935 the `PEM_write_bio*` family of functions. These are mainly used within the
4936 OpenSSL command line applications, so any application which processes data
4937 from an untrusted source and outputs it as a PEM file should be considered
4938 vulnerable to this issue. User applications that call these APIs directly
4939 with large amounts of untrusted data may also be vulnerable.
4941 This issue was reported by Guido Vranken.
4946 * Fix EVP_EncryptUpdate overflow
4948 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4949 is able to supply very large amounts of input data after a previous call to
4950 EVP_EncryptUpdate() with a partial block then a length check can overflow
4951 resulting in a heap corruption. Following an analysis of all OpenSSL
4952 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4953 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4954 the first called function after an EVP_EncryptInit(), and therefore that
4955 specific call must be safe. The second form is where the length passed to
4956 EVP_EncryptUpdate() can be seen from the code to be some small value and
4957 therefore there is no possibility of an overflow. Since all instances are
4958 one of these two forms, it is believed that there can be no overflows in
4959 internal code due to this problem. It should be noted that
4960 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4961 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4962 of these calls have also been analysed too and it is believed there are no
4963 instances in internal usage where an overflow could occur.
4965 This issue was reported by Guido Vranken.
4970 * Prevent ASN.1 BIO excessive memory allocation
4972 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4973 a short invalid encoding can cause allocation of large amounts of memory
4974 potentially consuming excessive resources or exhausting memory.
4976 Any application parsing untrusted data through d2i BIO functions is
4977 affected. The memory based functions such as d2i_X509() are *not* affected.
4978 Since the memory based functions are used by the TLS library, TLS
4979 applications are not affected.
4981 This issue was reported by Brian Carpenter.
4988 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4989 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4990 in arbitrary stack data being returned in the buffer.
4992 This issue was reported by Guido Vranken.
4997 * Modify behavior of ALPN to invoke callback after SNI/servername
4998 callback, such that updates to the SSL_CTX affect ALPN.
5002 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5007 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5008 methods are enabled and ssl2 is disabled the methods return NULL.
5012 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5014 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5015 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5016 provide any "EXPORT" or "LOW" strength ciphers.
5020 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5021 is by default disabled at build-time. Builds that are not configured with
5022 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5023 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5024 will need to explicitly call either of:
5026 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5028 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5030 as appropriate. Even if either of those is used, or the application
5031 explicitly uses the version-specific SSLv2_method() or its client and
5032 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5033 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5034 ciphers, and SSLv2 56-bit DES are no longer available.
5039 * Fix a double-free in DSA code
5041 A double free bug was discovered when OpenSSL parses malformed DSA private
5042 keys and could lead to a DoS attack or memory corruption for applications
5043 that receive DSA private keys from untrusted sources. This scenario is
5046 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5052 * Disable SRP fake user seed to address a server memory leak.
5054 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5056 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5057 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5058 was changed to ignore the "fake user" SRP seed, even if the seed
5061 Users should use SRP_VBASE_get1_by_user instead. Note that in
5062 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5063 also that even though configuring the SRP seed attempts to hide
5064 invalid usernames by continuing the handshake with fake
5065 credentials, this behaviour is not constant time and no strong
5066 guarantees are made that the handshake is indistinguishable from
5067 that of a valid user.
5072 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5074 In the BN_hex2bn function the number of hex digits is calculated using an
5075 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5076 large values of `i` this can result in `bn_expand` not allocating any
5077 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5078 field as NULL leading to a subsequent NULL ptr deref. For very large values
5079 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5080 In this case memory is allocated to the internal BIGNUM data field, but it
5081 is insufficiently sized leading to heap corruption. A similar issue exists
5082 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5083 is ever called by user applications with very large untrusted hex/dec data.
5084 This is anticipated to be a rare occurrence.
5086 All OpenSSL internal usage of these functions use data that is not expected
5087 to be untrusted, e.g. config file data or application command line
5088 arguments. If user developed applications generate config file data based
5089 on untrusted data then it is possible that this could also lead to security
5090 consequences. This is also anticipated to be rare.
5092 This issue was reported to OpenSSL by Guido Vranken.
5097 * Fix memory issues in `BIO_*printf` functions
5099 The internal `fmtstr` function used in processing a "%s" format string in
5100 the `BIO_*printf` functions could overflow while calculating the length of a
5101 string and cause an OOB read when printing very long strings.
5103 Additionally the internal `doapr_outch` function can attempt to write to an
5104 OOB memory location (at an offset from the NULL pointer) in the event of a
5105 memory allocation failure. In 1.0.2 and below this could be caused where
5106 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5107 could be in processing a very long "%s" format string. Memory leaks can
5110 The first issue may mask the second issue dependent on compiler behaviour.
5111 These problems could enable attacks where large amounts of untrusted data
5112 is passed to the `BIO_*printf` functions. If applications use these functions
5113 in this way then they could be vulnerable. OpenSSL itself uses these
5114 functions when printing out human-readable dumps of ASN.1 data. Therefore
5115 applications that print this data could be vulnerable if the data is from
5116 untrusted sources. OpenSSL command line applications could also be
5117 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5118 as command line arguments.
5120 Libssl is not considered directly vulnerable. Additionally certificates etc
5121 received via remote connections via libssl are also unlikely to be able to
5122 trigger these issues because of message size limits enforced within libssl.
5124 This issue was reported to OpenSSL Guido Vranken.
5129 * Side channel attack on modular exponentiation
5131 A side-channel attack was found which makes use of cache-bank conflicts on
5132 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5133 of RSA keys. The ability to exploit this issue is limited as it relies on
5134 an attacker who has control of code in a thread running on the same
5135 hyper-threaded core as the victim thread which is performing decryptions.
5137 This issue was reported to OpenSSL by Yuval Yarom, The University of
5138 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5139 Nadia Heninger, University of Pennsylvania with more information at
5140 <http://cachebleed.info>.
5145 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5146 if no keysize is specified with default_bits. This fixes an
5147 omission in an earlier change that changed all RSA/DSA key generation
5148 commands to use 2048 bits by default.
5152 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5154 * DH small subgroups
5156 Historically OpenSSL only ever generated DH parameters based on "safe"
5157 primes. More recently (in version 1.0.2) support was provided for
5158 generating X9.42 style parameter files such as those required for RFC 5114
5159 support. The primes used in such files may not be "safe". Where an
5160 application is using DH configured with parameters based on primes that are
5161 not "safe" then an attacker could use this fact to find a peer's private
5162 DH exponent. This attack requires that the attacker complete multiple
5163 handshakes in which the peer uses the same private DH exponent. For example
5164 this could be used to discover a TLS server's private DH exponent if it's
5165 reusing the private DH exponent or it's using a static DH ciphersuite.
5167 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5168 TLS. It is not on by default. If the option is not set then the server
5169 reuses the same private DH exponent for the life of the server process and
5170 would be vulnerable to this attack. It is believed that many popular
5171 applications do set this option and would therefore not be at risk.
5173 The fix for this issue adds an additional check where a "q" parameter is
5174 available (as is the case in X9.42 based parameters). This detects the
5175 only known attack, and is the only possible defense for static DH
5176 ciphersuites. This could have some performance impact.
5178 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5179 default and cannot be disabled. This could have some performance impact.
5181 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5186 * SSLv2 doesn't block disabled ciphers
5188 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5189 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5190 been disabled, provided that the SSLv2 protocol was not also disabled via
5193 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5194 and Sebastian Schinzel.
5199 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5201 * BN_mod_exp may produce incorrect results on x86_64
5203 There is a carry propagating bug in the x86_64 Montgomery squaring
5204 procedure. No EC algorithms are affected. Analysis suggests that attacks
5205 against RSA and DSA as a result of this defect would be very difficult to
5206 perform and are not believed likely. Attacks against DH are considered just
5207 feasible (although very difficult) because most of the work necessary to
5208 deduce information about a private key may be performed offline. The amount
5209 of resources required for such an attack would be very significant and
5210 likely only accessible to a limited number of attackers. An attacker would
5211 additionally need online access to an unpatched system using the target
5212 private key in a scenario with persistent DH parameters and a private
5213 key that is shared between multiple clients. For example this can occur by
5214 default in OpenSSL DHE based SSL/TLS ciphersuites.
5216 This issue was reported to OpenSSL by Hanno Böck.
5221 * Certificate verify crash with missing PSS parameter
5223 The signature verification routines will crash with a NULL pointer
5224 dereference if presented with an ASN.1 signature using the RSA PSS
5225 algorithm and absent mask generation function parameter. Since these
5226 routines are used to verify certificate signature algorithms this can be
5227 used to crash any certificate verification operation and exploited in a
5228 DoS attack. Any application which performs certificate verification is
5229 vulnerable including OpenSSL clients and servers which enable client
5232 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5237 * X509_ATTRIBUTE memory leak
5239 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5240 memory. This structure is used by the PKCS#7 and CMS routines so any
5241 application which reads PKCS#7 or CMS data from untrusted sources is
5242 affected. SSL/TLS is not affected.
5244 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5250 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5251 This changes the decoding behaviour for some invalid messages,
5252 though the change is mostly in the more lenient direction, and
5253 legacy behaviour is preserved as much as possible.
5257 * In DSA_generate_parameters_ex, if the provided seed is too short,
5260 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5262 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5264 * Alternate chains certificate forgery
5266 During certificate verification, OpenSSL will attempt to find an
5267 alternative certificate chain if the first attempt to build such a chain
5268 fails. An error in the implementation of this logic can mean that an
5269 attacker could cause certain checks on untrusted certificates to be
5270 bypassed, such as the CA flag, enabling them to use a valid leaf
5271 certificate to act as a CA and "issue" an invalid certificate.
5273 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5278 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5280 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5281 incompatibility in the handling of HMAC. The previous ABI has now been
5286 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5288 * Malformed ECParameters causes infinite loop
5290 When processing an ECParameters structure OpenSSL enters an infinite loop
5291 if the curve specified is over a specially malformed binary polynomial
5294 This can be used to perform denial of service against any
5295 system which processes public keys, certificate requests or
5296 certificates. This includes TLS clients and TLS servers with
5297 client authentication enabled.
5299 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5304 * Exploitable out-of-bounds read in X509_cmp_time
5306 X509_cmp_time does not properly check the length of the ASN1_TIME
5307 string and can read a few bytes out of bounds. In addition,
5308 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5311 An attacker can use this to craft malformed certificates and CRLs of
5312 various sizes and potentially cause a segmentation fault, resulting in
5313 a DoS on applications that verify certificates or CRLs. TLS clients
5314 that verify CRLs are affected. TLS clients and servers with client
5315 authentication enabled may be affected if they use custom verification
5318 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5319 independently by Hanno Böck.
5324 * PKCS7 crash with missing EnvelopedContent
5326 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5327 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5328 with missing content and trigger a NULL pointer dereference on parsing.
5330 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5331 structures from untrusted sources are affected. OpenSSL clients and
5332 servers are not affected.
5334 This issue was reported to OpenSSL by Michal Zalewski (Google).
5339 * CMS verify infinite loop with unknown hash function
5341 When verifying a signedData message the CMS code can enter an infinite loop
5342 if presented with an unknown hash function OID. This can be used to perform
5343 denial of service against any system which verifies signedData messages using
5345 This issue was reported to OpenSSL by Johannes Bauer.
5350 * Race condition handling NewSessionTicket
5352 If a NewSessionTicket is received by a multi-threaded client when attempting to
5353 reuse a previous ticket then a race condition can occur potentially leading to
5354 a double free of the ticket data.
5359 * Only support 256-bit or stronger elliptic curves with the
5360 'ecdh_auto' setting (server) or by default (client). Of supported
5361 curves, prefer P-256 (both).
5365 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5367 * ClientHello sigalgs DoS fix
5369 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5370 invalid signature algorithms extension a NULL pointer dereference will
5371 occur. This can be exploited in a DoS attack against the server.
5373 This issue was was reported to OpenSSL by David Ramos of Stanford
5377 *Stephen Henson and Matt Caswell*
5379 * Multiblock corrupted pointer fix
5381 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5382 feature only applies on 64 bit x86 architecture platforms that support AES
5383 NI instructions. A defect in the implementation of "multiblock" can cause
5384 OpenSSL's internal write buffer to become incorrectly set to NULL when
5385 using non-blocking IO. Typically, when the user application is using a
5386 socket BIO for writing, this will only result in a failed connection.
5387 However if some other BIO is used then it is likely that a segmentation
5388 fault will be triggered, thus enabling a potential DoS attack.
5390 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5395 * Segmentation fault in DTLSv1_listen fix
5397 The DTLSv1_listen function is intended to be stateless and processes the
5398 initial ClientHello from many peers. It is common for user code to loop
5399 over the call to DTLSv1_listen until a valid ClientHello is received with
5400 an associated cookie. A defect in the implementation of DTLSv1_listen means
5401 that state is preserved in the SSL object from one invocation to the next
5402 that can lead to a segmentation fault. Errors processing the initial
5403 ClientHello can trigger this scenario. An example of such an error could be
5404 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5407 This issue was reported to OpenSSL by Per Allansson.
5412 * Segmentation fault in ASN1_TYPE_cmp fix
5414 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5415 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5416 certificate signature algorithm consistency this can be used to crash any
5417 certificate verification operation and exploited in a DoS attack. Any
5418 application which performs certificate verification is vulnerable including
5419 OpenSSL clients and servers which enable client authentication.
5424 * Segmentation fault for invalid PSS parameters fix
5426 The signature verification routines will crash with a NULL pointer
5427 dereference if presented with an ASN.1 signature using the RSA PSS
5428 algorithm and invalid parameters. Since these routines are used to verify
5429 certificate signature algorithms this can be used to crash any
5430 certificate verification operation and exploited in a DoS attack. Any
5431 application which performs certificate verification is vulnerable including
5432 OpenSSL clients and servers which enable client authentication.
5434 This issue was was reported to OpenSSL by Brian Carpenter.
5439 * ASN.1 structure reuse memory corruption fix
5441 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5442 memory corruption via an invalid write. Such reuse is and has been
5443 strongly discouraged and is believed to be rare.
5445 Applications that parse structures containing CHOICE or ANY DEFINED BY
5446 components may be affected. Certificate parsing (d2i_X509 and related
5447 functions) are however not affected. OpenSSL clients and servers are
5453 * PKCS7 NULL pointer dereferences fix
5455 The PKCS#7 parsing code does not handle missing outer ContentInfo
5456 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5457 missing content and trigger a NULL pointer dereference on parsing.
5459 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5460 otherwise parse PKCS#7 structures from untrusted sources are
5461 affected. OpenSSL clients and servers are not affected.
5463 This issue was reported to OpenSSL by Michal Zalewski (Google).
5468 * DoS via reachable assert in SSLv2 servers fix
5470 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5471 servers that both support SSLv2 and enable export cipher suites by sending
5472 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5474 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5475 (OpenSSL development team).
5480 * Empty CKE with client auth and DHE fix
5482 If client auth is used then a server can seg fault in the event of a DHE
5483 ciphersuite being selected and a zero length ClientKeyExchange message
5484 being sent by the client. This could be exploited in a DoS attack.
5489 * Handshake with unseeded PRNG fix
5491 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5492 with an unseeded PRNG. The conditions are:
5493 - The client is on a platform where the PRNG has not been seeded
5494 automatically, and the user has not seeded manually
5495 - A protocol specific client method version has been used (i.e. not
5496 SSL_client_methodv23)
5497 - A ciphersuite is used that does not require additional random data from
5498 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5500 If the handshake succeeds then the client random that has been used will
5501 have been generated from a PRNG with insufficient entropy and therefore the
5502 output may be predictable.
5504 For example using the following command with an unseeded openssl will
5505 succeed on an unpatched platform:
5507 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5512 * Use After Free following d2i_ECPrivatekey error fix
5514 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5515 could cause a use after free condition. This, in turn, could cause a double
5516 free in several private key parsing functions (such as d2i_PrivateKey
5517 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5518 for applications that receive EC private keys from untrusted
5519 sources. This scenario is considered rare.
5521 This issue was discovered by the BoringSSL project and fixed in their
5527 * X509_to_X509_REQ NULL pointer deref fix
5529 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5530 the certificate key is invalid. This function is rarely used in practice.
5532 This issue was discovered by Brian Carpenter.
5537 * Removed the export ciphers from the DEFAULT ciphers
5541 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5543 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5544 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5545 So far those who have to target multiple platforms would compromise
5546 and argue that binary targeting say ARMv5 would still execute on
5547 ARMv8. "Universal" build resolves this compromise by providing
5548 near-optimal performance even on newer platforms.
5552 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5553 (other platforms pending).
5555 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5557 * Add support for the SignedCertificateTimestampList certificate and
5558 OCSP response extensions from RFC6962.
5562 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5563 for corner cases. (Certain input points at infinity could lead to
5564 bogus results, with non-infinity inputs mapped to infinity too.)
5568 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5569 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5570 common cases are optimized and there still is room for further
5571 improvements. Vector Permutation AES for Altivec is also added.
5575 * Add support for little-endian ppc64 Linux target.
5577 *Marcelo Cerri (IBM)*
5579 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5580 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5581 are optimized and there still is room for further improvements.
5582 Both 32- and 64-bit modes are supported.
5584 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5586 * Improved ARMv7 NEON support.
5590 * Support for SPARC Architecture 2011 crypto extensions, first
5591 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5592 SHA256/512, MD5, GHASH and modular exponentiation.
5594 *Andy Polyakov, David Miller*
5596 * Accelerated modular exponentiation for Intel processors, a.k.a.
5599 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5601 * Support for new and upcoming Intel processors, including AVX2,
5602 BMI and SHA ISA extensions. This includes additional "stitched"
5603 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5606 This work was sponsored by Intel Corp.
5610 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5611 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5612 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5616 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5617 this fixes a limitation in previous versions of OpenSSL.
5621 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5622 MGF1 digest and OAEP label.
5626 * Add EVP support for key wrapping algorithms, to avoid problems with
5627 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5628 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5629 algorithms and include tests cases.
5633 * Add functions to allocate and set the fields of an ECDSA_METHOD
5636 *Douglas E. Engert, Steve Henson*
5638 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5639 difference in days and seconds between two tm or ASN1_TIME structures.
5643 * Add -rev test option to s_server to just reverse order of characters
5644 received by client and send back to server. Also prints an abbreviated
5645 summary of the connection parameters.
5649 * New option -brief for s_client and s_server to print out a brief summary
5650 of connection parameters.
5654 * Add callbacks for arbitrary TLS extensions.
5656 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5658 * New option -crl_download in several openssl utilities to download CRLs
5659 from CRLDP extension in certificates.
5663 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5667 * New function X509_CRL_diff to generate a delta CRL from the difference
5668 of two full CRLs. Add support to "crl" utility.
5672 * New functions to set lookup_crls function and to retrieve
5673 X509_STORE from X509_STORE_CTX.
5677 * Print out deprecated issuer and subject unique ID fields in
5682 * Extend OCSP I/O functions so they can be used for simple general purpose
5683 HTTP as well as OCSP. New wrapper function which can be used to download
5684 CRLs using the OCSP API.
5688 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5692 * `SSL_CONF*` functions. These provide a common framework for application
5693 configuration using configuration files or command lines.
5697 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5698 message callback and prints the results. Needs compile time option
5699 "enable-ssl-trace". New options to s_client and s_server to enable
5704 * New ctrl and macro to retrieve supported points extensions.
5705 Print out extension in s_server and s_client.
5709 * New functions to retrieve certificate signature and signature
5714 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5719 * New Suite B modes for TLS code. These use and enforce the requirements
5720 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5721 only use Suite B curves. The Suite B modes can be set by using the
5722 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5726 * New chain verification flags for Suite B levels of security. Check
5727 algorithms are acceptable when flags are set in X509_verify_cert.
5731 * Make tls1_check_chain return a set of flags indicating checks passed
5732 by a certificate chain. Add additional tests to handle client
5733 certificates: checks for matching certificate type and issuer name
5738 * If an attempt is made to use a signature algorithm not in the peer
5739 preference list abort the handshake. If client has no suitable
5740 signature algorithms in response to a certificate request do not
5741 use the certificate.
5745 * If server EC tmp key is not in client preference list abort handshake.
5749 * Add support for certificate stores in CERT structure. This makes it
5750 possible to have different stores per SSL structure or one store in
5751 the parent SSL_CTX. Include distinct stores for certificate chain
5752 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5753 to build and store a certificate chain in CERT structure: returning
5754 an error if the chain cannot be built: this will allow applications
5755 to test if a chain is correctly configured.
5757 Note: if the CERT based stores are not set then the parent SSL_CTX
5758 store is used to retain compatibility with existing behaviour.
5762 * New function ssl_set_client_disabled to set a ciphersuite disabled
5763 mask based on the current session, check mask when sending client
5764 hello and checking the requested ciphersuite.
5768 * New ctrls to retrieve and set certificate types in a certificate
5769 request message. Print out received values in s_client. If certificate
5770 types is not set with custom values set sensible values based on
5771 supported signature algorithms.
5775 * Support for distinct client and server supported signature algorithms.
5779 * Add certificate callback. If set this is called whenever a certificate
5780 is required by client or server. An application can decide which
5781 certificate chain to present based on arbitrary criteria: for example
5782 supported signature algorithms. Add very simple example to s_server.
5783 This fixes many of the problems and restrictions of the existing client
5784 certificate callback: for example you can now clear an existing
5785 certificate and specify the whole chain.
5789 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5790 the certificate can be used for (if anything). Set valid_flags field
5791 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5792 to have similar checks in it.
5794 Add new "cert_flags" field to CERT structure and include a "strict mode".
5795 This enforces some TLS certificate requirements (such as only permitting
5796 certificate signature algorithms contained in the supported algorithms
5797 extension) which some implementations ignore: this option should be used
5798 with caution as it could cause interoperability issues.
5802 * Update and tidy signature algorithm extension processing. Work out
5803 shared signature algorithms based on preferences and peer algorithms
5804 and print them out in s_client and s_server. Abort handshake if no
5805 shared signature algorithms.
5809 * Add new functions to allow customised supported signature algorithms
5810 for SSL and SSL_CTX structures. Add options to s_client and s_server
5815 * New function SSL_certs_clear() to delete all references to certificates
5816 from an SSL structure. Before this once a certificate had been added
5817 it couldn't be removed.
5821 * Integrate hostname, email address and IP address checking with certificate
5822 verification. New verify options supporting checking in openssl utility.
5826 * Fixes and wildcard matching support to hostname and email checking
5827 functions. Add manual page.
5829 *Florian Weimer (Red Hat Product Security Team)*
5831 * New functions to check a hostname email or IP address against a
5832 certificate. Add options x509 utility to print results of checks against
5837 * Fix OCSP checking.
5839 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5841 * Initial experimental support for explicitly trusted non-root CAs.
5842 OpenSSL still tries to build a complete chain to a root but if an
5843 intermediate CA has a trust setting included that is used. The first
5844 setting is used: whether to trust (e.g., -addtrust option to the x509
5849 * Add -trusted_first option which attempts to find certificates in the
5850 trusted store even if an untrusted chain is also supplied.
5854 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5855 platform support for Linux and Android.
5859 * Support for linux-x32, ILP32 environment in x86_64 framework.
5863 * Experimental multi-implementation support for FIPS capable OpenSSL.
5864 When in FIPS mode the approved implementations are used as normal,
5865 when not in FIPS mode the internal unapproved versions are used instead.
5866 This means that the FIPS capable OpenSSL isn't forced to use the
5867 (often lower performance) FIPS implementations outside FIPS mode.
5871 * Transparently support X9.42 DH parameters when calling
5872 PEM_read_bio_DHparameters. This means existing applications can handle
5873 the new parameter format automatically.
5877 * Initial experimental support for X9.42 DH parameter format: mainly
5878 to support use of 'q' parameter for RFC5114 parameters.
5882 * Add DH parameters from RFC5114 including test data to dhtest.
5886 * Support for automatic EC temporary key parameter selection. If enabled
5887 the most preferred EC parameters are automatically used instead of
5888 hardcoded fixed parameters. Now a server just has to call:
5889 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5890 support ECDH and use the most appropriate parameters.
5894 * Enhance and tidy EC curve and point format TLS extension code. Use
5895 static structures instead of allocation if default values are used.
5896 New ctrls to set curves we wish to support and to retrieve shared curves.
5897 Print out shared curves in s_server. New options to s_server and s_client
5898 to set list of supported curves.
5902 * New ctrls to retrieve supported signature algorithms and
5903 supported curve values as an array of NIDs. Extend openssl utility
5904 to print out received values.
5908 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5909 between NIDs and the more common NIST names such as "P-256". Enhance
5910 ecparam utility and ECC method to recognise the NIST names for curves.
5914 * Enhance SSL/TLS certificate chain handling to support different
5915 chains for each certificate instead of one chain in the parent SSL_CTX.
5919 * Support for fixed DH ciphersuite client authentication: where both
5920 server and client use DH certificates with common parameters.
5924 * Support for fixed DH ciphersuites: those requiring DH server
5929 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5931 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5932 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5933 X509_CINF_get_signature were reverted post internal team review.
5938 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5940 * OCSP Status Request extension unbounded memory growth
5942 A malicious client can send an excessively large OCSP Status Request
5943 extension. If that client continually requests renegotiation, sending a
5944 large OCSP Status Request extension each time, then there will be unbounded
5945 memory growth on the server. This will eventually lead to a Denial Of
5946 Service attack through memory exhaustion. Servers with a default
5947 configuration are vulnerable even if they do not support OCSP. Builds using
5948 the "no-ocsp" build time option are not affected.
5950 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5955 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5958 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5964 * OOB write in MDC2_Update()
5966 An overflow can occur in MDC2_Update() either if called directly or
5967 through the EVP_DigestUpdate() function using MDC2. If an attacker
5968 is able to supply very large amounts of input data after a previous
5969 call to EVP_EncryptUpdate() with a partial block then a length check
5970 can overflow resulting in a heap corruption.
5972 The amount of data needed is comparable to SIZE_MAX which is impractical
5975 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5980 * Malformed SHA512 ticket DoS
5982 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5983 DoS attack where a malformed ticket will result in an OOB read which will
5986 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5987 a custom server callback and ticket lookup mechanism.
5989 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5994 * OOB write in BN_bn2dec()
5996 The function BN_bn2dec() does not check the return value of BN_div_word().
5997 This can cause an OOB write if an application uses this function with an
5998 overly large BIGNUM. This could be a problem if an overly large certificate
5999 or CRL is printed out from an untrusted source. TLS is not affected because
6000 record limits will reject an oversized certificate before it is parsed.
6002 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6007 * OOB read in TS_OBJ_print_bio()
6009 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6010 the total length the OID text representation would use and not the amount
6011 of data written. This will result in OOB reads when large OIDs are
6014 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6019 * Pointer arithmetic undefined behaviour
6021 Avoid some undefined pointer arithmetic
6023 A common idiom in the codebase is to check limits in the following manner:
6026 Where "p" points to some malloc'd data of SIZE bytes and
6029 "len" here could be from some externally supplied data (e.g. from a TLS
6032 The rules of C pointer arithmetic are such that "p + len" is only well
6033 defined where len <= SIZE. Therefore the above idiom is actually
6034 undefined behaviour.
6036 For example this could cause problems if some malloc implementation
6037 provides an address for "p" such that "p + len" actually overflows for
6038 values of len that are too big and therefore p + len < limit.
6040 This issue was reported to OpenSSL by Guido Vranken
6045 * Constant time flag not preserved in DSA signing
6047 Operations in the DSA signing algorithm should run in constant time in
6048 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6049 implementation means that a non-constant time codepath is followed for
6050 certain operations. This has been demonstrated through a cache-timing
6051 attack to be sufficient for an attacker to recover the private DSA key.
6053 This issue was reported by César Pereida (Aalto University), Billy Brumley
6054 (Tampere University of Technology), and Yuval Yarom (The University of
6055 Adelaide and NICTA).
6060 * DTLS buffered message DoS
6062 In a DTLS connection where handshake messages are delivered out-of-order
6063 those messages that OpenSSL is not yet ready to process will be buffered
6064 for later use. Under certain circumstances, a flaw in the logic means that
6065 those messages do not get removed from the buffer even though the handshake
6066 has been completed. An attacker could force up to approx. 15 messages to
6067 remain in the buffer when they are no longer required. These messages will
6068 be cleared when the DTLS connection is closed. The default maximum size for
6069 a message is 100k. Therefore the attacker could force an additional 1500k
6070 to be consumed per connection. By opening many simulataneous connections an
6071 attacker could cause a DoS attack through memory exhaustion.
6073 This issue was reported to OpenSSL by Quan Luo.
6078 * DTLS replay protection DoS
6080 A flaw in the DTLS replay attack protection mechanism means that records
6081 that arrive for future epochs update the replay protection "window" before
6082 the MAC for the record has been validated. This could be exploited by an
6083 attacker by sending a record for the next epoch (which does not have to
6084 decrypt or have a valid MAC), with a very large sequence number. This means
6085 that all subsequent legitimate packets are dropped causing a denial of
6086 service for a specific DTLS connection.
6088 This issue was reported to OpenSSL by the OCAP audit team.
6093 * Certificate message OOB reads
6095 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6096 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6097 theoretical DoS risk but this has not been observed in practice on common
6100 The messages affected are client certificate, client certificate request
6101 and server certificate. As a result the attack can only be performed
6102 against a client or a server which enables client authentication.
6104 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6109 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6111 * Prevent padding oracle in AES-NI CBC MAC check
6113 A MITM attacker can use a padding oracle attack to decrypt traffic
6114 when the connection uses an AES CBC cipher and the server support
6117 This issue was introduced as part of the fix for Lucky 13 padding
6118 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6119 constant time by making sure that always the same bytes are read and
6120 compared against either the MAC or padding bytes. But it no longer
6121 checked that there was enough data to have both the MAC and padding
6124 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6129 * Fix EVP_EncodeUpdate overflow
6131 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6132 Base64 encoding of binary data. If an attacker is able to supply very large
6133 amounts of input data then a length check can overflow resulting in a heap
6136 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6137 the `PEM_write_bio*` family of functions. These are mainly used within the
6138 OpenSSL command line applications, so any application which processes data
6139 from an untrusted source and outputs it as a PEM file should be considered
6140 vulnerable to this issue. User applications that call these APIs directly
6141 with large amounts of untrusted data may also be vulnerable.
6143 This issue was reported by Guido Vranken.
6148 * Fix EVP_EncryptUpdate overflow
6150 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6151 is able to supply very large amounts of input data after a previous call to
6152 EVP_EncryptUpdate() with a partial block then a length check can overflow
6153 resulting in a heap corruption. Following an analysis of all OpenSSL
6154 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6155 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6156 the first called function after an EVP_EncryptInit(), and therefore that
6157 specific call must be safe. The second form is where the length passed to
6158 EVP_EncryptUpdate() can be seen from the code to be some small value and
6159 therefore there is no possibility of an overflow. Since all instances are
6160 one of these two forms, it is believed that there can be no overflows in
6161 internal code due to this problem. It should be noted that
6162 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6163 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6164 of these calls have also been analysed too and it is believed there are no
6165 instances in internal usage where an overflow could occur.
6167 This issue was reported by Guido Vranken.
6172 * Prevent ASN.1 BIO excessive memory allocation
6174 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6175 a short invalid encoding can casuse allocation of large amounts of memory
6176 potentially consuming excessive resources or exhausting memory.
6178 Any application parsing untrusted data through d2i BIO functions is
6179 affected. The memory based functions such as d2i_X509() are *not* affected.
6180 Since the memory based functions are used by the TLS library, TLS
6181 applications are not affected.
6183 This issue was reported by Brian Carpenter.
6190 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6191 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6192 in arbitrary stack data being returned in the buffer.
6194 This issue was reported by Guido Vranken.
6199 * Modify behavior of ALPN to invoke callback after SNI/servername
6200 callback, such that updates to the SSL_CTX affect ALPN.
6204 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6209 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6210 methods are enabled and ssl2 is disabled the methods return NULL.
6214 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6216 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6217 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6218 provide any "EXPORT" or "LOW" strength ciphers.
6222 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6223 is by default disabled at build-time. Builds that are not configured with
6224 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6225 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6226 will need to explicitly call either of:
6228 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6230 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6232 as appropriate. Even if either of those is used, or the application
6233 explicitly uses the version-specific SSLv2_method() or its client and
6234 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6235 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6236 ciphers, and SSLv2 56-bit DES are no longer available.
6241 * Fix a double-free in DSA code
6243 A double free bug was discovered when OpenSSL parses malformed DSA private
6244 keys and could lead to a DoS attack or memory corruption for applications
6245 that receive DSA private keys from untrusted sources. This scenario is
6248 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6254 * Disable SRP fake user seed to address a server memory leak.
6256 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6258 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6259 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6260 was changed to ignore the "fake user" SRP seed, even if the seed
6263 Users should use SRP_VBASE_get1_by_user instead. Note that in
6264 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6265 also that even though configuring the SRP seed attempts to hide
6266 invalid usernames by continuing the handshake with fake
6267 credentials, this behaviour is not constant time and no strong
6268 guarantees are made that the handshake is indistinguishable from
6269 that of a valid user.
6274 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6276 In the BN_hex2bn function the number of hex digits is calculated using an
6277 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6278 large values of `i` this can result in `bn_expand` not allocating any
6279 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6280 field as NULL leading to a subsequent NULL ptr deref. For very large values
6281 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6282 In this case memory is allocated to the internal BIGNUM data field, but it
6283 is insufficiently sized leading to heap corruption. A similar issue exists
6284 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6285 is ever called by user applications with very large untrusted hex/dec data.
6286 This is anticipated to be a rare occurrence.
6288 All OpenSSL internal usage of these functions use data that is not expected
6289 to be untrusted, e.g. config file data or application command line
6290 arguments. If user developed applications generate config file data based
6291 on untrusted data then it is possible that this could also lead to security
6292 consequences. This is also anticipated to be rare.
6294 This issue was reported to OpenSSL by Guido Vranken.
6299 * Fix memory issues in `BIO_*printf` functions
6301 The internal `fmtstr` function used in processing a "%s" format string in
6302 the `BIO_*printf` functions could overflow while calculating the length of a
6303 string and cause an OOB read when printing very long strings.
6305 Additionally the internal `doapr_outch` function can attempt to write to an
6306 OOB memory location (at an offset from the NULL pointer) in the event of a
6307 memory allocation failure. In 1.0.2 and below this could be caused where
6308 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6309 could be in processing a very long "%s" format string. Memory leaks can
6312 The first issue may mask the second issue dependent on compiler behaviour.
6313 These problems could enable attacks where large amounts of untrusted data
6314 is passed to the `BIO_*printf` functions. If applications use these functions
6315 in this way then they could be vulnerable. OpenSSL itself uses these
6316 functions when printing out human-readable dumps of ASN.1 data. Therefore
6317 applications that print this data could be vulnerable if the data is from
6318 untrusted sources. OpenSSL command line applications could also be
6319 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6320 as command line arguments.
6322 Libssl is not considered directly vulnerable. Additionally certificates etc
6323 received via remote connections via libssl are also unlikely to be able to
6324 trigger these issues because of message size limits enforced within libssl.
6326 This issue was reported to OpenSSL Guido Vranken.
6331 * Side channel attack on modular exponentiation
6333 A side-channel attack was found which makes use of cache-bank conflicts on
6334 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6335 of RSA keys. The ability to exploit this issue is limited as it relies on
6336 an attacker who has control of code in a thread running on the same
6337 hyper-threaded core as the victim thread which is performing decryptions.
6339 This issue was reported to OpenSSL by Yuval Yarom, The University of
6340 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6341 Nadia Heninger, University of Pennsylvania with more information at
6342 <http://cachebleed.info>.
6347 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6348 if no keysize is specified with default_bits. This fixes an
6349 omission in an earlier change that changed all RSA/DSA key generation
6350 commands to use 2048 bits by default.
6354 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6356 * Protection for DH small subgroup attacks
6358 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6359 switched on by default and cannot be disabled. This could have some
6364 * SSLv2 doesn't block disabled ciphers
6366 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6367 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6368 been disabled, provided that the SSLv2 protocol was not also disabled via
6371 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6372 and Sebastian Schinzel.
6377 * Reject DH handshakes with parameters shorter than 1024 bits.
6381 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6383 * Certificate verify crash with missing PSS parameter
6385 The signature verification routines will crash with a NULL pointer
6386 dereference if presented with an ASN.1 signature using the RSA PSS
6387 algorithm and absent mask generation function parameter. Since these
6388 routines are used to verify certificate signature algorithms this can be
6389 used to crash any certificate verification operation and exploited in a
6390 DoS attack. Any application which performs certificate verification is
6391 vulnerable including OpenSSL clients and servers which enable client
6394 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6399 * X509_ATTRIBUTE memory leak
6401 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6402 memory. This structure is used by the PKCS#7 and CMS routines so any
6403 application which reads PKCS#7 or CMS data from untrusted sources is
6404 affected. SSL/TLS is not affected.
6406 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6412 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6413 This changes the decoding behaviour for some invalid messages,
6414 though the change is mostly in the more lenient direction, and
6415 legacy behaviour is preserved as much as possible.
6419 * In DSA_generate_parameters_ex, if the provided seed is too short,
6420 use a random seed, as already documented.
6422 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6424 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6426 * Alternate chains certificate forgery
6428 During certificate verfification, OpenSSL will attempt to find an
6429 alternative certificate chain if the first attempt to build such a chain
6430 fails. An error in the implementation of this logic can mean that an
6431 attacker could cause certain checks on untrusted certificates to be
6432 bypassed, such as the CA flag, enabling them to use a valid leaf
6433 certificate to act as a CA and "issue" an invalid certificate.
6435 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6441 * Race condition handling PSK identify hint
6443 If PSK identity hints are received by a multi-threaded client then
6444 the values are wrongly updated in the parent SSL_CTX structure. This can
6445 result in a race condition potentially leading to a double free of the
6451 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6453 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6454 incompatibility in the handling of HMAC. The previous ABI has now been
6457 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6459 * Malformed ECParameters causes infinite loop
6461 When processing an ECParameters structure OpenSSL enters an infinite loop
6462 if the curve specified is over a specially malformed binary polynomial
6465 This can be used to perform denial of service against any
6466 system which processes public keys, certificate requests or
6467 certificates. This includes TLS clients and TLS servers with
6468 client authentication enabled.
6470 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6475 * Exploitable out-of-bounds read in X509_cmp_time
6477 X509_cmp_time does not properly check the length of the ASN1_TIME
6478 string and can read a few bytes out of bounds. In addition,
6479 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6482 An attacker can use this to craft malformed certificates and CRLs of
6483 various sizes and potentially cause a segmentation fault, resulting in
6484 a DoS on applications that verify certificates or CRLs. TLS clients
6485 that verify CRLs are affected. TLS clients and servers with client
6486 authentication enabled may be affected if they use custom verification
6489 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6490 independently by Hanno Böck.
6495 * PKCS7 crash with missing EnvelopedContent
6497 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6498 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6499 with missing content and trigger a NULL pointer dereference on parsing.
6501 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6502 structures from untrusted sources are affected. OpenSSL clients and
6503 servers are not affected.
6505 This issue was reported to OpenSSL by Michal Zalewski (Google).
6510 * CMS verify infinite loop with unknown hash function
6512 When verifying a signedData message the CMS code can enter an infinite loop
6513 if presented with an unknown hash function OID. This can be used to perform
6514 denial of service against any system which verifies signedData messages using
6516 This issue was reported to OpenSSL by Johannes Bauer.
6521 * Race condition handling NewSessionTicket
6523 If a NewSessionTicket is received by a multi-threaded client when attempting to
6524 reuse a previous ticket then a race condition can occur potentially leading to
6525 a double free of the ticket data.
6530 * Reject DH handshakes with parameters shorter than 768 bits.
6532 *Kurt Roeckx and Emilia Kasper*
6534 * dhparam: generate 2048-bit parameters by default.
6536 *Kurt Roeckx and Emilia Kasper*
6538 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6540 * Segmentation fault in ASN1_TYPE_cmp fix
6542 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6543 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6544 certificate signature algorithm consistency this can be used to crash any
6545 certificate verification operation and exploited in a DoS attack. Any
6546 application which performs certificate verification is vulnerable including
6547 OpenSSL clients and servers which enable client authentication.
6552 * ASN.1 structure reuse memory corruption fix
6554 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6555 memory corruption via an invalid write. Such reuse is and has been
6556 strongly discouraged and is believed to be rare.
6558 Applications that parse structures containing CHOICE or ANY DEFINED BY
6559 components may be affected. Certificate parsing (d2i_X509 and related
6560 functions) are however not affected. OpenSSL clients and servers are
6566 * PKCS7 NULL pointer dereferences fix
6568 The PKCS#7 parsing code does not handle missing outer ContentInfo
6569 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6570 missing content and trigger a NULL pointer dereference on parsing.
6572 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6573 otherwise parse PKCS#7 structures from untrusted sources are
6574 affected. OpenSSL clients and servers are not affected.
6576 This issue was reported to OpenSSL by Michal Zalewski (Google).
6581 * DoS via reachable assert in SSLv2 servers fix
6583 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6584 servers that both support SSLv2 and enable export cipher suites by sending
6585 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6587 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6588 (OpenSSL development team).
6593 * Use After Free following d2i_ECPrivatekey error fix
6595 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6596 could cause a use after free condition. This, in turn, could cause a double
6597 free in several private key parsing functions (such as d2i_PrivateKey
6598 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6599 for applications that receive EC private keys from untrusted
6600 sources. This scenario is considered rare.
6602 This issue was discovered by the BoringSSL project and fixed in their
6608 * X509_to_X509_REQ NULL pointer deref fix
6610 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6611 the certificate key is invalid. This function is rarely used in practice.
6613 This issue was discovered by Brian Carpenter.
6618 * Removed the export ciphers from the DEFAULT ciphers
6622 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6624 * Build fixes for the Windows and OpenVMS platforms
6626 *Matt Caswell and Richard Levitte*
6628 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6630 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6631 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6632 dereference. This could lead to a Denial Of Service attack. Thanks to
6633 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6638 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6639 dtls1_buffer_record function under certain conditions. In particular this
6640 could occur if an attacker sent repeated DTLS records with the same
6641 sequence number but for the next epoch. The memory leak could be exploited
6642 by an attacker in a Denial of Service attack through memory exhaustion.
6643 Thanks to Chris Mueller for reporting this issue.
6648 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6649 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6650 method would be set to NULL which could later result in a NULL pointer
6651 dereference. Thanks to Frank Schmirler for reporting this issue.
6656 * Abort handshake if server key exchange message is omitted for ephemeral
6659 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6660 reporting this issue.
6665 * Remove non-export ephemeral RSA code on client and server. This code
6666 violated the TLS standard by allowing the use of temporary RSA keys in
6667 non-export ciphersuites and could be used by a server to effectively
6668 downgrade the RSA key length used to a value smaller than the server
6669 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6670 INRIA or reporting this issue.
6675 * Fixed issue where DH client certificates are accepted without verification.
6676 An OpenSSL server will accept a DH certificate for client authentication
6677 without the certificate verify message. This effectively allows a client to
6678 authenticate without the use of a private key. This only affects servers
6679 which trust a client certificate authority which issues certificates
6680 containing DH keys: these are extremely rare and hardly ever encountered.
6681 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6687 * Ensure that the session ID context of an SSL is updated when its
6688 SSL_CTX is updated via SSL_set_SSL_CTX.
6690 The session ID context is typically set from the parent SSL_CTX,
6691 and can vary with the CTX.
6695 * Fix various certificate fingerprint issues.
6697 By using non-DER or invalid encodings outside the signed portion of a
6698 certificate the fingerprint can be changed without breaking the signature.
6699 Although no details of the signed portion of the certificate can be changed
6700 this can cause problems with some applications: e.g. those using the
6701 certificate fingerprint for blacklists.
6703 1. Reject signatures with non zero unused bits.
6705 If the BIT STRING containing the signature has non zero unused bits reject
6706 the signature. All current signature algorithms require zero unused bits.
6708 2. Check certificate algorithm consistency.
6710 Check the AlgorithmIdentifier inside TBS matches the one in the
6711 certificate signature. NB: this will result in signature failure
6712 errors for some broken certificates.
6714 Thanks to Konrad Kraszewski from Google for reporting this issue.
6716 3. Check DSA/ECDSA signatures use DER.
6718 Re-encode DSA/ECDSA signatures and compare with the original received
6719 signature. Return an error if there is a mismatch.
6721 This will reject various cases including garbage after signature
6722 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6723 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6724 (negative or with leading zeroes).
6726 Further analysis was conducted and fixes were developed by Stephen Henson
6727 of the OpenSSL core team.
6733 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6734 results on some platforms, including x86_64. This bug occurs at random
6735 with a very low probability, and is not known to be exploitable in any
6736 way, though its exact impact is difficult to determine. Thanks to Pieter
6737 Wuille (Blockstream) who reported this issue and also suggested an initial
6738 fix. Further analysis was conducted by the OpenSSL development team and
6739 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6740 the OpenSSL core team.
6745 * Do not resume sessions on the server if the negotiated protocol
6746 version does not match the session's version. Resuming with a different
6747 version, while not strictly forbidden by the RFC, is of questionable
6748 sanity and breaks all known clients.
6750 *David Benjamin, Emilia Käsper*
6752 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6753 early CCS messages during renegotiation. (Note that because
6754 renegotiation is encrypted, this early CCS was not exploitable.)
6758 * Tighten client-side session ticket handling during renegotiation:
6759 ensure that the client only accepts a session ticket if the server sends
6760 the extension anew in the ServerHello. Previously, a TLS client would
6761 reuse the old extension state and thus accept a session ticket if one was
6762 announced in the initial ServerHello.
6764 Similarly, ensure that the client requires a session ticket if one
6765 was advertised in the ServerHello. Previously, a TLS client would
6766 ignore a missing NewSessionTicket message.
6770 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6774 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6775 sends a carefully crafted handshake message, to cause OpenSSL to fail
6776 to free up to 64k of memory causing a memory leak. This could be
6777 exploited in a Denial Of Service attack. This issue affects OpenSSL
6778 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6779 whether SRTP is used or configured. Implementations of OpenSSL that
6780 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6782 The fix was developed by the OpenSSL team.
6787 * Session Ticket Memory Leak.
6789 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6790 integrity of that ticket is first verified. In the event of a session
6791 ticket integrity check failing, OpenSSL will fail to free memory
6792 causing a memory leak. By sending a large number of invalid session
6793 tickets an attacker could exploit this issue in a Denial Of Service
6799 * Build option no-ssl3 is incomplete.
6801 When OpenSSL is configured with "no-ssl3" as a build option, servers
6802 could accept and complete a SSL 3.0 handshake, and clients could be
6803 configured to send them.
6806 *Akamai and the OpenSSL team*
6808 * Add support for TLS_FALLBACK_SCSV.
6809 Client applications doing fallback retries should call
6810 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6813 *Adam Langley, Bodo Moeller*
6815 * Add additional DigestInfo checks.
6817 Re-encode DigestInto in DER and check against the original when
6818 verifying RSA signature: this will reject any improperly encoded
6819 DigestInfo structures.
6821 Note: this is a precautionary measure and no attacks are currently known.
6825 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6827 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6828 SRP code can be overrun an internal buffer. Add sanity check that
6829 g, A, B < N to SRP code.
6831 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6832 Group for discovering this issue.
6837 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6838 TLS 1.0 instead of higher protocol versions when the ClientHello message
6839 is badly fragmented. This allows a man-in-the-middle attacker to force a
6840 downgrade to TLS 1.0 even if both the server and the client support a
6841 higher protocol version, by modifying the client's TLS records.
6843 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6844 researching this issue.
6849 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6850 to a denial of service attack. A malicious server can crash the client
6851 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6852 ciphersuite and sending carefully crafted handshake messages.
6854 Thanks to Felix Gröbert (Google) for discovering and researching this
6860 * By sending carefully crafted DTLS packets an attacker could cause openssl
6861 to leak memory. This can be exploited through a Denial of Service attack.
6862 Thanks to Adam Langley for discovering and researching this issue.
6867 * An attacker can force openssl to consume large amounts of memory whilst
6868 processing DTLS handshake messages. This can be exploited through a
6869 Denial of Service attack.
6870 Thanks to Adam Langley for discovering and researching this issue.
6875 * An attacker can force an error condition which causes openssl to crash
6876 whilst processing DTLS packets due to memory being freed twice. This
6877 can be exploited through a Denial of Service attack.
6878 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6884 * If a multithreaded client connects to a malicious server using a resumed
6885 session and the server sends an ec point format extension it could write
6886 up to 255 bytes to freed memory.
6888 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6894 * A malicious server can crash an OpenSSL client with a null pointer
6895 dereference (read) by specifying an SRP ciphersuite even though it was not
6896 properly negotiated with the client. This can be exploited through a
6897 Denial of Service attack.
6899 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6900 discovering and researching this issue.
6905 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6906 X509_name_oneline, X509_name_print_ex et al. to leak some information
6907 from the stack. Applications may be affected if they echo pretty printing
6908 output to the attacker.
6910 Thanks to Ivan Fratric (Google) for discovering this issue.
6913 *Emilia Käsper, and Steve Henson*
6915 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6916 for corner cases. (Certain input points at infinity could lead to
6917 bogus results, with non-infinity inputs mapped to infinity too.)
6921 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6923 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6924 handshake can force the use of weak keying material in OpenSSL
6925 SSL/TLS clients and servers.
6927 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6928 researching this issue. ([CVE-2014-0224])
6930 *KIKUCHI Masashi, Steve Henson*
6932 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6933 OpenSSL DTLS client the code can be made to recurse eventually crashing
6936 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6939 *Imre Rad, Steve Henson*
6941 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6942 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6943 client or server. This is potentially exploitable to run arbitrary
6944 code on a vulnerable client or server.
6946 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6948 *Jüri Aedla, Steve Henson*
6950 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6951 are subject to a denial of service attack.
6953 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6954 this issue. ([CVE-2014-3470])
6956 *Felix Gröbert, Ivan Fratric, Steve Henson*
6958 * Harmonize version and its documentation. -f flag is used to display
6961 *mancha <mancha1@zoho.com>*
6963 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6964 in i2d_ECPrivateKey.
6966 *mancha <mancha1@zoho.com>*
6968 * Fix some double frees. These are not thought to be exploitable.
6970 *mancha <mancha1@zoho.com>*
6972 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6974 * A missing bounds check in the handling of the TLS heartbeat extension
6975 can be used to reveal up to 64k of memory to a connected client or
6978 Thanks for Neel Mehta of Google Security for discovering this bug and to
6979 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6980 preparing the fix ([CVE-2014-0160])
6982 *Adam Langley, Bodo Moeller*
6984 * Fix for the attack described in the paper "Recovering OpenSSL
6985 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6986 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6987 <http://eprint.iacr.org/2014/140>
6989 Thanks to Yuval Yarom and Naomi Benger for discovering this
6990 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6992 *Yuval Yarom and Naomi Benger*
6994 * TLS pad extension: draft-agl-tls-padding-03
6996 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6997 TLS client Hello record length value would otherwise be > 255 and
6998 less that 512 pad with a dummy extension containing zeroes so it
6999 is at least 512 bytes long.
7001 *Adam Langley, Steve Henson*
7003 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7005 * Fix for TLS record tampering bug. A carefully crafted invalid
7006 handshake could crash OpenSSL with a NULL pointer exception.
7007 Thanks to Anton Johansson for reporting this issues.
7010 * Keep original DTLS digest and encryption contexts in retransmission
7011 structures so we can use the previous session parameters if they need
7012 to be resent. ([CVE-2013-6450])
7016 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7017 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7018 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7019 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7020 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7021 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7023 *Rob Stradling, Adam Langley*
7025 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7027 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7028 supporting platforms or when small records were transferred.
7030 *Andy Polyakov, Steve Henson*
7032 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7034 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7036 This addresses the flaw in CBC record processing discovered by
7037 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7038 at: <http://www.isg.rhul.ac.uk/tls/>
7040 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7041 Security Group at Royal Holloway, University of London
7042 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7043 Emilia Käsper for the initial patch.
7046 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7048 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7049 ciphersuites which can be exploited in a denial of service attack.
7050 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7051 and detecting this bug and to Wolfgang Ettlinger
7052 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7057 * Return an error when checking OCSP signatures when key is NULL.
7058 This fixes a DoS attack. ([CVE-2013-0166])
7062 * Make openssl verify return errors.
7064 *Chris Palmer <palmer@google.com> and Ben Laurie*
7066 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7067 the right response is stapled. Also change SSL_get_certificate()
7068 so it returns the certificate actually sent.
7069 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7071 *Rob Stradling <rob.stradling@comodo.com>*
7073 * Fix possible deadlock when decoding public keys.
7077 * Don't use TLS 1.0 record version number in initial client hello
7082 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7084 * Sanity check record length before skipping explicit IV in TLS
7085 1.2, 1.1 and DTLS to fix DoS attack.
7087 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7088 fuzzing as a service testing platform.
7093 * Initialise tkeylen properly when encrypting CMS messages.
7094 Thanks to Solar Designer of Openwall for reporting this issue.
7098 * In FIPS mode don't try to use composite ciphers as they are not
7103 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7105 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7106 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7107 mean any application compiled against OpenSSL 1.0.0 headers setting
7108 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7109 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7110 0x10000000L Any application which was previously compiled against
7111 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7112 will need to be recompiled as a result. Letting be results in
7113 inability to disable specifically TLS 1.1 and in client context,
7114 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7118 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7119 disable just protocol X, but all protocols above X *if* there are
7120 protocols *below* X still enabled. In more practical terms it means
7121 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7122 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7123 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7128 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7130 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7131 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7132 in CRYPTO_realloc_clean.
7134 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7135 issue and to Adam Langley <agl@chromium.org> for fixing it.
7138 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7140 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7144 * Workarounds for some broken servers that "hang" if a client hello
7145 record length exceeds 255 bytes.
7147 1. Do not use record version number > TLS 1.0 in initial client
7148 hello: some (but not all) hanging servers will now work.
7149 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7150 the number of ciphers sent in the client hello. This should be
7151 set to an even number, such as 50, for example by passing:
7152 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7153 Most broken servers should now work.
7154 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7155 TLS 1.2 client support entirely.
7159 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7163 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7165 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7166 STRING form instead of a DigestInfo.
7170 * The format used for MDC2 RSA signatures is inconsistent between EVP
7171 and the RSA_sign/RSA_verify functions. This was made more apparent when
7172 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7173 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7174 the correct format in RSA_verify so both forms transparently work.
7178 * Some servers which support TLS 1.0 can choke if we initially indicate
7179 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7180 encrypted premaster secret. As a workaround use the maximum permitted
7181 client version in client hello, this should keep such servers happy
7182 and still work with previous versions of OpenSSL.
7186 * Add support for TLS/DTLS heartbeats.
7188 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7190 * Add support for SCTP.
7192 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7194 * Improved PRNG seeding for VOS.
7196 *Paul Green <Paul.Green@stratus.com>*
7198 * Extensive assembler packs updates, most notably:
7200 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7201 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7202 - x86_64: bit-sliced AES implementation;
7203 - ARM: NEON support, contemporary platforms optimizations;
7204 - s390x: z196 support;
7205 - `*`: GHASH and GF(2^m) multiplication implementations;
7209 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7210 (removal of unnecessary code)
7212 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7214 * Add TLS key material exporter from RFC 5705.
7218 * Add DTLS-SRTP negotiation from RFC 5764.
7222 * Add Next Protocol Negotiation,
7223 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7224 disabled with a no-npn flag to config or Configure. Code donated
7227 *Adam Langley <agl@google.com> and Ben Laurie*
7229 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7230 NIST-P256, NIST-P521, with constant-time single point multiplication on
7231 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7232 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7233 Code made available under Apache License version 2.0.
7235 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7236 line to include this in your build of OpenSSL, and run "make depend" (or
7237 "make update"). This enables the following EC_METHODs:
7239 EC_GFp_nistp224_method()
7240 EC_GFp_nistp256_method()
7241 EC_GFp_nistp521_method()
7243 EC_GROUP_new_by_curve_name() will automatically use these (while
7244 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7247 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7249 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7250 all platforms. Move ssize_t definition from e_os.h to the public
7251 header file e_os2.h as it now appears in public header file cms.h
7255 * New -sigopt option to the ca, req and x509 utilities. Additional
7256 signature parameters can be passed using this option and in
7261 * Add RSA PSS signing function. This will generate and set the
7262 appropriate AlgorithmIdentifiers for PSS based on those in the
7263 corresponding EVP_MD_CTX structure. No application support yet.
7267 * Support for companion algorithm specific ASN1 signing routines.
7268 New function ASN1_item_sign_ctx() signs a pre-initialised
7269 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7270 the appropriate parameters.
7274 * Add new algorithm specific ASN1 verification initialisation function
7275 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7276 handling will be the same no matter what EVP_PKEY_METHOD is used.
7277 Add a PSS handler to support verification of PSS signatures: checked
7278 against a number of sample certificates.
7282 * Add signature printing for PSS. Add PSS OIDs.
7284 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7286 * Add algorithm specific signature printing. An individual ASN1 method
7287 can now print out signatures instead of the standard hex dump.
7289 More complex signatures (e.g. PSS) can print out more meaningful
7290 information. Include DSA version that prints out the signature
7295 * Password based recipient info support for CMS library: implementing
7300 * Split password based encryption into PBES2 and PBKDF2 functions. This
7301 neatly separates the code into cipher and PBE sections and is required
7302 for some algorithms that split PBES2 into separate pieces (such as
7303 password based CMS).
7307 * Session-handling fixes:
7308 - Fix handling of connections that are resuming with a session ID,
7309 but also support Session Tickets.
7310 - Fix a bug that suppressed issuing of a new ticket if the client
7311 presented a ticket with an expired session.
7312 - Try to set the ticket lifetime hint to something reasonable.
7313 - Make tickets shorter by excluding irrelevant information.
7314 - On the client side, don't ignore renewed tickets.
7316 *Adam Langley, Bodo Moeller (Google)*
7318 * Fix PSK session representation.
7322 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7324 This work was sponsored by Intel.
7328 * Add GCM support to TLS library. Some custom code is needed to split
7329 the IV between the fixed (from PRF) and explicit (from TLS record)
7330 portions. This adds all GCM ciphersuites supported by RFC5288 and
7331 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7332 add a special AESGCM string for GCM only.
7336 * Expand range of ctrls for AES GCM. Permit setting invocation
7337 field on decrypt and retrieval of invocation field only on encrypt.
7341 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7342 As required by RFC5289 these ciphersuites cannot be used if for
7343 versions of TLS earlier than 1.2.
7347 * For FIPS capable OpenSSL interpret a NULL default public key method
7348 as unset and return the appropriate default but do *not* set the default.
7349 This means we can return the appropriate method in applications that
7350 switch between FIPS and non-FIPS modes.
7354 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7355 ENGINE is used then we cannot handle that in the FIPS module so we
7356 keep original code iff non-FIPS operations are allowed.
7360 * Add -attime option to openssl utilities.
7362 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7364 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7368 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7369 FIPS EC methods unconditionally for now.
7373 * New build option no-ec2m to disable characteristic 2 code.
7377 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7378 all cases can be covered as some introduce binary incompatibilities.
7382 * Redirect RSA operations to FIPS module including keygen,
7383 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7387 * Add similar low-level API blocking to ciphers.
7391 * low-level digest APIs are not approved in FIPS mode: any attempt
7392 to use these will cause a fatal error. Applications that *really* want
7393 to use them can use the `private_*` version instead.
7397 * Redirect cipher operations to FIPS module for FIPS builds.
7401 * Redirect digest operations to FIPS module for FIPS builds.
7405 * Update build system to add "fips" flag which will link in fipscanister.o
7406 for static and shared library builds embedding a signature if needed.
7410 * Output TLS supported curves in preference order instead of numerical
7411 order. This is currently hardcoded for the highest order curves first.
7412 This should be configurable so applications can judge speed vs strength.
7416 * Add TLS v1.2 server support for client authentication.
7420 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7425 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7426 FIPS modules versions.
7430 * Add TLS v1.2 client side support for client authentication. Keep cache
7431 of handshake records longer as we don't know the hash algorithm to use
7432 until after the certificate request message is received.
7436 * Initial TLS v1.2 client support. Add a default signature algorithms
7437 extension including all the algorithms we support. Parse new signature
7438 format in client key exchange. Relax some ECC signing restrictions for
7439 TLS v1.2 as indicated in RFC5246.
7443 * Add server support for TLS v1.2 signature algorithms extension. Switch
7444 to new signature format when needed using client digest preference.
7445 All server ciphersuites should now work correctly in TLS v1.2. No client
7446 support yet and no support for client certificates.
7450 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7451 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7452 ciphersuites. At present only RSA key exchange ciphersuites work with
7453 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7454 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7455 and version checking.
7459 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7460 with this defined it will not be affected by any changes to ssl internal
7461 structures. Add several utility functions to allow openssl application
7462 to work with OPENSSL_NO_SSL_INTERN defined.
7466 * A long standing patch to add support for SRP from EdelWeb (Peter
7467 Sylvester and Christophe Renou) was integrated.
7468 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7469 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7472 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7476 * Permit abbreviated handshakes when renegotiating using the function
7477 SSL_renegotiate_abbreviated().
7479 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7481 * Add call to ENGINE_register_all_complete() to
7482 ENGINE_load_builtin_engines(), so some implementations get used
7483 automatically instead of needing explicit application support.
7487 * Add support for TLS key exporter as described in RFC5705.
7489 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7491 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7492 a few changes are required:
7494 Add SSL_OP_NO_TLSv1_1 flag.
7495 Add TLSv1_1 methods.
7496 Update version checking logic to handle version 1.1.
7497 Add explicit IV handling (ported from DTLS code).
7498 Add command line options to s_client/s_server.
7505 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7507 * X509_ATTRIBUTE memory leak
7509 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7510 memory. This structure is used by the PKCS#7 and CMS routines so any
7511 application which reads PKCS#7 or CMS data from untrusted sources is
7512 affected. SSL/TLS is not affected.
7514 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7520 * Race condition handling PSK identify hint
7522 If PSK identity hints are received by a multi-threaded client then
7523 the values are wrongly updated in the parent SSL_CTX structure. This can
7524 result in a race condition potentially leading to a double free of the
7530 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7532 * Malformed ECParameters causes infinite loop
7534 When processing an ECParameters structure OpenSSL enters an infinite loop
7535 if the curve specified is over a specially malformed binary polynomial
7538 This can be used to perform denial of service against any
7539 system which processes public keys, certificate requests or
7540 certificates. This includes TLS clients and TLS servers with
7541 client authentication enabled.
7543 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7548 * Exploitable out-of-bounds read in X509_cmp_time
7550 X509_cmp_time does not properly check the length of the ASN1_TIME
7551 string and can read a few bytes out of bounds. In addition,
7552 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7555 An attacker can use this to craft malformed certificates and CRLs of
7556 various sizes and potentially cause a segmentation fault, resulting in
7557 a DoS on applications that verify certificates or CRLs. TLS clients
7558 that verify CRLs are affected. TLS clients and servers with client
7559 authentication enabled may be affected if they use custom verification
7562 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7563 independently by Hanno Böck.
7568 * PKCS7 crash with missing EnvelopedContent
7570 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7571 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7572 with missing content and trigger a NULL pointer dereference on parsing.
7574 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7575 structures from untrusted sources are affected. OpenSSL clients and
7576 servers are not affected.
7578 This issue was reported to OpenSSL by Michal Zalewski (Google).
7583 * CMS verify infinite loop with unknown hash function
7585 When verifying a signedData message the CMS code can enter an infinite loop
7586 if presented with an unknown hash function OID. This can be used to perform
7587 denial of service against any system which verifies signedData messages using
7589 This issue was reported to OpenSSL by Johannes Bauer.
7594 * Race condition handling NewSessionTicket
7596 If a NewSessionTicket is received by a multi-threaded client when attempting to
7597 reuse a previous ticket then a race condition can occur potentially leading to
7598 a double free of the ticket data.
7603 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7605 * Segmentation fault in ASN1_TYPE_cmp fix
7607 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7608 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7609 certificate signature algorithm consistency this can be used to crash any
7610 certificate verification operation and exploited in a DoS attack. Any
7611 application which performs certificate verification is vulnerable including
7612 OpenSSL clients and servers which enable client authentication.
7617 * ASN.1 structure reuse memory corruption fix
7619 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7620 memory corruption via an invalid write. Such reuse is and has been
7621 strongly discouraged and is believed to be rare.
7623 Applications that parse structures containing CHOICE or ANY DEFINED BY
7624 components may be affected. Certificate parsing (d2i_X509 and related
7625 functions) are however not affected. OpenSSL clients and servers are
7631 * PKCS7 NULL pointer dereferences fix
7633 The PKCS#7 parsing code does not handle missing outer ContentInfo
7634 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7635 missing content and trigger a NULL pointer dereference on parsing.
7637 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7638 otherwise parse PKCS#7 structures from untrusted sources are
7639 affected. OpenSSL clients and servers are not affected.
7641 This issue was reported to OpenSSL by Michal Zalewski (Google).
7646 * DoS via reachable assert in SSLv2 servers fix
7648 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7649 servers that both support SSLv2 and enable export cipher suites by sending
7650 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7652 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7653 (OpenSSL development team).
7658 * Use After Free following d2i_ECPrivatekey error fix
7660 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7661 could cause a use after free condition. This, in turn, could cause a double
7662 free in several private key parsing functions (such as d2i_PrivateKey
7663 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7664 for applications that receive EC private keys from untrusted
7665 sources. This scenario is considered rare.
7667 This issue was discovered by the BoringSSL project and fixed in their
7673 * X509_to_X509_REQ NULL pointer deref fix
7675 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7676 the certificate key is invalid. This function is rarely used in practice.
7678 This issue was discovered by Brian Carpenter.
7683 * Removed the export ciphers from the DEFAULT ciphers
7687 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7689 * Build fixes for the Windows and OpenVMS platforms
7691 *Matt Caswell and Richard Levitte*
7693 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7695 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7696 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7697 dereference. This could lead to a Denial Of Service attack. Thanks to
7698 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7703 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7704 dtls1_buffer_record function under certain conditions. In particular this
7705 could occur if an attacker sent repeated DTLS records with the same
7706 sequence number but for the next epoch. The memory leak could be exploited
7707 by an attacker in a Denial of Service attack through memory exhaustion.
7708 Thanks to Chris Mueller for reporting this issue.
7713 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7714 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7715 method would be set to NULL which could later result in a NULL pointer
7716 dereference. Thanks to Frank Schmirler for reporting this issue.
7721 * Abort handshake if server key exchange message is omitted for ephemeral
7724 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7725 reporting this issue.
7730 * Remove non-export ephemeral RSA code on client and server. This code
7731 violated the TLS standard by allowing the use of temporary RSA keys in
7732 non-export ciphersuites and could be used by a server to effectively
7733 downgrade the RSA key length used to a value smaller than the server
7734 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7735 INRIA or reporting this issue.
7740 * Fixed issue where DH client certificates are accepted without verification.
7741 An OpenSSL server will accept a DH certificate for client authentication
7742 without the certificate verify message. This effectively allows a client to
7743 authenticate without the use of a private key. This only affects servers
7744 which trust a client certificate authority which issues certificates
7745 containing DH keys: these are extremely rare and hardly ever encountered.
7746 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7752 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7753 results on some platforms, including x86_64. This bug occurs at random
7754 with a very low probability, and is not known to be exploitable in any
7755 way, though its exact impact is difficult to determine. Thanks to Pieter
7756 Wuille (Blockstream) who reported this issue and also suggested an initial
7757 fix. Further analysis was conducted by the OpenSSL development team and
7758 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7759 the OpenSSL core team.
7764 * Fix various certificate fingerprint issues.
7766 By using non-DER or invalid encodings outside the signed portion of a
7767 certificate the fingerprint can be changed without breaking the signature.
7768 Although no details of the signed portion of the certificate can be changed
7769 this can cause problems with some applications: e.g. those using the
7770 certificate fingerprint for blacklists.
7772 1. Reject signatures with non zero unused bits.
7774 If the BIT STRING containing the signature has non zero unused bits reject
7775 the signature. All current signature algorithms require zero unused bits.
7777 2. Check certificate algorithm consistency.
7779 Check the AlgorithmIdentifier inside TBS matches the one in the
7780 certificate signature. NB: this will result in signature failure
7781 errors for some broken certificates.
7783 Thanks to Konrad Kraszewski from Google for reporting this issue.
7785 3. Check DSA/ECDSA signatures use DER.
7787 Reencode DSA/ECDSA signatures and compare with the original received
7788 signature. Return an error if there is a mismatch.
7790 This will reject various cases including garbage after signature
7791 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7792 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7793 (negative or with leading zeroes).
7795 Further analysis was conducted and fixes were developed by Stephen Henson
7796 of the OpenSSL core team.
7802 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7804 * Session Ticket Memory Leak.
7806 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7807 integrity of that ticket is first verified. In the event of a session
7808 ticket integrity check failing, OpenSSL will fail to free memory
7809 causing a memory leak. By sending a large number of invalid session
7810 tickets an attacker could exploit this issue in a Denial Of Service
7816 * Build option no-ssl3 is incomplete.
7818 When OpenSSL is configured with "no-ssl3" as a build option, servers
7819 could accept and complete a SSL 3.0 handshake, and clients could be
7820 configured to send them.
7823 *Akamai and the OpenSSL team*
7825 * Add support for TLS_FALLBACK_SCSV.
7826 Client applications doing fallback retries should call
7827 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7830 *Adam Langley, Bodo Moeller*
7832 * Add additional DigestInfo checks.
7834 Reencode DigestInto in DER and check against the original when
7835 verifying RSA signature: this will reject any improperly encoded
7836 DigestInfo structures.
7838 Note: this is a precautionary measure and no attacks are currently known.
7842 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7844 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7845 to a denial of service attack. A malicious server can crash the client
7846 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7847 ciphersuite and sending carefully crafted handshake messages.
7849 Thanks to Felix Gröbert (Google) for discovering and researching this
7855 * By sending carefully crafted DTLS packets an attacker could cause openssl
7856 to leak memory. This can be exploited through a Denial of Service attack.
7857 Thanks to Adam Langley for discovering and researching this issue.
7862 * An attacker can force openssl to consume large amounts of memory whilst
7863 processing DTLS handshake messages. This can be exploited through a
7864 Denial of Service attack.
7865 Thanks to Adam Langley for discovering and researching this issue.
7870 * An attacker can force an error condition which causes openssl to crash
7871 whilst processing DTLS packets due to memory being freed twice. This
7872 can be exploited through a Denial of Service attack.
7873 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7879 * If a multithreaded client connects to a malicious server using a resumed
7880 session and the server sends an ec point format extension it could write
7881 up to 255 bytes to freed memory.
7883 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7889 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7890 X509_name_oneline, X509_name_print_ex et al. to leak some information
7891 from the stack. Applications may be affected if they echo pretty printing
7892 output to the attacker.
7894 Thanks to Ivan Fratric (Google) for discovering this issue.
7897 *Emilia Käsper, and Steve Henson*
7899 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7900 for corner cases. (Certain input points at infinity could lead to
7901 bogus results, with non-infinity inputs mapped to infinity too.)
7905 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7907 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7908 handshake can force the use of weak keying material in OpenSSL
7909 SSL/TLS clients and servers.
7911 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7912 researching this issue. ([CVE-2014-0224])
7914 *KIKUCHI Masashi, Steve Henson*
7916 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7917 OpenSSL DTLS client the code can be made to recurse eventually crashing
7920 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7923 *Imre Rad, Steve Henson*
7925 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7926 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7927 client or server. This is potentially exploitable to run arbitrary
7928 code on a vulnerable client or server.
7930 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7932 *Jüri Aedla, Steve Henson*
7934 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7935 are subject to a denial of service attack.
7937 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7938 this issue. ([CVE-2014-3470])
7940 *Felix Gröbert, Ivan Fratric, Steve Henson*
7942 * Harmonize version and its documentation. -f flag is used to display
7945 *mancha <mancha1@zoho.com>*
7947 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7948 in i2d_ECPrivateKey.
7950 *mancha <mancha1@zoho.com>*
7952 * Fix some double frees. These are not thought to be exploitable.
7954 *mancha <mancha1@zoho.com>*
7956 * Fix for the attack described in the paper "Recovering OpenSSL
7957 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7958 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7959 <http://eprint.iacr.org/2014/140>
7961 Thanks to Yuval Yarom and Naomi Benger for discovering this
7962 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7964 *Yuval Yarom and Naomi Benger*
7966 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7968 * Keep original DTLS digest and encryption contexts in retransmission
7969 structures so we can use the previous session parameters if they need
7970 to be resent. ([CVE-2013-6450])
7974 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7975 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7976 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7977 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7978 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7979 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7981 *Rob Stradling, Adam Langley*
7983 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7985 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7987 This addresses the flaw in CBC record processing discovered by
7988 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7989 at: <http://www.isg.rhul.ac.uk/tls/>
7991 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7992 Security Group at Royal Holloway, University of London
7993 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7994 Emilia Käsper for the initial patch.
7997 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7999 * Return an error when checking OCSP signatures when key is NULL.
8000 This fixes a DoS attack. ([CVE-2013-0166])
8004 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8005 the right response is stapled. Also change SSL_get_certificate()
8006 so it returns the certificate actually sent.
8007 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8008 (This is a backport)
8010 *Rob Stradling <rob.stradling@comodo.com>*
8012 * Fix possible deadlock when decoding public keys.
8016 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8018 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8021 * Sanity check record length before skipping explicit IV in DTLS
8024 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8025 fuzzing as a service testing platform.
8030 * Initialise tkeylen properly when encrypting CMS messages.
8031 Thanks to Solar Designer of Openwall for reporting this issue.
8035 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8037 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8038 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8039 in CRYPTO_realloc_clean.
8041 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8042 issue and to Adam Langley <agl@chromium.org> for fixing it.
8045 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8047 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8049 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8050 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8051 content decryption and always return the same error. Note: this attack
8052 needs on average 2^20 messages so it only affects automated senders. The
8053 old behaviour can be re-enabled in the CMS code by setting the
8054 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8055 an MMA defence is not necessary.
8056 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8057 this issue. ([CVE-2012-0884])
8061 * Fix CVE-2011-4619: make sure we really are receiving a
8062 client hello before rejecting multiple SGC restarts. Thanks to
8063 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8067 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8069 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8070 Thanks to Antonio Martin, Enterprise Secure Access Research and
8071 Development, Cisco Systems, Inc. for discovering this bug and
8072 preparing a fix. ([CVE-2012-0050])
8076 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8078 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8079 of the Vaudenay padding oracle attack on CBC mode encryption
8080 which enables an efficient plaintext recovery attack against
8081 the OpenSSL implementation of DTLS. Their attack exploits timing
8082 differences arising during decryption processing. A research
8083 paper describing this attack can be found at:
8084 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8085 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8086 Security Group at Royal Holloway, University of London
8087 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8088 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8089 for preparing the fix. ([CVE-2011-4108])
8091 *Robin Seggelmann, Michael Tuexen*
8093 * Clear bytes used for block padding of SSL 3.0 records.
8096 *Adam Langley (Google)*
8098 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8099 Kadianakis <desnacked@gmail.com> for discovering this issue and
8100 Adam Langley for preparing the fix. ([CVE-2011-4619])
8102 *Adam Langley (Google)*
8104 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8106 *Andrey Kulikov <amdeich@gmail.com>*
8108 * Prevent malformed RFC3779 data triggering an assertion failure.
8109 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8110 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8112 *Rob Austein <sra@hactrn.net>*
8114 * Improved PRNG seeding for VOS.
8116 *Paul Green <Paul.Green@stratus.com>*
8118 * Fix ssl_ciph.c set-up race.
8120 *Adam Langley (Google)*
8122 * Fix spurious failures in ecdsatest.c.
8124 *Emilia Käsper (Google)*
8126 * Fix the BIO_f_buffer() implementation (which was mixing different
8127 interpretations of the `..._len` fields).
8129 *Adam Langley (Google)*
8131 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8132 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8133 threads won't reuse the same blinding coefficients.
8135 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8136 lock to call BN_BLINDING_invert_ex, and avoids one use of
8137 BN_BLINDING_update for each BN_BLINDING structure (previously,
8138 the last update always remained unused).
8140 *Emilia Käsper (Google)*
8142 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8144 *Bob Buckholz (Google)*
8146 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8148 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8149 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8151 *Kaspar Brand <ossl@velox.ch>*
8153 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8154 for multi-threaded use of ECDH. ([CVE-2011-3210])
8156 *Adam Langley (Google)*
8158 * Fix x509_name_ex_d2i memory leak on bad inputs.
8162 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8163 signature public key algorithm by using OID xref utilities instead.
8164 Before this you could only use some ECC ciphersuites with SHA1 only.
8168 * Add protection against ECDSA timing attacks as mentioned in the paper
8169 by Billy Bob Brumley and Nicola Tuveri, see:
8170 <http://eprint.iacr.org/2011/232.pdf>
8172 *Billy Bob Brumley and Nicola Tuveri*
8174 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8176 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8178 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8180 * Fix bug in string printing code: if *any* escaping is enabled we must
8181 escape the escape character (backslash) or the resulting string is
8186 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8188 * Disable code workaround for ancient and obsolete Netscape browsers
8189 and servers: an attacker can use it in a ciphersuite downgrade attack.
8190 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8194 * Fixed J-PAKE implementation error, originally discovered by
8195 Sebastien Martini, further info and confirmation from Stefan
8196 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8200 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8202 * Fix extension code to avoid race conditions which can result in a buffer
8203 overrun vulnerability: resumed sessions must not be modified as they can
8204 be shared by multiple threads. CVE-2010-3864
8208 * Fix WIN32 build system to correctly link an ENGINE directory into
8213 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8215 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8218 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8220 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8222 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8223 context. The operation can be customised via the ctrl mechanism in
8224 case ENGINEs want to include additional functionality.
8228 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8232 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8233 output hashes compatible with older versions of OpenSSL.
8235 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8237 * Fix compression algorithm handling: if resuming a session use the
8238 compression algorithm of the resumed session instead of determining
8239 it from client hello again. Don't allow server to change algorithm.
8243 * Add load_crls() function to commands tidying load_certs() too. Add option
8244 to verify utility to allow additional CRLs to be included.
8248 * Update OCSP request code to permit adding custom headers to the request:
8249 some responders need this.
8253 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8256 *Julia Lawall <julia@diku.dk>*
8258 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8259 needlessly dereferenced structures, used obsolete functions and
8260 didn't handle all updated verify codes correctly.
8264 * Disable MD2 in the default configuration.
8268 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8269 indicate the initial BIO being pushed or popped. This makes it possible
8270 to determine whether the BIO is the one explicitly called or as a result
8271 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8272 it handles reference counts correctly and doesn't zero out the I/O bio
8273 when it is not being explicitly popped. WARNING: applications which
8274 included workarounds for the old buggy behaviour will need to be modified
8275 or they could free up already freed BIOs.
8279 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8280 renaming to all platforms (within the 0.9.8 branch, this was
8281 done conditionally on Netware platforms to avoid a name clash).
8283 *Guenter <lists@gknw.net>*
8285 * Add ECDHE and PSK support to DTLS.
8287 *Michael Tuexen <tuexen@fh-muenster.de>*
8289 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8294 * Add "missing" function EVP_MD_flags() (without this the only way to
8295 retrieve a digest flags is by accessing the structure directly. Update
8296 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8297 or cipher is registered as in the "from" argument. Print out all
8298 registered digests in the dgst usage message instead of manually
8299 attempting to work them out.
8303 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8304 this allows the use of compression and extensions. Change default cipher
8305 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8306 by default unless an application cipher string requests it.
8310 * Alter match criteria in PKCS12_parse(). It used to try to use local
8311 key ids to find matching certificates and keys but some PKCS#12 files
8312 don't follow the (somewhat unwritten) rules and this strategy fails.
8313 Now just gather all certificates together and the first private key
8314 then look for the first certificate that matches the key.
8318 * Support use of registered digest and cipher names for dgst and cipher
8319 commands instead of having to add each one as a special case. So now
8326 openssl dgst -sha256 foo
8328 and this works for ENGINE based algorithms too.
8332 * Update Gost ENGINE to support parameter files.
8334 *Victor B. Wagner <vitus@cryptocom.ru>*
8336 * Support GeneralizedTime in ca utility.
8338 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8340 * Enhance the hash format used for certificate directory links. The new
8341 form uses the canonical encoding (meaning equivalent names will work
8342 even if they aren't identical) and uses SHA1 instead of MD5. This form
8343 is incompatible with the older format and as a result c_rehash should
8344 be used to rebuild symbolic links.
8348 * Make PKCS#8 the default write format for private keys, replacing the
8349 traditional format. This form is standardised, more secure and doesn't
8350 include an implicit MD5 dependency.
8354 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8355 committed to OpenSSL should pass this lot as a minimum.
8359 * Add session ticket override functionality for use by EAP-FAST.
8361 *Jouni Malinen <j@w1.fi>*
8363 * Modify HMAC functions to return a value. Since these can be implemented
8364 in an ENGINE errors can occur.
8368 * Type-checked OBJ_bsearch_ex.
8372 * Type-checked OBJ_bsearch. Also some constification necessitated
8373 by type-checking. Still to come: TXT_DB, bsearch(?),
8374 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8379 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8380 seconds to a tm structure directly, instead of going through OS
8381 specific date routines. This avoids any issues with OS routines such
8382 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8383 and X509_time_adj_ex() to cover the extended range. The existing
8384 X509_time_adj() is still usable and will no longer have any date issues.
8388 * Delta CRL support. New use deltas option which will attempt to locate
8389 and search any appropriate delta CRLs available.
8391 This work was sponsored by Google.
8395 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8396 code and add additional score elements. Validate alternate CRL paths
8397 as part of the CRL checking and indicate a new error "CRL path validation
8398 error" in this case. Applications wanting additional details can use
8399 the verify callback and check the new "parent" field. If this is not
8400 NULL CRL path validation is taking place. Existing applications won't
8401 see this because it requires extended CRL support which is off by
8404 This work was sponsored by Google.
8408 * Support for freshest CRL extension.
8410 This work was sponsored by Google.
8414 * Initial indirect CRL support. Currently only supported in the CRLs
8415 passed directly and not via lookup. Process certificate issuer
8416 CRL entry extension and lookup CRL entries by bother issuer name
8417 and serial number. Check and process CRL issuer entry in IDP extension.
8419 This work was sponsored by Google.
8423 * Add support for distinct certificate and CRL paths. The CRL issuer
8424 certificate is validated separately in this case. Only enabled if
8425 an extended CRL support flag is set: this flag will enable additional
8426 CRL functionality in future.
8428 This work was sponsored by Google.
8432 * Add support for policy mappings extension.
8434 This work was sponsored by Google.
8438 * Fixes to pathlength constraint, self issued certificate handling,
8439 policy processing to align with RFC3280 and PKITS tests.
8441 This work was sponsored by Google.
8445 * Support for name constraints certificate extension. DN, email, DNS
8446 and URI types are currently supported.
8448 This work was sponsored by Google.
8452 * To cater for systems that provide a pointer-based thread ID rather
8453 than numeric, deprecate the current numeric thread ID mechanism and
8454 replace it with a structure and associated callback type. This
8455 mechanism allows a numeric "hash" to be extracted from a thread ID in
8456 either case, and on platforms where pointers are larger than 'long',
8457 mixing is done to help ensure the numeric 'hash' is usable even if it
8458 can't be guaranteed unique. The default mechanism is to use "&errno"
8459 as a pointer-based thread ID to distinguish between threads.
8461 Applications that want to provide their own thread IDs should now use
8462 CRYPTO_THREADID_set_callback() to register a callback that will call
8463 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8465 Note that ERR_remove_state() is now deprecated, because it is tied
8466 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8467 to free the current thread's error state should be replaced by
8468 ERR_remove_thread_state(NULL).
8470 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8471 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8472 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8473 application was previously providing a numeric thread callback that
8474 was inappropriate for distinguishing threads, then uniqueness might
8475 have been obtained with &errno that happened immediately in the
8476 intermediate development versions of OpenSSL; this is no longer the
8477 case, the numeric thread callback will now override the automatic use
8480 *Geoff Thorpe, with help from Bodo Moeller*
8482 * Initial support for different CRL issuing certificates. This covers a
8483 simple case where the self issued certificates in the chain exist and
8484 the real CRL issuer is higher in the existing chain.
8486 This work was sponsored by Google.
8490 * Removed effectively defunct crypto/store from the build.
8494 * Revamp of STACK to provide stronger type-checking. Still to come:
8495 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8496 ASN1_STRING, CONF_VALUE.
8500 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8501 RAM on SSL connections. This option can save about 34k per idle SSL.
8505 * Revamp of LHASH to provide stronger type-checking. Still to come:
8506 STACK, TXT_DB, bsearch, qsort.
8510 * Initial support for Cryptographic Message Syntax (aka CMS) based
8511 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8512 support for data, signedData, compressedData, digestedData and
8513 encryptedData, envelopedData types included. Scripts to check against
8514 RFC4134 examples draft and interop and consistency checks of many
8515 content types and variants.
8519 * Add options to enc utility to support use of zlib compression BIO.
8523 * Extend mk1mf to support importing of options and assembly language
8524 files from Configure script, currently only included in VC-WIN32.
8525 The assembly language rules can now optionally generate the source
8526 files from the associated perl scripts.
8530 * Implement remaining functionality needed to support GOST ciphersuites.
8531 Interop testing has been performed using CryptoPro implementations.
8533 *Victor B. Wagner <vitus@cryptocom.ru>*
8535 * s390x assembler pack.
8539 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8544 * Implement Opaque PRF Input TLS extension as specified in
8545 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8546 official specification yet and no extension type assignment by
8547 IANA exists, this extension (for now) will have to be explicitly
8548 enabled when building OpenSSL by providing the extension number
8549 to use. For example, specify an option
8551 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8553 to the "config" or "Configure" script to enable the extension,
8554 assuming extension number 0x9527 (which is a completely arbitrary
8555 and unofficial assignment based on the MD5 hash of the Internet
8556 Draft). Note that by doing so, you potentially lose
8557 interoperability with other TLS implementations since these might
8558 be using the same extension number for other purposes.
8560 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8561 opaque PRF input value to use in the handshake. This will create
8562 an internal copy of the length-'len' string at 'src', and will
8563 return non-zero for success.
8565 To get more control and flexibility, provide a callback function
8568 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8569 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8573 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8576 Callback function 'cb' will be called in handshakes, and is
8577 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8578 Argument 'arg' is for application purposes (the value as given to
8579 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8580 be provided to the callback function). The callback function
8581 has to return non-zero to report success: usually 1 to use opaque
8582 PRF input just if possible, or 2 to enforce use of the opaque PRF
8583 input. In the latter case, the library will abort the handshake
8584 if opaque PRF input is not successfully negotiated.
8586 Arguments 'peerinput' and 'len' given to the callback function
8587 will always be NULL and 0 in the case of a client. A server will
8588 see the client's opaque PRF input through these variables if
8589 available (NULL and 0 otherwise). Note that if the server
8590 provides an opaque PRF input, the length must be the same as the
8591 length of the client's opaque PRF input.
8593 Note that the callback function will only be called when creating
8594 a new session (session resumption can resume whatever was
8595 previously negotiated), and will not be called in SSL 2.0
8596 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8597 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8598 for applications that need to enforce opaque PRF input.
8602 * Update ssl code to support digests other than SHA1+MD5 for handshake
8605 *Victor B. Wagner <vitus@cryptocom.ru>*
8607 * Add RFC4507 support to OpenSSL. This includes the corrections in
8608 RFC4507bis. The encrypted ticket format is an encrypted encoded
8609 SSL_SESSION structure, that way new session features are automatically
8612 If a client application caches session in an SSL_SESSION structure
8613 support is transparent because tickets are now stored in the encoded
8616 The SSL_CTX structure automatically generates keys for ticket
8617 protection in servers so again support should be possible
8618 with no application modification.
8620 If a client or server wishes to disable RFC4507 support then the option
8621 SSL_OP_NO_TICKET can be set.
8623 Add a TLS extension debugging callback to allow the contents of any client
8624 or server extensions to be examined.
8626 This work was sponsored by Google.
8630 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8631 OpenSSL should now compile cleanly on gcc 4.2
8633 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8635 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8636 support including streaming MAC support: this is required for GOST
8637 ciphersuite support.
8639 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8641 * Add option -stream to use PKCS#7 streaming in smime utility. New
8642 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8643 to output in BER and PEM format.
8647 * Experimental support for use of HMAC via EVP_PKEY interface. This
8648 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8649 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8650 ENGINE support for HMAC keys which are unextractable. New -mac and
8651 -macopt options to dgst utility.
8655 * New option -sigopt to dgst utility. Update dgst to use
8656 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8657 alternative signing parameters such as X9.31 or PSS in the dgst
8662 * Change ssl_cipher_apply_rule(), the internal function that does
8663 the work each time a ciphersuite string requests enabling
8664 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8665 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8666 the order of disabled ciphersuites such that those ciphersuites
8667 that most recently went from enabled to disabled not only stay
8668 in order with respect to each other, but also have higher priority
8669 than other disabled ciphersuites the next time ciphersuites are
8672 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8673 the same ciphersuites as with "HIGH" alone, but in a specific
8674 order where the PSK ciphersuites come first (since they are the
8675 most recently disabled ciphersuites when "HIGH" is parsed).
8677 Also, change ssl_create_cipher_list() (using this new
8678 functionality) such that between otherwise identical
8679 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8684 * Change ssl_create_cipher_list() so that it automatically
8685 arranges the ciphersuites in reasonable order before starting
8686 to process the rule string. Thus, the definition for "DEFAULT"
8687 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8688 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8689 This makes it much easier to arrive at a reasonable default order
8690 in applications for which anonymous ciphers are OK (meaning
8691 that you can't actually use DEFAULT).
8693 *Bodo Moeller; suggested by Victor Duchovni*
8695 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8696 processing) into multiple integers instead of setting
8697 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8698 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8699 (These masks as well as the individual bit definitions are hidden
8700 away into the non-exported interface ssl/ssl_locl.h, so this
8701 change to the definition of the SSL_CIPHER structure shouldn't
8702 affect applications.) This give us more bits for each of these
8703 categories, so there is no longer a need to coagulate AES128 and
8704 AES256 into a single algorithm bit, and to coagulate Camellia128
8705 and Camellia256 into a single algorithm bit, which has led to all
8708 Thus, among other things, the kludge introduced in 0.9.7m and
8709 0.9.8e for masking out AES256 independently of AES128 or masking
8710 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8712 With the change, we also introduce new ciphersuite aliases that
8713 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8718 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8719 Use the leftmost N bytes of the signature input if the input is
8720 larger than the prime q (with N being the size in bytes of q).
8724 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8725 it yet and it is largely untested.
8729 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8733 * Initial incomplete changes to avoid need for function casts in OpenSSL
8734 some compilers (gcc 4.2 and later) reject their use. Safestack is
8735 reimplemented. Update ASN1 to avoid use of legacy functions.
8739 * Win32/64 targets are linked with Winsock2.
8743 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8744 to external functions. This can be used to increase CRL handling
8745 efficiency especially when CRLs are very large by (for example) storing
8746 the CRL revoked certificates in a database.
8750 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8751 new CRLs added to a directory can be used. New command line option
8752 -verify_return_error to s_client and s_server. This causes real errors
8753 to be returned by the verify callback instead of carrying on no matter
8754 what. This reflects the way a "real world" verify callback would behave.
8758 * GOST engine, supporting several GOST algorithms and public key formats.
8759 Kindly donated by Cryptocom.
8763 * Partial support for Issuing Distribution Point CRL extension. CRLs
8764 partitioned by DP are handled but no indirect CRL or reason partitioning
8765 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8766 selected via a scoring technique which handles IDP and AKID in CRLs.
8770 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8771 will ultimately be used for all verify operations: this will remove the
8772 X509_STORE dependency on certificate verification and allow alternative
8773 lookup methods. X509_STORE based implementations of these two callbacks.
8777 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8778 Modify get_crl() to find a valid (unexpired) CRL if possible.
8782 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8783 this would be called X509_CRL_cmp() but that name is already used by
8784 a function that just compares CRL issuer names. Cache several CRL
8785 extensions in X509_CRL structure and cache CRLDP in X509.
8789 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8790 this maps equivalent X509_NAME structures into a consistent structure.
8791 Name comparison can then be performed rapidly using memcmp().
8795 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8800 * Allow digests to supply their own micalg string for S/MIME type using
8801 the ctrl EVP_MD_CTRL_MICALG.
8805 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8806 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8807 ctrl. It can then customise the structure before and/or after signing
8812 * New function OBJ_add_sigid() to allow application defined signature OIDs
8813 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8814 to free up any added signature OIDs.
8818 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8819 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8820 digest and cipher tables. New options added to openssl utility:
8821 list-message-digest-algorithms and list-cipher-algorithms.
8825 * Change the array representation of binary polynomials: the list
8826 of degrees of non-zero coefficients is now terminated with -1.
8827 Previously it was terminated with 0, which was also part of the
8828 value; thus, the array representation was not applicable to
8829 polynomials where t^0 has coefficient zero. This change makes
8830 the array representation useful in a more general context.
8834 * Various modifications and fixes to SSL/TLS cipher string
8835 handling. For ECC, the code now distinguishes between fixed ECDH
8836 with RSA certificates on the one hand and with ECDSA certificates
8837 on the other hand, since these are separate ciphersuites. The
8838 unused code for Fortezza ciphersuites has been removed.
8840 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8841 (not "ECDHE"). For consistency with the code for DH
8842 certificates, use of ECDH certificates is now considered ECDH
8843 authentication, not RSA or ECDSA authentication (the latter is
8844 merely the CA's signing algorithm and not actively used in the
8847 The temporary ciphersuite alias "ECCdraft" is no longer
8848 available, and ECC ciphersuites are no longer excluded from "ALL"
8849 and "DEFAULT". The following aliases now exist for RFC 4492
8850 ciphersuites, most of these by analogy with the DH case:
8852 kECDHr - ECDH cert, signed with RSA
8853 kECDHe - ECDH cert, signed with ECDSA
8854 kECDH - ECDH cert (signed with either RSA or ECDSA)
8855 kEECDH - ephemeral ECDH
8856 ECDH - ECDH cert or ephemeral ECDH
8862 AECDH - anonymous ECDH
8863 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8867 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8868 Use correct micalg parameters depending on digest(s) in signed message.
8872 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8873 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8877 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8878 an engine to register a method. Add ENGINE lookups for methods and
8879 functional reference processing.
8883 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8884 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8889 * New -resign option to smime utility. This adds one or more signers
8890 to an existing PKCS#7 signedData structure. Also -md option to use an
8891 alternative message digest algorithm for signing.
8895 * Tidy up PKCS#7 routines and add new functions to make it easier to
8896 create PKCS7 structures containing multiple signers. Update smime
8897 application to support multiple signers.
8901 * New -macalg option to pkcs12 utility to allow setting of an alternative
8906 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8907 Reorganize PBE internals to lookup from a static table using NIDs,
8908 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8909 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8910 PRF which will be automatically used with PBES2.
8914 * Replace the algorithm specific calls to generate keys in "req" with the
8919 * Update PKCS#7 enveloped data routines to use new API. This is now
8920 supported by any public key method supporting the encrypt operation. A
8921 ctrl is added to allow the public key algorithm to examine or modify
8922 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8927 * Add a ctrl to asn1 method to allow a public key algorithm to express
8928 a default digest type to use. In most cases this will be SHA1 but some
8929 algorithms (such as GOST) need to specify an alternative digest. The
8930 return value indicates how strong the preference is 1 means optional and
8931 2 is mandatory (that is it is the only supported type). Modify
8932 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8933 use the default md. Update openssl utilities to use the default digest
8934 type for signing if it is not explicitly indicated.
8938 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8939 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8940 signing method from the key type. This effectively removes the link
8941 between digests and public key types.
8945 * Add an OID cross reference table and utility functions. Its purpose is to
8946 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8947 rsaEncryption. This will allow some of the algorithm specific hackery
8948 needed to use the correct OID to be removed.
8952 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8953 structures for PKCS7_sign(). They are now set up by the relevant public
8958 * Add provisional EC pkey method with support for ECDSA and ECDH.
8962 * Add support for key derivation (agreement) in the API, DH method and
8967 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8968 public and private key formats. As a side effect these add additional
8969 command line functionality not previously available: DSA signatures can be
8970 generated and verified using pkeyutl and DH key support and generation in
8977 *Oliver Tappe <zooey@hirschkaefer.de>*
8979 * New make target "install_html_docs" installs HTML renditions of the
8982 *Oliver Tappe <zooey@hirschkaefer.de>*
8984 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8985 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8986 support key and parameter generation and add initial key generation
8987 functionality for RSA.
8991 * Add functions for main EVP_PKEY_method operations. The undocumented
8992 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8993 `EVP_PKEY_{encrypt,decrypt}_old`.
8997 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8998 key API, doesn't do much yet.
9002 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9003 public key algorithms. New option to openssl utility:
9004 "list-public-key-algorithms" to print out info.
9008 * Implement the Supported Elliptic Curves Extension for
9009 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9013 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9014 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9018 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9019 utilities such as rsa, dsa, dsaparam etc except they process any key
9024 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9025 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9026 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9031 * Initial support for pluggable public key ASN1.
9032 De-spaghettify the public key ASN1 handling. Move public and private
9033 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9034 algorithm specific handling to a single module within the relevant
9035 algorithm directory. Add functions to allow (near) opaque processing
9036 of public and private key structures.
9040 * Implement the Supported Point Formats Extension for
9041 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9045 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9046 for the psk identity [hint] and the psk callback functions to the
9047 SSL_SESSION, SSL and SSL_CTX structure.
9050 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9054 SSL_CTX_use_psk_identity_hint
9055 SSL_get_psk_identity_hint
9056 SSL_get_psk_identity
9057 SSL_use_psk_identity_hint
9059 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9061 * Add RFC 3161 compliant time stamp request creation, response generation
9062 and response verification functionality.
9064 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9066 * Add initial support for TLS extensions, specifically for the server_name
9067 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9068 have new members for a host name. The SSL data structure has an
9069 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9070 stored in that context to allow for session resumption, even after the
9071 SSL has been switched to a new SSL_CTX in reaction to a client's
9072 server_name extension.
9074 New functions (subject to change):
9076 SSL_get_servername()
9077 SSL_get_servername_type()
9080 New CTRL codes and macros (subject to change):
9082 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9083 - SSL_CTX_set_tlsext_servername_callback()
9084 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9085 - SSL_CTX_set_tlsext_servername_arg()
9086 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9088 openssl s_client has a new '-servername ...' option.
9090 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9091 '-key2 ...', '-servername_fatal' (subject to change). This allows
9092 testing the HostName extension for a specific single host name ('-cert'
9093 and '-key' remain fallbacks for handshakes without HostName
9094 negotiation). If the unrecognized_name alert has to be sent, this by
9095 default is a warning; it becomes fatal with the '-servername_fatal'
9098 *Peter Sylvester, Remy Allais, Christophe Renou*
9100 * Whirlpool hash implementation is added.
9104 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9105 bn(64,32). Because of instruction set limitations it doesn't have
9106 any negative impact on performance. This was done mostly in order
9107 to make it possible to share assembler modules, such as bn_mul_mont
9108 implementations, between 32- and 64-bit builds without hassle.
9112 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9113 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9118 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9119 dedicated Montgomery multiplication procedure, is introduced.
9120 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9121 "64-bit" performance on certain 32-bit targets.
9125 * New option SSL_OP_NO_COMP to disable use of compression selectively
9126 in SSL structures. New SSL ctrl to set maximum send fragment size.
9127 Save memory by setting the I/O buffer sizes dynamically instead of
9128 using the maximum available value.
9132 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9133 in addition to the text details.
9137 * Very, very preliminary EXPERIMENTAL support for printing of general
9138 ASN1 structures. This currently produces rather ugly output and doesn't
9139 handle several customised structures at all.
9143 * Integrated support for PVK file format and some related formats such
9144 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9145 these in the 'rsa' and 'dsa' utilities.
9149 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9153 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9154 place for the (very old) "NETSCAPE" format certificates which are now
9155 handled using new ASN1 code equivalents.
9159 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9160 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9161 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9165 * Modify CRL distribution points extension code to print out previously
9166 unsupported fields. Enhance extension setting code to allow setting of
9171 * Add print and set support for Issuing Distribution Point CRL extension.
9175 * Change 'Configure' script to enable Camellia by default.
9182 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9184 * When rejecting SSL/TLS records due to an incorrect version number, never
9185 update s->server with a new major version number. As of
9186 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9187 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9188 the previous behavior could result in a read attempt at NULL when
9189 receiving specific incorrect SSL/TLS records once record payload
9190 protection is active. ([CVE-2010-0740])
9192 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9194 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9195 could be crashed if the relevant tables were not present (e.g. chrooted).
9197 *Tomas Hoger <thoger@redhat.com>*
9199 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9201 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9203 *Martin Olsson, Neel Mehta*
9205 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9206 accommodate for stack sorting, always a write lock!).
9210 * On some versions of WIN32 Heap32Next is very slow. This can cause
9211 excessive delays in the RAND_poll(): over a minute. As a workaround
9212 include a time check in the inner Heap32Next loop too.
9216 * The code that handled flushing of data in SSL/TLS originally used the
9217 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9218 the problem outlined in PR#1949. The fix suggested there however can
9219 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9220 of Apache). So instead simplify the code to flush unconditionally.
9221 This should be fine since flushing with no data to flush is a no op.
9225 * Handle TLS versions 2.0 and later properly and correctly use the
9226 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9227 off ancient servers have a habit of sticking around for a while...
9231 * Modify compression code so it frees up structures without using the
9232 ex_data callbacks. This works around a problem where some applications
9233 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9234 restarting) then use compression (e.g. SSL with compression) later.
9235 This results in significant per-connection memory leaks and
9236 has caused some security issues including CVE-2008-1678 and
9241 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9242 change when encrypting or decrypting.
9246 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9247 connect and renegotiate with servers which do not support RI.
9248 Until RI is more widely deployed this option is enabled by default.
9252 * Add "missing" ssl ctrls to clear options and mode.
9256 * If client attempts to renegotiate and doesn't support RI respond with
9257 a no_renegotiation alert as required by RFC5746. Some renegotiating
9258 TLS clients will continue a connection gracefully when they receive
9259 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9260 waiting for a server hello which it will never receive. Now we treat a
9261 received no_renegotiation alert as a fatal error. This is because
9262 applications requesting a renegotiation might well expect it to succeed
9263 and would have no code in place to handle the server denying it so the
9264 only safe thing to do is to terminate the connection.
9268 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9269 peer supports secure renegotiation and 0 otherwise. Print out peer
9270 renegotiation support in s_client/s_server.
9274 * Replace the highly broken and deprecated SPKAC certification method with
9275 the updated NID creation version. This should correctly handle UTF8.
9279 * Implement RFC5746. Re-enable renegotiation but require the extension
9280 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9281 turns out to be a bad idea. It has been replaced by
9282 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9283 SSL_CTX_set_options(). This is really not recommended unless you
9284 know what you are doing.
9286 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9288 * Fixes to stateless session resumption handling. Use initial_ctx when
9289 issuing and attempting to decrypt tickets in case it has changed during
9290 servername handling. Use a non-zero length session ID when attempting
9291 stateless session resumption: this makes it possible to determine if
9292 a resumption has occurred immediately after receiving server hello
9293 (several places in OpenSSL subtly assume this) instead of later in
9298 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9299 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9300 fixes for a few places where the return code is not checked
9303 *Julia Lawall <julia@diku.dk>*
9305 * Add --strict-warnings option to Configure script to include devteam
9306 warnings in other configurations.
9310 * Add support for --libdir option and LIBDIR variable in makefiles. This
9311 makes it possible to install openssl libraries in locations which
9312 have names other than "lib", for example "/usr/lib64" which some
9315 *Steve Henson, based on patch from Jeremy Utley*
9317 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9318 X690 8.9.12 and can produce some misleading textual output of OIDs.
9320 *Steve Henson, reported by Dan Kaminsky*
9322 * Delete MD2 from algorithm tables. This follows the recommendation in
9323 several standards that it is not used in new applications due to
9324 several cryptographic weaknesses. For binary compatibility reasons
9325 the MD2 API is still compiled in by default.
9329 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9334 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9335 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9338 *Guenter <lists@gknw.net>*
9340 * Fix the server certificate chain building code to use X509_verify_cert(),
9341 it used to have an ad-hoc builder which was unable to cope with anything
9342 other than a simple chain.
9344 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9346 * Don't check self signed certificate signatures in X509_verify_cert()
9347 by default (a flag can override this): it just wastes time without
9348 adding any security. As a useful side effect self signed root CAs
9349 with non-FIPS digests are now usable in FIPS mode.
9353 * In dtls1_process_out_of_seq_message() the check if the current message
9354 is already buffered was missing. For every new message was memory
9355 allocated, allowing an attacker to perform an denial of service attack
9356 with sending out of seq handshake messages until there is no memory
9357 left. Additionally every future message was buffered, even if the
9358 sequence number made no sense and would be part of another handshake.
9359 So only messages with sequence numbers less than 10 in advance will be
9360 buffered. ([CVE-2009-1378])
9362 *Robin Seggelmann, discovered by Daniel Mentz*
9364 * Records are buffered if they arrive with a future epoch to be
9365 processed after finishing the corresponding handshake. There is
9366 currently no limitation to this buffer allowing an attacker to perform
9367 a DOS attack with sending records with future epochs until there is no
9368 memory left. This patch adds the pqueue_size() function to determine
9369 the size of a buffer and limits the record buffer to 100 entries.
9372 *Robin Seggelmann, discovered by Daniel Mentz*
9374 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9375 parent structure is freed. ([CVE-2009-1379])
9379 * Handle non-blocking I/O properly in SSL_shutdown() call.
9381 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9383 * Add `2.5.4.*` OIDs
9385 *Ilya O. <vrghost@gmail.com>*
9387 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9389 * Disable renegotiation completely - this fixes a severe security
9390 problem ([CVE-2009-3555]) at the cost of breaking all
9391 renegotiation. Renegotiation can be re-enabled by setting
9392 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9393 run-time. This is really not recommended unless you know what
9398 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9400 * Don't set val to NULL when freeing up structures, it is freed up by
9401 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9402 zeroing past the valid field. ([CVE-2009-0789])
9404 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9406 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9407 checked correctly. This would allow some invalid signed attributes to
9408 appear to verify correctly. ([CVE-2009-0591])
9410 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9412 * Reject UniversalString and BMPString types with invalid lengths. This
9413 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9414 a legal length. ([CVE-2009-0590])
9418 * Set S/MIME signing as the default purpose rather than setting it
9419 unconditionally. This allows applications to override it at the store
9424 * Permit restricted recursion of ASN1 strings. This is needed in practice
9425 to handle some structures.
9429 * Improve efficiency of mem_gets: don't search whole buffer each time
9432 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9434 * New -hex option for openssl rand.
9438 * Print out UTF8String and NumericString when parsing ASN1.
9442 * Support NumericString type for name components.
9446 * Allow CC in the environment to override the automatically chosen
9447 compiler. Note that nothing is done to ensure flags work with the
9452 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9454 * Properly check EVP_VerifyFinal() and similar return values
9457 *Ben Laurie, Bodo Moeller, Google Security Team*
9459 * Enable TLS extensions by default.
9463 * Allow the CHIL engine to be loaded, whether the application is
9464 multithreaded or not. (This does not release the developer from the
9465 obligation to set up the dynamic locking callbacks.)
9467 *Sander Temme <sander@temme.net>*
9469 * Use correct exit code if there is an error in dgst command.
9471 *Steve Henson; problem pointed out by Roland Dirlewanger*
9473 * Tweak Configure so that you need to say "experimental-jpake" to enable
9474 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9478 * Add experimental JPAKE support, including demo authentication in
9479 s_client and s_server.
9483 * Set the comparison function in v3_addr_canonize().
9485 *Rob Austein <sra@hactrn.net>*
9487 * Add support for XMPP STARTTLS in s_client.
9489 *Philip Paeps <philip@freebsd.org>*
9491 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9492 to ensure that even with this option, only ciphersuites in the
9493 server's preference list will be accepted. (Note that the option
9494 applies only when resuming a session, so the earlier behavior was
9495 just about the algorithm choice for symmetric cryptography.)
9499 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9501 * Fix NULL pointer dereference if a DTLS server received
9502 ChangeCipherSpec as first record ([CVE-2009-1386]).
9506 * Fix a state transition in s3_srvr.c and d1_srvr.c
9507 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9511 * The fix in 0.9.8c that supposedly got rid of unsafe
9512 double-checked locking was incomplete for RSA blinding,
9513 addressing just one layer of what turns out to have been
9514 doubly unsafe triple-checked locking.
9516 So now fix this for real by retiring the MONT_HELPER macro
9517 in crypto/rsa/rsa_eay.c.
9519 *Bodo Moeller; problem pointed out by Marius Schilder*
9521 * Various precautionary measures:
9523 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9525 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9526 (NB: This would require knowledge of the secret session ticket key
9527 to exploit, in which case you'd be SOL either way.)
9529 - Change bn_nist.c so that it will properly handle input BIGNUMs
9530 outside the expected range.
9532 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9535 *Neel Mehta, Bodo Moeller*
9537 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9538 the load fails. Useful for distros.
9540 *Ben Laurie and the FreeBSD team*
9542 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9546 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9550 * Expand ENGINE to support engine supplied SSL client certificate functions.
9552 This work was sponsored by Logica.
9556 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9557 keystores. Support for SSL/TLS client authentication too.
9558 Not compiled unless enable-capieng specified to Configure.
9560 This work was sponsored by Logica.
9564 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9565 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9566 attribute creation routines such as certificate requests and PKCS#12
9571 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9573 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9574 handshake which could lead to a client crash as found using the
9575 Codenomicon TLS test suite ([CVE-2008-1672])
9577 *Steve Henson, Mark Cox*
9579 * Fix double free in TLS server name extensions which could lead to
9580 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9584 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9586 Clear the error queue to ensure that error entries left from
9587 older function calls do not interfere with the correct operation.
9589 *Lutz Jaenicke, Erik de Castro Lopo*
9591 * Remove root CA certificates of commercial CAs:
9593 The OpenSSL project does not recommend any specific CA and does not
9594 have any policy with respect to including or excluding any CA.
9595 Therefore it does not make any sense to ship an arbitrary selection
9596 of root CA certificates with the OpenSSL software.
9600 * RSA OAEP patches to fix two separate invalid memory reads.
9601 The first one involves inputs when 'lzero' is greater than
9602 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9603 before the beginning of from). The second one involves inputs where
9604 the 'db' section contains nothing but zeroes (there is a one-byte
9605 invalid read after the end of 'db').
9607 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9609 * Partial backport from 0.9.9-dev:
9611 Introduce bn_mul_mont (dedicated Montgomery multiplication
9612 procedure) as a candidate for BIGNUM assembler implementation.
9613 While 0.9.9-dev uses assembler for various architectures, only
9614 x86_64 is available by default here in the 0.9.8 branch, and
9615 32-bit x86 is available through a compile-time setting.
9617 To try the 32-bit x86 assembler implementation, use Configure
9618 option "enable-montasm" (which exists only for this backport).
9620 As "enable-montasm" for 32-bit x86 disclaims code stability
9621 anyway, in this constellation we activate additional code
9622 backported from 0.9.9-dev for further performance improvements,
9623 namely BN_from_montgomery_word. (To enable this otherwise,
9624 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9626 *Andy Polyakov (backport partially by Bodo Moeller)*
9628 * Add TLS session ticket callback. This allows an application to set
9629 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9630 values. This is useful for key rollover for example where several key
9631 sets may exist with different names.
9635 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9636 This was broken until now in 0.9.8 releases, such that the only way
9637 a registered ENGINE could be used (assuming it initialises
9638 successfully on the host) was to explicitly set it as the default
9639 for the relevant algorithms. This is in contradiction with 0.9.7
9640 behaviour and the documentation. With this fix, when an ENGINE is
9641 registered into a given algorithm's table of implementations, the
9642 'uptodate' flag is reset so that auto-discovery will be used next
9643 time a new context for that algorithm attempts to select an
9646 *Ian Lister (tweaked by Geoff Thorpe)*
9648 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9649 implementation in the following ways:
9651 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9654 Lack of BER streaming support means one pass streaming processing is
9655 only supported if data is detached: setting the streaming flag is
9656 ignored for embedded content.
9658 CMS support is disabled by default and must be explicitly enabled
9659 with the enable-cms configuration option.
9663 * Update the GMP engine glue to do direct copies between BIGNUM and
9664 mpz_t when openssl and GMP use the same limb size. Otherwise the
9665 existing "conversion via a text string export" trick is still used.
9667 *Paul Sheer <paulsheer@gmail.com>*
9669 * Zlib compression BIO. This is a filter BIO which compressed and
9670 uncompresses any data passed through it.
9674 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9675 RFC3394 compatible AES key wrapping.
9679 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9680 sets string data without copying. X509_ALGOR_set0() and
9681 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9682 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9683 from an X509_ATTRIBUTE structure optionally checking it occurs only
9684 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9689 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9690 to get the expected BN_FLG_CONSTTIME behavior.
9692 *Bodo Moeller (Google)*
9696 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9697 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9698 - added some more tests to do_tests.pl
9699 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9700 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9701 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9702 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9703 - various changes to netware.pl to enable gcc-cross builds on Win32
9705 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9706 - various changes to fix missing prototype warnings
9707 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9708 - added AES, WHIRLPOOL and CPUID assembler code to build files
9709 - added missing AES assembler make rules to mk1mf.pl
9710 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9712 *Guenter Knauf <eflash@gmx.net>*
9714 * Implement certificate status request TLS extension defined in RFC3546.
9715 A client can set the appropriate parameters and receive the encoded
9716 OCSP response via a callback. A server can query the supplied parameters
9717 and set the encoded OCSP response in the callback. Add simplified examples
9718 to s_client and s_server.
9722 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9725 + Binary incompatibility of ssl_ctx_st structure
9726 + DTLS interoperation with non-compliant servers
9727 + Don't call get_session_cb() without proposed session
9728 + Fix ia64 assembler code
9730 *Andy Polyakov, Steve Henson*
9732 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9734 * DTLS Handshake overhaul. There were longstanding issues with
9735 OpenSSL DTLS implementation, which were making it impossible for
9736 RFC 4347 compliant client to communicate with OpenSSL server.
9737 Unfortunately just fixing these incompatibilities would "cut off"
9738 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9739 server keeps tolerating non RFC compliant syntax. The opposite is
9740 not true, 0.9.8f client can not communicate with earlier server.
9741 This update even addresses CVE-2007-4995.
9745 * Changes to avoid need for function casts in OpenSSL: some compilers
9746 (gcc 4.2 and later) reject their use.
9747 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9750 * Add RFC4507 support to OpenSSL. This includes the corrections in
9751 RFC4507bis. The encrypted ticket format is an encrypted encoded
9752 SSL_SESSION structure, that way new session features are automatically
9755 If a client application caches session in an SSL_SESSION structure
9756 support is transparent because tickets are now stored in the encoded
9759 The SSL_CTX structure automatically generates keys for ticket
9760 protection in servers so again support should be possible
9761 with no application modification.
9763 If a client or server wishes to disable RFC4507 support then the option
9764 SSL_OP_NO_TICKET can be set.
9766 Add a TLS extension debugging callback to allow the contents of any client
9767 or server extensions to be examined.
9769 This work was sponsored by Google.
9773 * Add initial support for TLS extensions, specifically for the server_name
9774 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9775 have new members for a host name. The SSL data structure has an
9776 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9777 stored in that context to allow for session resumption, even after the
9778 SSL has been switched to a new SSL_CTX in reaction to a client's
9779 server_name extension.
9781 New functions (subject to change):
9783 SSL_get_servername()
9784 SSL_get_servername_type()
9787 New CTRL codes and macros (subject to change):
9789 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9790 - SSL_CTX_set_tlsext_servername_callback()
9791 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9792 - SSL_CTX_set_tlsext_servername_arg()
9793 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9795 openssl s_client has a new '-servername ...' option.
9797 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9798 '-key2 ...', '-servername_fatal' (subject to change). This allows
9799 testing the HostName extension for a specific single host name ('-cert'
9800 and '-key' remain fallbacks for handshakes without HostName
9801 negotiation). If the unrecognized_name alert has to be sent, this by
9802 default is a warning; it becomes fatal with the '-servername_fatal'
9805 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9807 * Add AES and SSE2 assembly language support to VC++ build.
9811 * Mitigate attack on final subtraction in Montgomery reduction.
9815 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9816 (which previously caused an internal error).
9820 * Squeeze another 10% out of IGE mode when in != out.
9824 * AES IGE mode speedup.
9826 *Dean Gaudet (Google)*
9828 * Add the Korean symmetric 128-bit cipher SEED (see
9829 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9830 add SEED ciphersuites from RFC 4162:
9832 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9833 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9834 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9835 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9837 To minimize changes between patchlevels in the OpenSSL 0.9.8
9838 series, SEED remains excluded from compilation unless OpenSSL
9839 is configured with 'enable-seed'.
9841 *KISA, Bodo Moeller*
9843 * Mitigate branch prediction attacks, which can be practical if a
9844 single processor is shared, allowing a spy process to extract
9845 information. For detailed background information, see
9846 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9847 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9848 and Necessary Software Countermeasures"). The core of the change
9849 are new versions BN_div_no_branch() and
9850 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9851 respectively, which are slower, but avoid the security-relevant
9852 conditional branches. These are automatically called by BN_div()
9853 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9854 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9855 remove a conditional branch.
9857 BN_FLG_CONSTTIME is the new name for the previous
9858 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9859 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9860 in the exponent causes BN_mod_exp_mont() to use the alternative
9861 implementation in BN_mod_exp_mont_consttime().) The old name
9862 remains as a deprecated alias.
9864 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9865 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9866 constant-time implementations for more than just exponentiation.
9867 Here too the old name is kept as a deprecated alias.
9869 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9870 the BN_BLINDING structure gets an independent copy of the
9871 modulus. This means that the previous `BIGNUM *m` argument to
9872 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9873 essentially becomes `const BIGNUM *m`, although we can't actually
9874 change this in the header file before 0.9.9. It allows
9875 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9876 enable BN_FLG_CONSTTIME.
9878 *Matthew D Wood (Intel Corp)*
9880 * In the SSL/TLS server implementation, be strict about session ID
9881 context matching (which matters if an application uses a single
9882 external cache for different purposes). Previously,
9883 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9884 set. This did ensure strict client verification, but meant that,
9885 with applications using a single external cache for quite
9886 different requirements, clients could circumvent ciphersuite
9887 restrictions for a given session ID context by starting a session
9888 in a different context.
9892 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9893 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9894 authentication-only ciphersuites.
9898 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9899 not complete and could lead to a possible single byte overflow
9900 ([CVE-2007-5135]) [Ben Laurie]
9902 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9904 * Since AES128 and AES256 (and similarly Camellia128 and
9905 Camellia256) share a single mask bit in the logic of
9906 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9907 kludge to work properly if AES128 is available and AES256 isn't
9908 (or if Camellia128 is available and Camellia256 isn't).
9912 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9913 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9914 When a point or a seed is encoded in a BIT STRING, we need to
9915 prevent the removal of trailing zero bits to get the proper DER
9916 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9917 of a NamedBitList, for which trailing 0 bits need to be removed.)
9921 * Have SSL/TLS server implementation tolerate "mismatched" record
9922 protocol version while receiving ClientHello even if the
9923 ClientHello is fragmented. (The server can't insist on the
9924 particular protocol version it has chosen before the ServerHello
9925 message has informed the client about his choice.)
9929 * Add RFC 3779 support.
9931 *Rob Austein for ARIN, Ben Laurie*
9933 * Load error codes if they are not already present instead of using a
9934 static variable. This allows them to be cleanly unloaded and reloaded.
9935 Improve header file function name parsing.
9939 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9940 or CAPABILITY handshake as required by RFCs.
9944 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9946 * Introduce limits to prevent malicious keys being able to
9947 cause a denial of service. ([CVE-2006-2940])
9949 *Steve Henson, Bodo Moeller*
9951 * Fix ASN.1 parsing of certain invalid structures that can result
9952 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9954 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9955 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9957 * Fix SSL client code which could crash if connecting to a
9958 malicious SSLv2 server. ([CVE-2006-4343])
9960 *Tavis Ormandy and Will Drewry, Google Security Team*
9962 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9963 match only those. Before that, "AES256-SHA" would be interpreted
9964 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9965 the same strength classification in 0.9.7h) as we currently only
9966 have a single AES bit in the ciphersuite description bitmap.
9967 That change, however, also applied to ciphersuite strings such as
9968 "RC4-MD5" that intentionally matched multiple ciphersuites --
9969 namely, SSL 2.0 ciphersuites in addition to the more common ones
9970 from SSL 3.0/TLS 1.0.
9972 So we change the selection algorithm again: Naming an explicit
9973 ciphersuite selects this one ciphersuite, and any other similar
9974 ciphersuite (same bitmap) from *other* protocol versions.
9975 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9976 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9978 Since SSL 2.0 does not have any ciphersuites for which the
9979 128/256 bit distinction would be relevant, this works for now.
9980 The proper fix will be to use different bits for AES128 and
9981 AES256, which would have avoided the problems from the beginning;
9982 however, bits are scarce, so we can only do this in a new release
9983 (not just a patchlevel) when we can change the SSL_CIPHER
9984 definition to split the single 'unsigned long mask' bitmap into
9985 multiple values to extend the available space.
9989 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9991 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9992 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9994 * Add AES IGE and biIGE modes.
9998 * Change the Unix randomness entropy gathering to use poll() when
9999 possible instead of select(), since the latter has some
10000 undesirable limitations.
10002 *Darryl Miles via Richard Levitte and Bodo Moeller*
10004 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10005 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10006 cannot be implicitly activated as part of, e.g., the "AES" alias.
10007 However, please upgrade to OpenSSL 0.9.9[-dev] for
10008 non-experimental use of the ECC ciphersuites to get TLS extension
10009 support, which is required for curve and point format negotiation
10010 to avoid potential handshake problems.
10014 * Disable rogue ciphersuites:
10016 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10017 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10018 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10020 The latter two were purportedly from
10021 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10024 Also deactivate the remaining ciphersuites from
10025 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10026 unofficial, and the ID has long expired.
10030 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10031 dual-core machines) and other potential thread-safety issues.
10035 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10036 versions), which is now available for royalty-free use
10037 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10038 Also, add Camellia TLS ciphersuites from RFC 4132.
10040 To minimize changes between patchlevels in the OpenSSL 0.9.8
10041 series, Camellia remains excluded from compilation unless OpenSSL
10042 is configured with 'enable-camellia'.
10046 * Disable the padding bug check when compression is in use. The padding
10047 bug check assumes the first packet is of even length, this is not
10048 necessarily true if compression is enabled and can result in false
10049 positives causing handshake failure. The actual bug test is ancient
10050 code so it is hoped that implementations will either have fixed it by
10051 now or any which still have the bug do not support compression.
10055 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10057 * When applying a cipher rule check to see if string match is an explicit
10058 cipher suite and only match that one cipher suite if it is.
10062 * Link in manifests for VC++ if needed.
10064 *Austin Ziegler <halostatue@gmail.com>*
10066 * Update support for ECC-based TLS ciphersuites according to
10067 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10068 TLS extensions, which are supported starting with the 0.9.9
10069 branch, not in the OpenSSL 0.9.8 branch).
10073 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10074 opaque EVP_CIPHER_CTX handling.
10078 * Fixes and enhancements to zlib compression code. We now only use
10079 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10080 to conform with the standards mentioned here:
10081 <http://www.zlib.net/DLL_FAQ.txt>
10082 Static zlib linking now works on Windows and the new --with-zlib-include
10083 --with-zlib-lib options to Configure can be used to supply the location
10084 of the headers and library. Gracefully handle case where zlib library
10089 * Several fixes and enhancements to the OID generation code. The old code
10090 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10091 handle numbers larger than ULONG_MAX, truncated printing and had a
10092 non standard OBJ_obj2txt() behaviour.
10096 * Add support for building of engines under engine/ as shared libraries
10097 under VC++ build system.
10101 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10102 Hopefully, we will not see any false combination of paths any more.
10106 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10108 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10109 (part of SSL_OP_ALL). This option used to disable the
10110 countermeasure against man-in-the-middle protocol-version
10111 rollback in the SSL 2.0 server implementation, which is a bad
10112 idea. ([CVE-2005-2969])
10114 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10115 for Information Security, National Institute of Advanced Industrial
10116 Science and Technology [AIST], Japan)*
10118 * Add two function to clear and return the verify parameter flags.
10122 * Keep cipherlists sorted in the source instead of sorting them at
10123 runtime, thus removing the need for a lock.
10127 * Avoid some small subgroup attacks in Diffie-Hellman.
10129 *Nick Mathewson and Ben Laurie*
10131 * Add functions for well-known primes.
10135 * Extended Windows CE support.
10137 *Satoshi Nakamura and Andy Polyakov*
10139 * Initialize SSL_METHOD structures at compile time instead of during
10140 runtime, thus removing the need for a lock.
10144 * Make PKCS7_decrypt() work even if no certificate is supplied by
10145 attempting to decrypt each encrypted key in turn. Add support to
10150 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10152 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10155 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10159 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10160 key into the same file any more.
10164 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10168 * Add -utf8 command line and config file option to 'ca'.
10170 *Stefan <stf@udoma.org*
10172 * Removed the macro des_crypt(), as it seems to conflict with some
10173 libraries. Use DES_crypt().
10177 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10178 involves renaming the source and generated shared-libs for
10179 both. The engines will accept the corrected or legacy ids
10180 ('ncipher' and '4758_cca' respectively) when binding. NB,
10181 this only applies when building 'shared'.
10183 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10185 * Add attribute functions to EVP_PKEY structure. Modify
10186 PKCS12_create() to recognize a CSP name attribute and
10187 use it. Make -CSP option work again in pkcs12 utility.
10191 * Add new functionality to the bn blinding code:
10192 - automatic re-creation of the BN_BLINDING parameters after
10193 a fixed number of uses (currently 32)
10194 - add new function for parameter creation
10195 - introduce flags to control the update behaviour of the
10196 BN_BLINDING parameters
10197 - hide BN_BLINDING structure
10198 Add a second BN_BLINDING slot to the RSA structure to improve
10199 performance when a single RSA object is shared among several
10204 * Add support for DTLS.
10206 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10208 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10209 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10213 * Remove buggy and incomplete DH cert support from
10214 ssl/ssl_rsa.c and ssl/s3_both.c
10218 * Use SHA-1 instead of MD5 as the default digest algorithm for
10219 the `apps/openssl` commands.
10223 * Compile clean with "-Wall -Wmissing-prototypes
10224 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10225 DEBUG_SAFESTACK must also be set.
10229 * Change ./Configure so that certain algorithms can be disabled by default.
10230 The new counterpiece to "no-xxx" is "enable-xxx".
10232 The patented RC5 and MDC2 algorithms will now be disabled unless
10233 "enable-rc5" and "enable-mdc2", respectively, are specified.
10235 (IDEA remains enabled despite being patented. This is because IDEA
10236 is frequently required for interoperability, and there is no license
10237 fee for non-commercial use. As before, "no-idea" can be used to
10238 avoid this algorithm.)
10242 * Add processing of proxy certificates (see RFC 3820). This work was
10243 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10244 EGEE (Enabling Grids for E-science in Europe).
10248 * RC4 performance overhaul on modern architectures/implementations, such
10249 as Intel P4, IA-64 and AMD64.
10253 * New utility extract-section.pl. This can be used specify an alternative
10254 section number in a pod file instead of having to treat each file as
10255 a separate case in Makefile. This can be done by adding two lines to the
10258 =for comment openssl_section:XXX
10260 The blank line is mandatory.
10264 * New arguments -certform, -keyform and -pass for s_client and s_server
10265 to allow alternative format key and certificate files and passphrase
10270 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10271 update associated structures and add various utility functions.
10273 Add new policy related verify parameters, include policy checking in
10274 standard verify code. Enhance 'smime' application with extra parameters
10275 to support policy checking and print out.
10279 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10280 Nehemiah processors. These extensions support AES encryption in hardware
10281 as well as RNG (though RNG support is currently disabled).
10283 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10285 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10289 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10291 *Andy Polyakov and a number of other people*
10293 * Improved PowerPC platform support. Most notably BIGNUM assembler
10294 implementation contributed by IBM.
10296 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10298 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10299 exponent rather than 'unsigned long'. There is a corresponding change to
10300 the new 'rsa_keygen' element of the RSA_METHOD structure.
10302 *Jelte Jansen, Geoff Thorpe*
10304 * Functionality for creating the initial serial number file is now
10305 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10307 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10308 number file to 1, which is bound to cause problems. To avoid
10309 the problems while respecting compatibility between different 0.9.7
10310 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10311 CA.pl for serial number initialization. With the new release 0.9.8,
10312 we can fix the problem directly in the 'ca' utility.)
10316 * Reduced header interdependencies by declaring more opaque objects in
10317 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10318 give fewer recursive includes, which could break lazy source code - so
10319 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10320 developers should define this symbol when building and using openssl to
10321 ensure they track the recommended behaviour, interfaces, [etc], but
10322 backwards-compatible behaviour prevails when this isn't defined.
10326 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10330 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10331 This will generate a random key of the appropriate length based on the
10332 cipher context. The EVP_CIPHER can provide its own random key generation
10333 routine to support keys of a specific form. This is used in the des and
10334 3des routines to generate a key of the correct parity. Update S/MIME
10335 code to use new functions and hence generate correct parity DES keys.
10336 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10337 valid (weak or incorrect parity).
10341 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10342 as looking them up. This is useful when the verified structure may contain
10343 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10344 present unless the new PKCS7_NO_CRL flag is asserted.
10348 * Extend ASN1 oid configuration module. It now additionally accepts the
10351 shortName = some long name, 1.2.3.4
10355 * Reimplemented the BN_CTX implementation. There is now no more static
10356 limitation on the number of variables it can handle nor the depth of the
10357 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10358 information can now expand as required, and rather than having a single
10359 static array of bignums, BN_CTX now uses a linked-list of such arrays
10360 allowing it to expand on demand whilst maintaining the usefulness of
10361 BN_CTX's "bundling".
10365 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10366 to allow all RSA operations to function using a single BN_CTX.
10370 * Preliminary support for certificate policy evaluation and checking. This
10371 is initially intended to pass the tests outlined in "Conformance Testing
10372 of Relying Party Client Certificate Path Processing Logic" v1.07.
10376 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10377 remained unused and not that useful. A variety of other little bignum
10378 tweaks and fixes have also been made continuing on from the audit (see
10383 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10384 associated ASN1, EVP and SSL functions and old ASN1 macros.
10388 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10389 and this should never fail. So the return value from the use of
10390 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10391 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10395 * BN_CTX_get() should return zero-valued bignums, providing the same
10396 initialised value as BN_new().
10398 *Geoff Thorpe, suggested by Ulf Möller*
10400 * Support for inhibitAnyPolicy certificate extension.
10404 * An audit of the BIGNUM code is underway, for which debugging code is
10405 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10406 is considered valid when processing BIGNUMs, and causes execution to
10407 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10408 further steps are taken to deliberately pollute unused data in BIGNUM
10409 structures to try and expose faulty code further on. For now, openssl will
10410 (in its default mode of operation) continue to tolerate the inconsistent
10411 forms that it has tolerated in the past, but authors and packagers should
10412 consider trying openssl and their own applications when compiled with
10413 these debugging symbols defined. It will help highlight potential bugs in
10414 their own code, and will improve the test coverage for OpenSSL itself. At
10415 some point, these tighter rules will become openssl's default to improve
10416 maintainability, though the assert()s and other overheads will remain only
10417 in debugging configurations. See bn.h for more details.
10419 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10421 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10422 that can only be obtained through BN_CTX_new() (which implicitly
10423 initialises it). The presence of this function only made it possible
10424 to overwrite an existing structure (and cause memory leaks).
10428 * Because of the callback-based approach for implementing LHASH as a
10429 template type, lh_insert() adds opaque objects to hash-tables and
10430 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10431 to clean up those corresponding objects before destroying the hash table
10432 (and losing the object pointers). So some over-zealous constifications in
10433 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10434 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10435 prototyped to have "const" restrictions on the object pointers they are
10436 given (and so aren't required to cast them away any more).
10440 * The tmdiff.h API was so ugly and minimal that our own timing utility
10441 (speed) prefers to use its own implementation. The two implementations
10442 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10443 its object type properly exposed (MS_TM) instead of casting to/from
10444 `char *`. This may still change yet if someone realises MS_TM and
10446 aren't necessarily the greatest nomenclatures - but this is what was used
10447 internally to the implementation so I've used that for now.
10451 * Ensure that deprecated functions do not get compiled when
10452 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10453 the self-tests were still using deprecated key-generation functions so
10454 these have been updated also.
10458 * Reorganise PKCS#7 code to separate the digest location functionality
10459 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10460 New function PKCS7_set_digest() to set the digest type for PKCS#7
10461 digestedData type. Add additional code to correctly generate the
10462 digestedData type and add support for this type in PKCS7 initialization
10467 * New function PKCS7_set0_type_other() this initializes a PKCS7
10468 structure of type "other".
10472 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10473 sure the loop does correctly stop and breaking ("division by zero")
10474 modulus operations are not performed. The (pre-generated) prime
10475 table crypto/bn/bn_prime.h was already correct, but it could not be
10476 re-generated on some platforms because of the "division by zero"
10477 situation in the script.
10479 *Ralf S. Engelschall*
10481 * Update support for ECC-based TLS ciphersuites according to
10482 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10483 SHA-1 now is only used for "small" curves (where the
10484 representation of a field element takes up to 24 bytes); for
10485 larger curves, the field element resulting from ECDH is directly
10486 used as premaster secret.
10488 *Douglas Stebila (Sun Microsystems Laboratories)*
10490 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10491 curve secp160r1 to the tests.
10493 *Douglas Stebila (Sun Microsystems Laboratories)*
10495 * Add the possibility to load symbols globally with DSO.
10497 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10499 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10500 control of the error stack.
10504 * Add support for STORE in ENGINE.
10508 * Add the STORE type. The intention is to provide a common interface
10509 to certificate and key stores, be they simple file-based stores, or
10510 HSM-type store, or LDAP stores, or...
10511 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10515 * Add a generic structure called OPENSSL_ITEM. This can be used to
10516 pass a list of arguments to any function as well as provide a way
10517 for a function to pass data back to the caller.
10521 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10522 works like BUF_strdup() but can be used to duplicate a portion of
10523 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10528 * Add the function sk_find_ex() which works like sk_find(), but will
10529 return an index to an element even if an exact match couldn't be
10530 found. The index is guaranteed to point at the element where the
10531 searched-for key would be inserted to preserve sorting order.
10535 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10536 takes an extra flags argument for optional functionality. Currently,
10537 the following flags are defined:
10539 OBJ_BSEARCH_VALUE_ON_NOMATCH
10540 This one gets OBJ_bsearch_ex() to return a pointer to the first
10541 element where the comparing function returns a negative or zero
10544 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10545 This one gets OBJ_bsearch_ex() to return a pointer to the first
10546 element where the comparing function returns zero. This is useful
10547 if there are more than one element where the comparing function
10552 * Make it possible to create self-signed certificates with 'openssl ca'
10553 in such a way that the self-signed certificate becomes part of the
10554 CA database and uses the same mechanisms for serial number generation
10555 as all other certificate signing. The new flag '-selfsign' enables
10556 this functionality. Adapt CA.sh and CA.pl.in.
10560 * Add functionality to check the public key of a certificate request
10561 against a given private. This is useful to check that a certificate
10562 request can be signed by that key (self-signing).
10566 * Make it possible to have multiple active certificates with the same
10567 subject in the CA index file. This is done only if the keyword
10568 'unique_subject' is set to 'no' in the main CA section (default
10569 if 'CA_default') of the configuration file. The value is saved
10570 with the database itself in a separate index attribute file,
10571 named like the index file with '.attr' appended to the name.
10575 * Generate multi-valued AVAs using '+' notation in config files for
10580 * Support for nameConstraints certificate extension.
10584 * Support for policyConstraints certificate extension.
10588 * Support for policyMappings certificate extension.
10592 * Make sure the default DSA_METHOD implementation only uses its
10593 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10594 and change its own handlers to be NULL so as to remove unnecessary
10595 indirection. This lets alternative implementations fallback to the
10596 default implementation more easily.
10600 * Support for directoryName in GeneralName related extensions
10605 * Make it possible to link applications using Makefile.shared.
10606 Make that possible even when linking against static libraries!
10610 * Support for single pass processing for S/MIME signing. This now
10611 means that S/MIME signing can be done from a pipe, in addition
10612 cleartext signing (multipart/signed type) is effectively streaming
10613 and the signed data does not need to be all held in memory.
10615 This is done with a new flag PKCS7_STREAM. When this flag is set
10616 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10617 is done after the data is output (and digests calculated) in
10618 SMIME_write_PKCS7().
10622 * Add full support for -rpath/-R, both in shared libraries and
10623 applications, at least on the platforms where it's known how
10628 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10629 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10630 will now compute a table of multiples of the generator that
10631 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10632 faster (notably in the case of a single point multiplication,
10633 scalar * generator).
10635 *Nils Larsch, Bodo Moeller*
10637 * IPv6 support for certificate extensions. The various extensions
10638 which use the IP:a.b.c.d can now take IPv6 addresses using the
10639 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10644 * Added an ENGINE that implements RSA by performing private key
10645 exponentiations with the GMP library. The conversions to and from
10646 GMP's mpz_t format aren't optimised nor are any montgomery forms
10647 cached, and on x86 it appears OpenSSL's own performance has caught up.
10648 However there are likely to be other architectures where GMP could
10649 provide a boost. This ENGINE is not built in by default, but it can be
10650 specified at Configure time and should be accompanied by the necessary
10651 linker additions, eg;
10652 ./config -DOPENSSL_USE_GMP -lgmp
10656 * "openssl engine" will not display ENGINE/DSO load failure errors when
10657 testing availability of engines with "-t" - the old behaviour is
10658 produced by increasing the feature's verbosity with "-tt".
10662 * ECDSA routines: under certain error conditions uninitialized BN objects
10663 could be freed. Solution: make sure initialization is performed early
10664 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10669 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10670 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10671 software implementations. For DSA and DH, parameter generation can
10672 also be overridden by providing the appropriate method callbacks.
10676 * Change the "progress" mechanism used in key-generation and
10677 primality testing to functions that take a new BN_GENCB pointer in
10678 place of callback/argument pairs. The new API functions have `_ex`
10679 postfixes and the older functions are reimplemented as wrappers for
10680 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10681 declarations of the old functions to help (graceful) attempts to
10682 migrate to the new functions. Also, the new key-generation API
10683 functions operate on a caller-supplied key-structure and return
10684 success/failure rather than returning a key or NULL - this is to
10685 help make "keygen" another member function of RSA_METHOD etc.
10687 Example for using the new callback interface:
10689 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10690 void *my_arg = ...;
10693 BN_GENCB_set(&my_cb, my_callback, my_arg);
10695 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10696 /* For the meaning of a, b in calls to my_callback(), see the
10697 * documentation of the function that calls the callback.
10698 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10699 * my_callback should return 1 if it wants BN_is_prime_ex()
10700 * to continue, or 0 to stop.
10705 * Change the ZLIB compression method to be stateful, and make it
10706 available to TLS with the number defined in
10707 draft-ietf-tls-compression-04.txt.
10711 * Add the ASN.1 structures and functions for CertificatePair, which
10712 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10714 CertificatePair ::= SEQUENCE {
10715 forward [0] Certificate OPTIONAL,
10716 reverse [1] Certificate OPTIONAL,
10717 -- at least one of the pair shall be present -- }
10719 Also implement the PEM functions to read and write certificate
10720 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10722 This needed to be defined, mostly for the sake of the LDAP
10723 attribute crossCertificatePair, but may prove useful elsewhere as
10728 * Make it possible to inhibit symlinking of shared libraries in
10729 Makefile.shared, for Cygwin's sake.
10733 * Extend the BIGNUM API by creating a function
10734 void BN_set_negative(BIGNUM *a, int neg);
10735 and a macro that behave like
10736 int BN_is_negative(const BIGNUM *a);
10738 to avoid the need to access 'a->neg' directly in applications.
10742 * Implement fast modular reduction for pseudo-Mersenne primes
10743 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10744 EC_GROUP_new_curve_GFp() will now automatically use this
10747 *Nils Larsch <nla@trustcenter.de>*
10749 * Add new lock type (CRYPTO_LOCK_BN).
10753 * Change the ENGINE framework to automatically load engines
10754 dynamically from specific directories unless they could be
10755 found to already be built in or loaded. Move all the
10756 current engines except for the cryptodev one to a new
10757 directory engines/.
10758 The engines in engines/ are built as shared libraries if
10759 the "shared" options was given to ./Configure or ./config.
10760 Otherwise, they are inserted in libcrypto.a.
10761 /usr/local/ssl/engines is the default directory for dynamic
10762 engines, but that can be overridden at configure time through
10763 the usual use of --prefix and/or --openssldir, and at run
10764 time with the environment variable OPENSSL_ENGINES.
10766 *Geoff Thorpe and Richard Levitte*
10768 * Add Makefile.shared, a helper makefile to build shared
10769 libraries. Adapt Makefile.org.
10773 * Add version info to Win32 DLLs.
10775 *Peter 'Luna' Runestig" <peter@runestig.com>*
10777 * Add new 'medium level' PKCS#12 API. Certificates and keys
10778 can be added using this API to created arbitrary PKCS#12
10779 files while avoiding the low-level API.
10781 New options to PKCS12_create(), key or cert can be NULL and
10782 will then be omitted from the output file. The encryption
10783 algorithm NIDs can be set to -1 for no encryption, the mac
10784 iteration count can be set to 0 to omit the mac.
10786 Enhance pkcs12 utility by making the -nokeys and -nocerts
10787 options work when creating a PKCS#12 file. New option -nomac
10788 to omit the mac, NONE can be set for an encryption algorithm.
10789 New code is modified to use the enhanced PKCS12_create()
10790 instead of the low-level API.
10794 * Extend ASN1 encoder to support indefinite length constructed
10795 encoding. This can output sequences tags and octet strings in
10796 this form. Modify pk7_asn1.c to support indefinite length
10797 encoding. This is experimental and needs additional code to
10798 be useful, such as an ASN1 bio and some enhanced streaming
10801 Extend template encode functionality so that tagging is passed
10802 down to the template encoder.
10806 * Let 'openssl req' fail if an argument to '-newkey' is not
10807 recognized instead of using RSA as a default.
10811 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10812 As these are not official, they are not included in "ALL";
10813 the "ECCdraft" ciphersuite group alias can be used to select them.
10815 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10817 * Add ECDH engine support.
10819 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10821 * Add ECDH in new directory crypto/ecdh/.
10823 *Douglas Stebila (Sun Microsystems Laboratories)*
10825 * Let BN_rand_range() abort with an error after 100 iterations
10826 without success (which indicates a broken PRNG).
10830 * Change BN_mod_sqrt() so that it verifies that the input value
10831 is really the square of the return value. (Previously,
10832 BN_mod_sqrt would show GIGO behaviour.)
10836 * Add named elliptic curves over binary fields from X9.62, SECG,
10837 and WAP/WTLS; add OIDs that were still missing.
10839 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10841 * Extend the EC library for elliptic curves over binary fields
10842 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10845 EC_GF2m_simple_method
10849 EC_GROUP_new_curve_GF2m
10850 EC_GROUP_set_curve_GF2m
10851 EC_GROUP_get_curve_GF2m
10852 EC_POINT_set_affine_coordinates_GF2m
10853 EC_POINT_get_affine_coordinates_GF2m
10854 EC_POINT_set_compressed_coordinates_GF2m
10856 Point compression for binary fields is disabled by default for
10857 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10860 As binary polynomials are represented as BIGNUMs, various members
10861 of the EC_GROUP and EC_POINT data structures can be shared
10862 between the implementations for prime fields and binary fields;
10863 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10864 are essentially identical to their `..._GFp` counterparts.
10865 (For simplicity, the `..._GFp` prefix has been dropped from
10866 various internal method names.)
10868 An internal 'field_div' method (similar to 'field_mul' and
10869 'field_sqr') has been added; this is used only for binary fields.
10871 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10873 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10874 through methods ('mul', 'precompute_mult').
10876 The generic implementations (now internally called 'ec_wNAF_mul'
10877 and 'ec_wNAF_precomputed_mult') remain the default if these
10878 methods are undefined.
10880 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10882 * New function EC_GROUP_get_degree, which is defined through
10883 EC_METHOD. For curves over prime fields, this returns the bit
10884 length of the modulus.
10886 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10888 * New functions EC_GROUP_dup, EC_POINT_dup.
10889 (These simply call ..._new and ..._copy).
10891 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10893 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10894 Polynomials are represented as BIGNUMs (where the sign bit is not
10895 used) in the following functions [macros]:
10898 BN_GF2m_sub [= BN_GF2m_add]
10899 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10900 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10901 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10903 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10904 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10905 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10906 BN_GF2m_cmp [= BN_ucmp]
10908 (Note that only the 'mod' functions are actually for fields GF(2^m).
10909 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10911 For some functions, an the irreducible polynomial defining a
10912 field can be given as an 'unsigned int[]' with strictly
10913 decreasing elements giving the indices of those bits that are set;
10914 i.e., p[] represents the polynomial
10915 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10917 p[0] > p[1] > ... > p[k] = 0.
10918 This applies to the following functions:
10921 BN_GF2m_mod_mul_arr
10922 BN_GF2m_mod_sqr_arr
10923 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10924 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10925 BN_GF2m_mod_exp_arr
10926 BN_GF2m_mod_sqrt_arr
10927 BN_GF2m_mod_solve_quad_arr
10931 Conversion can be performed by the following functions:
10936 bntest.c has additional tests for binary polynomial arithmetic.
10938 Two implementations for BN_GF2m_mod_div() are available.
10939 The default algorithm simply uses BN_GF2m_mod_inv() and
10940 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10941 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10942 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10944 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10946 * Add new error code 'ERR_R_DISABLED' that can be used when some
10947 functionality is disabled at compile-time.
10949 *Douglas Stebila <douglas.stebila@sun.com>*
10951 * Change default behaviour of 'openssl asn1parse' so that more
10952 information is visible when viewing, e.g., a certificate:
10954 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10955 mode the content of non-printable OCTET STRINGs is output in a
10956 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10957 avoid the appearance of a printable string.
10959 *Nils Larsch <nla@trustcenter.de>*
10961 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10963 EC_GROUP_set_asn1_flag()
10964 EC_GROUP_get_asn1_flag()
10965 EC_GROUP_set_point_conversion_form()
10966 EC_GROUP_get_point_conversion_form()
10967 These control ASN1 encoding details:
10968 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10969 has been set to OPENSSL_EC_NAMED_CURVE.
10970 - Points are encoded in uncompressed form by default; options for
10971 asn1_for are as for point2oct, namely
10972 POINT_CONVERSION_COMPRESSED
10973 POINT_CONVERSION_UNCOMPRESSED
10974 POINT_CONVERSION_HYBRID
10976 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10978 EC_GROUP_set_seed()
10979 EC_GROUP_get0_seed()
10980 EC_GROUP_get_seed_len()
10981 This is used only for ASN1 purposes (so far).
10983 *Nils Larsch <nla@trustcenter.de>*
10985 * Add 'field_type' member to EC_METHOD, which holds the NID
10986 of the appropriate field type OID. The new function
10987 EC_METHOD_get_field_type() returns this value.
10989 *Nils Larsch <nla@trustcenter.de>*
10992 EC_POINT_point2bn()
10993 EC_POINT_bn2point()
10994 EC_POINT_point2hex()
10995 EC_POINT_hex2point()
10996 providing useful interfaces to EC_POINT_point2oct() and
10997 EC_POINT_oct2point().
10999 *Nils Larsch <nla@trustcenter.de>*
11001 * Change internals of the EC library so that the functions
11002 EC_GROUP_set_generator()
11003 EC_GROUP_get_generator()
11004 EC_GROUP_get_order()
11005 EC_GROUP_get_cofactor()
11006 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11007 to methods, which would lead to unnecessary code duplication when
11008 adding different types of curves.
11010 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11012 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11013 arithmetic, and such that modified wNAFs are generated
11014 (which avoid length expansion in many cases).
11018 * Add a function EC_GROUP_check_discriminant() (defined via
11019 EC_METHOD) that verifies that the curve discriminant is non-zero.
11021 Add a function EC_GROUP_check() that makes some sanity tests
11022 on a EC_GROUP, its generator and order. This includes
11023 EC_GROUP_check_discriminant().
11025 *Nils Larsch <nla@trustcenter.de>*
11027 * Add ECDSA in new directory crypto/ecdsa/.
11029 Add applications 'openssl ecparam' and 'openssl ecdsa'
11030 (these are based on 'openssl dsaparam' and 'openssl dsa').
11032 ECDSA support is also included in various other files across the
11033 library. Most notably,
11034 - 'openssl req' now has a '-newkey ecdsa:file' option;
11035 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11036 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11037 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11038 them suitable for ECDSA where domain parameters must be
11039 extracted before the specific public key;
11040 - ECDSA engine support has been added.
11042 *Nils Larsch <nla@trustcenter.de>*
11044 * Include some named elliptic curves, and add OIDs from X9.62,
11045 SECG, and WAP/WTLS. Each curve can be obtained from the new
11047 EC_GROUP_new_by_curve_name(),
11048 and the list of available named curves can be obtained with
11049 EC_get_builtin_curves().
11050 Also add a 'curve_name' member to EC_GROUP objects, which can be
11052 EC_GROUP_set_curve_name()
11053 EC_GROUP_get_curve_name()
11055 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11057 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11058 was actually never needed) and in BN_mul(). The removal in BN_mul()
11059 required a small change in bn_mul_part_recursive() and the addition
11060 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11061 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11062 bn_sub_words() and bn_add_words() except they take arrays with
11067 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11069 * Cleanse PEM buffers before freeing them since they may contain
11072 *Benjamin Bennett <ben@psc.edu>*
11074 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11075 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11076 authentication-only ciphersuites.
11080 * Since AES128 and AES256 share a single mask bit in the logic of
11081 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11082 kludge to work properly if AES128 is available and AES256 isn't.
11086 * Expand security boundary to match 1.1.1 module.
11090 * Remove redundant features: hash file source, editing of test vectors
11091 modify fipsld to use external fips_premain.c signature.
11095 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11096 run algorithm test programs.
11100 * Make algorithm test programs more tolerant of whitespace.
11104 * Have SSL/TLS server implementation tolerate "mismatched" record
11105 protocol version while receiving ClientHello even if the
11106 ClientHello is fragmented. (The server can't insist on the
11107 particular protocol version it has chosen before the ServerHello
11108 message has informed the client about his choice.)
11112 * Load error codes if they are not already present instead of using a
11113 static variable. This allows them to be cleanly unloaded and reloaded.
11117 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11119 * Introduce limits to prevent malicious keys being able to
11120 cause a denial of service. ([CVE-2006-2940])
11122 *Steve Henson, Bodo Moeller*
11124 * Fix ASN.1 parsing of certain invalid structures that can result
11125 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11127 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11128 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11130 * Fix SSL client code which could crash if connecting to a
11131 malicious SSLv2 server. ([CVE-2006-4343])
11133 *Tavis Ormandy and Will Drewry, Google Security Team*
11135 * Change ciphersuite string processing so that an explicit
11136 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11137 will no longer include "AES128-SHA"), and any other similar
11138 ciphersuite (same bitmap) from *other* protocol versions (so that
11139 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11140 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11141 changes from 0.9.8b and 0.9.8d.
11145 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11147 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11148 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11150 * Change the Unix randomness entropy gathering to use poll() when
11151 possible instead of select(), since the latter has some
11152 undesirable limitations.
11154 *Darryl Miles via Richard Levitte and Bodo Moeller*
11156 * Disable rogue ciphersuites:
11158 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11159 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11160 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11162 The latter two were purportedly from
11163 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11166 Also deactivate the remaining ciphersuites from
11167 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11168 unofficial, and the ID has long expired.
11172 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11173 dual-core machines) and other potential thread-safety issues.
11177 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11179 * Adapt fipsld and the build system to link against the validated FIPS
11180 module in FIPS mode.
11184 * Fixes for VC++ 2005 build under Windows.
11188 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11189 from a Windows bash shell such as MSYS. It is autodetected from the
11190 "config" script when run from a VC++ environment. Modify standard VC++
11191 build to use fipscanister.o from the GNU make build.
11195 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11197 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11198 The value now differs depending on if you build for FIPS or not.
11199 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11200 safely run with a non-FIPSed libcrypto, as it may crash because of
11201 the difference induced by this change.
11205 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11207 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11208 (part of SSL_OP_ALL). This option used to disable the
11209 countermeasure against man-in-the-middle protocol-version
11210 rollback in the SSL 2.0 server implementation, which is a bad
11211 idea. ([CVE-2005-2969])
11213 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11214 for Information Security, National Institute of Advanced Industrial
11215 Science and Technology [AIST, Japan)]*
11217 * Minimal support for X9.31 signatures and PSS padding modes. This is
11218 mainly for FIPS compliance and not fully integrated at this stage.
11222 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11223 the exponentiation using a fixed-length exponent. (Otherwise,
11224 the information leaked through timing could expose the secret key
11225 after many signatures; cf. Bleichenbacher's attack on DSA with
11230 * Make a new fixed-window mod_exp implementation the default for
11231 RSA, DSA, and DH private-key operations so that the sequence of
11232 squares and multiplies and the memory access pattern are
11233 independent of the particular secret key. This will mitigate
11234 cache-timing and potential related attacks.
11236 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11237 and this is automatically used by BN_mod_exp_mont() if the new flag
11238 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11239 will use this BN flag for private exponents unless the flag
11240 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11241 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11243 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11245 * Change the client implementation for SSLv23_method() and
11246 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11247 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11248 (Previously, the SSL 2.0 backwards compatible Client Hello
11249 message format would be used even with SSL_OP_NO_SSLv2.)
11253 * Add support for smime-type MIME parameter in S/MIME messages which some
11258 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11259 a threadsafe manner. Modify rsa code to use new function and add calls
11260 to dsa and dh code (which had race conditions before).
11264 * Include the fixed error library code in the C error file definitions
11265 instead of fixing them up at runtime. This keeps the error code
11266 structures constant.
11270 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11272 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11275 * Fixes for newer kerberos headers. NB: the casts are needed because
11276 the 'length' field is signed on one version and unsigned on another
11277 with no (?) obvious way to tell the difference, without these VC++
11278 complains. Also the "definition" of FAR (blank) is no longer included
11279 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11280 some needed definitions.
11284 * Undo Cygwin change.
11288 * Added support for proxy certificates according to RFC 3820.
11289 Because they may be a security thread to unaware applications,
11290 they must be explicitly allowed in run-time. See
11291 docs/HOWTO/proxy_certificates.txt for further information.
11295 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11297 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11298 server and client random values. Previously
11299 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11300 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11302 This change has negligible security impact because:
11304 1. Server and client random values still have 24 bytes of pseudo random
11307 2. Server and client random values are sent in the clear in the initial
11310 3. The master secret is derived using the premaster secret (48 bytes in
11311 size for static RSA ciphersuites) as well as client server and random
11314 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11317 *Stephen Henson, reported by UK NISCC*
11319 * Use Windows randomness collection on Cygwin.
11323 * Fix hang in EGD/PRNGD query when communication socket is closed
11324 prematurely by EGD/PRNGD.
11326 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11328 * Prompt for pass phrases when appropriate for PKCS12 input format.
11332 * Back-port of selected performance improvements from development
11333 branch, as well as improved support for PowerPC platforms.
11337 * Add lots of checks for memory allocation failure, error codes to indicate
11338 failure and freeing up memory if a failure occurs.
11340 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11342 * Add new -passin argument to dgst.
11346 * Perform some character comparisons of different types in X509_NAME_cmp:
11347 this is needed for some certificates that re-encode DNs into UTF8Strings
11348 (in violation of RFC3280) and can't or won't issue name rollover
11353 * Make an explicit check during certificate validation to see that
11354 the CA setting in each certificate on the chain is correct. As a
11355 side effect always do the following basic checks on extensions,
11356 not just when there's an associated purpose to the check:
11358 - if there is an unhandled critical extension (unless the user
11359 has chosen to ignore this fault)
11360 - if the path length has been exceeded (if one is set at all)
11361 - that certain extensions fit the associated purpose (if one has
11366 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11368 * Avoid a race condition when CRLs are checked in a multi threaded
11369 environment. This would happen due to the reordering of the revoked
11370 entries during signature checking and serial number lookup. Now the
11371 encoding is cached and the serial number sort performed under a lock.
11372 Add new STACK function sk_is_sorted().
11376 * Add Delta CRL to the extension code.
11380 * Various fixes to s3_pkt.c so alerts are sent properly.
11382 *David Holmes <d.holmes@f5.com>*
11384 * Reduce the chances of duplicate issuer name and serial numbers (in
11385 violation of RFC3280) using the OpenSSL certificate creation utilities.
11386 This is done by creating a random 64 bit value for the initial serial
11387 number when a serial number file is created or when a self signed
11388 certificate is created using 'openssl req -x509'. The initial serial
11389 number file is created using 'openssl x509 -next_serial' in CA.pl
11390 rather than being initialized to 1.
11394 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11396 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11397 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11399 *Joe Orton, Steve Henson*
11401 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11404 *Joe Orton, Steve Henson*
11406 * Make it possible to have multiple active certificates with the same
11407 subject in the CA index file. This is done only if the keyword
11408 'unique_subject' is set to 'no' in the main CA section (default
11409 if 'CA_default') of the configuration file. The value is saved
11410 with the database itself in a separate index attribute file,
11411 named like the index file with '.attr' appended to the name.
11415 * X509 verify fixes. Disable broken certificate workarounds when
11416 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11417 keyUsage extension present. Don't accept CRLs with unhandled critical
11418 extensions: since verify currently doesn't process CRL extensions this
11419 rejects a CRL with *any* critical extensions. Add new verify error codes
11424 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11425 A clarification of RFC2560 will require the use of OCTET STRINGs and
11426 some implementations cannot handle the current raw format. Since OpenSSL
11427 copies and compares OCSP nonces as opaque blobs without any attempt at
11428 parsing them this should not create any compatibility issues.
11432 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11433 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11434 this HMAC (and other) operations are several times slower than OpenSSL
11439 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11441 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11443 * Use the correct content when signing type "other".
11447 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11449 * Fix various bugs revealed by running the NISCC test suite:
11451 Stop out of bounds reads in the ASN1 code when presented with
11452 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11454 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11456 If verify callback ignores invalid public key errors don't try to check
11457 certificate signature with the NULL public key.
11461 * New -ignore_err option in ocsp application to stop the server
11462 exiting on the first error in a request.
11466 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11467 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11472 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11473 extra data after the compression methods not only for TLS 1.0
11474 but also for SSL 3.0 (as required by the specification).
11476 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11478 * Change X509_certificate_type() to mark the key as exported/exportable
11479 when it's 512 *bits* long, not 512 bytes.
11483 * Change AES_cbc_encrypt() so it outputs exact multiple of
11484 blocks during encryption.
11488 * Various fixes to base64 BIO and non blocking I/O. On write
11489 flushes were not handled properly if the BIO retried. On read
11490 data was not being buffered properly and had various logic bugs.
11491 This also affects blocking I/O when the data being decoded is a
11496 * Various S/MIME bugfixes and compatibility changes:
11497 output correct application/pkcs7 MIME type if
11498 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11499 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11500 of files as .eml work). Correctly handle very long lines in MIME
11505 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11507 * Countermeasure against the Klima-Pokorny-Rosa extension of
11508 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11509 a protocol version number mismatch like a decryption error
11510 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11514 * Turn on RSA blinding by default in the default implementation
11515 to avoid a timing attack. Applications that don't want it can call
11516 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11517 They would be ill-advised to do so in most cases.
11519 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11521 * Change RSA blinding code so that it works when the PRNG is not
11522 seeded (in this case, the secret RSA exponent is abused as
11523 an unpredictable seed -- if it is not unpredictable, there
11524 is no point in blinding anyway). Make RSA blinding thread-safe
11525 by remembering the creator's thread ID in rsa->blinding and
11526 having all other threads use local one-time blinding factors
11527 (this requires more computation than sharing rsa->blinding, but
11528 avoids excessive locking; and if an RSA object is not shared
11529 between threads, blinding will still be very fast).
11533 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11534 ENGINE as defaults for all supported algorithms irrespective of
11535 the 'flags' parameter. 'flags' is now honoured, so applications
11536 should make sure they are passing it correctly.
11540 * Target "mingw" now allows native Windows code to be generated in
11541 the Cygwin environment as well as with the MinGW compiler.
11545 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11547 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11548 via timing by performing a MAC computation even if incorrect
11549 block cipher padding has been found. This is a countermeasure
11550 against active attacks where the attacker has to distinguish
11551 between bad padding and a MAC verification error. ([CVE-2003-0078])
11553 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11554 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11555 Martin Vuagnoux (EPFL, Ilion)*
11557 * Make the no-err option work as intended. The intention with no-err
11558 is not to have the whole error stack handling routines removed from
11559 libcrypto, it's only intended to remove all the function name and
11560 reason texts, thereby removing some of the footprint that may not
11561 be interesting if those errors aren't displayed anyway.
11563 NOTE: it's still possible for any application or module to have its
11564 own set of error texts inserted. The routines are there, just not
11565 used by default when no-err is given.
11569 * Add support for FreeBSD on IA64.
11571 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11573 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11574 Kerberos function mit_des_cbc_cksum(). Before this change,
11575 the value returned by DES_cbc_cksum() was like the one from
11576 mit_des_cbc_cksum(), except the bytes were swapped.
11578 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11580 * Allow an application to disable the automatic SSL chain building.
11581 Before this a rather primitive chain build was always performed in
11582 ssl3_output_cert_chain(): an application had no way to send the
11583 correct chain if the automatic operation produced an incorrect result.
11585 Now the chain builder is disabled if either:
11587 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11589 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11591 The reasoning behind this is that an application would not want the
11592 auto chain building to take place if extra chain certificates are
11593 present and it might also want a means of sending no additional
11594 certificates (for example the chain has two certificates and the
11599 * Add the possibility to build without the ENGINE framework.
11601 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11603 * Under Win32 gmtime() can return NULL: check return value in
11604 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11608 * DSA routines: under certain error conditions uninitialized BN objects
11609 could be freed. Solution: make sure initialization is performed early
11610 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11611 Nils Larsch <nla@trustcenter.de> via PR#459)
11615 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11616 checked on reconnect on the client side, therefore session resumption
11617 could still fail with a "ssl session id is different" error. This
11618 behaviour is masked when SSL_OP_ALL is used due to
11619 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11620 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11621 followup to PR #377.
11625 * IA-32 assembler support enhancements: unified ELF targets, support
11626 for SCO/Caldera platforms, fix for Cygwin shared build.
11630 * Add support for FreeBSD on sparc64. As a consequence, support for
11631 FreeBSD on non-x86 processors is separate from x86 processors on
11632 the config script, much like the NetBSD support.
11634 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11636 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11638 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11641 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11642 code (06) was taken as the first octet of the session ID and the last
11643 octet was ignored consequently. As a result SSLv2 client side session
11644 caching could not have worked due to the session ID mismatch between
11646 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11651 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11652 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11657 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11658 seems that in spite of existing for more than a year, many application
11659 author have done nothing to provide the necessary callbacks, which
11660 means that this particular engine will not work properly anywhere.
11661 This is a very unfortunate situation which forces us, in the name
11662 of usability, to give the hw_ncipher.c a static lock, which is part
11664 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11665 appear in 0.9.8 or later. We EXPECT application authors to have
11666 dealt properly with this when 0.9.8 is released (unless we actually
11667 make such changes in the libcrypto locking code that changes will
11668 have to be made anyway).
11672 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11673 octets have been read, EOF or an error occurs. Without this change
11674 some truncated ASN1 structures will not produce an error.
11678 * Disable Heimdal support, since it hasn't been fully implemented.
11679 Still give the possibility to force the use of Heimdal, but with
11680 warnings and a request that patches get sent to openssl-dev.
11684 * Add the VC-CE target, introduce the WINCE sysname, and add
11685 INSTALL.WCE and appropriate conditionals to make it build.
11687 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11689 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11690 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11691 edit numbers of the version.
11693 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11695 * Introduce safe string copy and catenation functions
11696 (BUF_strlcpy() and BUF_strlcat()).
11698 *Ben Laurie (CHATS) and Richard Levitte*
11700 * Avoid using fixed-size buffers for one-line DNs.
11702 *Ben Laurie (CHATS)*
11704 * Add BUF_MEM_grow_clean() to avoid information leakage when
11705 resizing buffers containing secrets, and use where appropriate.
11707 *Ben Laurie (CHATS)*
11709 * Avoid using fixed size buffers for configuration file location.
11711 *Ben Laurie (CHATS)*
11713 * Avoid filename truncation for various CA files.
11715 *Ben Laurie (CHATS)*
11717 * Use sizeof in preference to magic numbers.
11719 *Ben Laurie (CHATS)*
11721 * Avoid filename truncation in cert requests.
11723 *Ben Laurie (CHATS)*
11725 * Add assertions to check for (supposedly impossible) buffer
11728 *Ben Laurie (CHATS)*
11730 * Don't cache truncated DNS entries in the local cache (this could
11731 potentially lead to a spoofing attack).
11733 *Ben Laurie (CHATS)*
11735 * Fix various buffers to be large enough for hex/decimal
11736 representations in a platform independent manner.
11738 *Ben Laurie (CHATS)*
11740 * Add CRYPTO_realloc_clean() to avoid information leakage when
11741 resizing buffers containing secrets, and use where appropriate.
11743 *Ben Laurie (CHATS)*
11745 * Add BIO_indent() to avoid much slightly worrying code to do
11748 *Ben Laurie (CHATS)*
11750 * Convert sprintf()/BIO_puts() to BIO_printf().
11752 *Ben Laurie (CHATS)*
11754 * buffer_gets() could terminate with the buffer only half
11757 *Ben Laurie (CHATS)*
11759 * Add assertions to prevent user-supplied crypto functions from
11760 overflowing internal buffers by having large block sizes, etc.
11762 *Ben Laurie (CHATS)*
11764 * New OPENSSL_assert() macro (similar to assert(), but enabled
11767 *Ben Laurie (CHATS)*
11769 * Eliminate unused copy of key in RC4.
11771 *Ben Laurie (CHATS)*
11773 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11775 *Ben Laurie (CHATS)*
11777 * Fix off-by-one error in EGD path.
11779 *Ben Laurie (CHATS)*
11781 * If RANDFILE path is too long, ignore instead of truncating.
11783 *Ben Laurie (CHATS)*
11785 * Eliminate unused and incorrectly sized X.509 structure
11788 *Ben Laurie (CHATS)*
11790 * Eliminate unused and dangerous function knumber().
11792 *Ben Laurie (CHATS)*
11794 * Eliminate unused and dangerous structure, KSSL_ERR.
11796 *Ben Laurie (CHATS)*
11798 * Protect against overlong session ID context length in an encoded
11799 session object. Since these are local, this does not appear to be
11802 *Ben Laurie (CHATS)*
11804 * Change from security patch (see 0.9.6e below) that did not affect
11805 the 0.9.6 release series:
11807 Remote buffer overflow in SSL3 protocol - an attacker could
11808 supply an oversized master key in Kerberos-enabled versions.
11811 *Ben Laurie (CHATS)*
11813 * Change the SSL kerb5 codes to match RFC 2712.
11817 * Make -nameopt work fully for req and add -reqopt switch.
11819 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11821 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11823 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11825 * Make sure tests can be performed even if the corresponding algorithms
11826 have been removed entirely. This was also the last step to make
11827 OpenSSL compilable with DJGPP under all reasonable conditions.
11829 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11831 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11832 to allow version independent disabling of normally unselected ciphers,
11833 which may be activated as a side-effect of selecting a single cipher.
11835 (E.g., cipher list string "RSA" enables ciphersuites that are left
11836 out of "ALL" because they do not provide symmetric encryption.
11837 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11839 *Lutz Jaenicke, Bodo Moeller*
11841 * Add appropriate support for separate platform-dependent build
11842 directories. The recommended way to make a platform-dependent
11843 build directory is the following (tested on Linux), maybe with
11846 # Place yourself outside of the OpenSSL source tree. In
11847 # this example, the environment variable OPENSSL_SOURCE
11848 # is assumed to contain the absolute OpenSSL source directory.
11849 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11850 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11851 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11852 mkdir -p `dirname $F`
11853 ln -s $OPENSSL_SOURCE/$F $F
11856 To be absolutely sure not to disturb the source tree, a "make clean"
11857 is a good thing. If it isn't successful, don't worry about it,
11858 it probably means the source directory is very clean.
11862 * Make sure any ENGINE control commands make local copies of string
11863 pointers passed to them whenever necessary. Otherwise it is possible
11864 the caller may have overwritten (or deallocated) the original string
11865 data when a later ENGINE operation tries to use the stored values.
11867 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11869 * Improve diagnostics in file reading and command-line digests.
11871 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11873 * Add AES modes CFB and OFB to the object database. Correct an
11874 error in AES-CFB decryption.
11878 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11879 allows existing EVP_CIPHER_CTX structures to be reused after
11880 calling `EVP_*Final()`. This behaviour is used by encryption
11881 BIOs and some applications. This has the side effect that
11882 applications must explicitly clean up cipher contexts with
11883 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11887 * Check the values of dna and dnb in bn_mul_recursive before calling
11888 bn_mul_comba (a non zero value means the a or b arrays do not contain
11889 n2 elements) and fallback to bn_mul_normal if either is not zero.
11893 * Fix escaping of non-ASCII characters when using the -subj option
11894 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11898 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11899 form for "surname", serialNumber has no short form.
11900 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11901 therefore remove "mail" short name for "internet 7".
11902 The OID for unique identifiers in X509 certificates is
11903 x500UniqueIdentifier, not uniqueIdentifier.
11904 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11908 * Add an "init" command to the ENGINE config module and auto initialize
11909 ENGINEs. Without any "init" command the ENGINE will be initialized
11910 after all ctrl commands have been executed on it. If init=1 the
11911 ENGINE is initialized at that point (ctrls before that point are run
11912 on the uninitialized ENGINE and after on the initialized one). If
11913 init=0 then the ENGINE will not be initialized at all.
11917 * Fix the 'app_verify_callback' interface so that the user-defined
11918 argument is actually passed to the callback: In the
11919 SSL_CTX_set_cert_verify_callback() prototype, the callback
11920 declaration has been changed from
11923 int (*cb)(X509_STORE_CTX *,void *);
11924 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11925 i=s->ctx->app_verify_callback(&ctx)
11926 has been changed into
11927 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11929 To update applications using SSL_CTX_set_cert_verify_callback(),
11930 a dummy argument can be added to their callback functions.
11932 *D. K. Smetters <smetters@parc.xerox.com>*
11934 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11936 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11938 * Add and OPENSSL_LOAD_CONF define which will cause
11939 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11940 This allows older applications to transparently support certain
11941 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11942 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11943 load the config file and OPENSSL_add_all_algorithms_conf() which will
11944 always load it have also been added.
11948 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11949 Adjust NIDs and EVP layer.
11951 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11953 * Config modules support in openssl utility.
11955 Most commands now load modules from the config file,
11956 though in a few (such as version) this isn't done
11957 because it couldn't be used for anything.
11959 In the case of ca and req the config file used is
11960 the same as the utility itself: that is the -config
11961 command line option can be used to specify an
11966 * Move default behaviour from OPENSSL_config(). If appname is NULL
11967 use "openssl_conf" if filename is NULL use default openssl config file.
11971 * Add an argument to OPENSSL_config() to allow the use of an alternative
11972 config section name. Add a new flag to tolerate a missing config file
11973 and move code to CONF_modules_load_file().
11977 * Support for crypto accelerator cards from Accelerated Encryption
11978 Processing, www.aep.ie. (Use engine 'aep')
11979 The support was copied from 0.9.6c [engine] and adapted/corrected
11980 to work with the new engine framework.
11982 *AEP Inc. and Richard Levitte*
11984 * Support for SureWare crypto accelerator cards from Baltimore
11985 Technologies. (Use engine 'sureware')
11986 The support was copied from 0.9.6c [engine] and adapted
11987 to work with the new engine framework.
11991 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11992 make the newer ENGINE framework commands for the CHIL engine work.
11994 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11996 * Make it possible to produce shared libraries on ReliantUNIX.
11998 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12000 * Add the configuration target debug-linux-ppro.
12001 Make 'openssl rsa' use the general key loading routines
12002 implemented in `apps.c`, and make those routines able to
12003 handle the key format FORMAT_NETSCAPE and the variant
12006 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12008 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12010 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12012 * Add -keyform to rsautl, and document -engine.
12014 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12016 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12017 BIO_R_NO_SUCH_FILE error code rather than the generic
12018 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12022 * Add new functions
12023 ERR_peek_last_error
12024 ERR_peek_last_error_line
12025 ERR_peek_last_error_line_data.
12026 These are similar to
12028 ERR_peek_error_line
12029 ERR_peek_error_line_data,
12030 but report on the latest error recorded rather than the first one
12031 still in the error queue.
12033 *Ben Laurie, Bodo Moeller*
12035 * default_algorithms option in ENGINE config module. This allows things
12037 default_algorithms = ALL
12038 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12042 * Preliminary ENGINE config module.
12046 * New experimental application configuration code.
12050 * Change the AES code to follow the same name structure as all other
12051 symmetric ciphers, and behave the same way. Move everything to
12052 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12054 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12056 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12058 *Ben Laurie and Theo de Raadt*
12060 * Add option to output public keys in req command.
12062 *Massimiliano Pala madwolf@openca.org*
12064 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12065 (up to about 10% better than before for P-192 and P-224).
12069 * New functions/macros
12071 SSL_CTX_set_msg_callback(ctx, cb)
12072 SSL_CTX_set_msg_callback_arg(ctx, arg)
12073 SSL_set_msg_callback(ssl, cb)
12074 SSL_set_msg_callback_arg(ssl, arg)
12076 to request calling a callback function
12078 void cb(int write_p, int version, int content_type,
12079 const void *buf, size_t len, SSL *ssl, void *arg)
12081 whenever a protocol message has been completely received
12082 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12083 protocol version according to which the SSL library interprets
12084 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12085 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12086 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12087 specification (change_cipher_spec(20), alert(21), handshake(22)).
12088 'buf' and 'len' point to the actual message, 'ssl' to the
12089 SSL object, and 'arg' is the application-defined value set by
12090 SSL[_CTX]_set_msg_callback_arg().
12092 'openssl s_client' and 'openssl s_server' have new '-msg' options
12093 to enable a callback that displays all protocol messages.
12097 * Change the shared library support so shared libraries are built as
12098 soon as the corresponding static library is finished, and thereby get
12099 openssl and the test programs linked against the shared library.
12100 This still only happens when the keyword "shard" has been given to
12101 the configuration scripts.
12103 NOTE: shared library support is still an experimental thing, and
12104 backward binary compatibility is still not guaranteed.
12106 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12108 * Add support for Subject Information Access extension.
12110 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12112 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12113 additional bytes when new memory had to be allocated, not just
12114 when reusing an existing buffer.
12118 * New command line and configuration option 'utf8' for the req command.
12119 This allows field values to be specified as UTF8 strings.
12123 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12124 runs for the former and machine-readable output for the latter.
12128 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12129 of the e-mail address in the DN (i.e., it will go into a certificate
12130 extension only). The new configuration file option 'email_in_dn = no'
12131 has the same effect.
12133 *Massimiliano Pala madwolf@openca.org*
12135 * Change all functions with names starting with `des_` to be starting
12136 with `DES_` instead. Add wrappers that are compatible with libdes,
12137 but are named `_ossl_old_des_*`. Finally, add macros that map the
12138 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12139 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12140 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12143 Since we provide two compatibility mappings, the user needs to
12144 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12145 compatibility is desired. The default (i.e., when that macro
12146 isn't defined) is OpenSSL 0.9.6c compatibility.
12148 There are also macros that enable and disable the support of old
12149 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12150 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12151 are defined, the default will apply: to support the old des routines.
12153 In either case, one must include openssl/des.h to get the correct
12154 definitions. Do not try to just include openssl/des_old.h, that
12157 NOTE: This is a major break of an old API into a new one. Software
12158 authors are encouraged to switch to the `DES_` style functions. Some
12159 time in the future, des_old.h and the libdes compatibility functions
12160 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12161 default), and then completely removed.
12165 * Test for certificates which contain unsupported critical extensions.
12166 If such a certificate is found during a verify operation it is
12167 rejected by default: this behaviour can be overridden by either
12168 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12169 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12170 X509_supported_extension() has also been added which returns 1 if a
12171 particular extension is supported.
12175 * Modify the behaviour of EVP cipher functions in similar way to digests
12176 to retain compatibility with existing code.
12180 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12181 compatibility with existing code. In particular the 'ctx' parameter does
12182 not have to be to be initialized before the call to EVP_DigestInit() and
12183 it is tidied up after a call to EVP_DigestFinal(). New function
12184 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12185 EVP_MD_CTX_copy() changed to not require the destination to be
12186 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12187 requires the destination to be valid.
12189 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12190 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12194 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12195 so that complete 'Handshake' protocol structures are kept in memory
12196 instead of overwriting 'msg_type' and 'length' with 'body' data.
12200 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12202 *Massimo Santin via Richard Levitte*
12204 * Major restructuring to the underlying ENGINE code. This includes
12205 reduction of linker bloat, separation of pure "ENGINE" manipulation
12206 (initialisation, etc) from functionality dealing with implementations
12207 of specific crypto interfaces. This change also introduces integrated
12208 support for symmetric ciphers and digest implementations - so ENGINEs
12209 can now accelerate these by providing EVP_CIPHER and EVP_MD
12210 implementations of their own. This is detailed in
12211 [crypto/engine/README.md](crypto/engine/README.md)
12212 as it couldn't be adequately described here. However, there are a few
12213 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12214 were changed in the original introduction of ENGINE code have now
12215 reverted back - the hooking from this code to ENGINE is now a good
12216 deal more passive and at run-time, operations deal directly with
12217 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12218 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12219 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12220 they were not being used by the framework as there is no concept of a
12221 BIGNUM_METHOD and they could not be generalised to the new
12222 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12223 ENGINE_cpy() has been removed as it cannot be consistently defined in
12228 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12232 * Change mkdef.pl to sort symbols that get the same entry number,
12233 and make sure the automatically generated functions `ERR_load_*`
12234 become part of libeay.num as well.
12238 * New function SSL_renegotiate_pending(). This returns true once
12239 renegotiation has been requested (either SSL_renegotiate() call
12240 or HelloRequest/ClientHello received from the peer) and becomes
12241 false once a handshake has been completed.
12242 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12243 sends a HelloRequest, but does not ensure that a handshake takes
12244 place. SSL_renegotiate_pending() is useful for checking if the
12245 client has followed the request.)
12249 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12250 By default, clients may request session resumption even during
12251 renegotiation (if session ID contexts permit); with this option,
12252 session resumption is possible only in the first handshake.
12254 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12255 more bits available for options that should not be part of
12256 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12260 * Add some demos for certificate and certificate request creation.
12264 * Make maximum certificate chain size accepted from the peer application
12265 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12266 "Douglas E. Engert" <deengert@anl.gov>.
12270 * Add support for shared libraries for Unixware-7
12271 (Boyd Lynn Gerber <gerberb@zenez.com>).
12275 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12276 be done prior to destruction. Use this to unload error strings from
12277 ENGINEs that load their own error strings. NB: This adds two new API
12278 functions to "get" and "set" this destroy handler in an ENGINE.
12282 * Alter all existing ENGINE implementations (except "openssl" and
12283 "openbsd") to dynamically instantiate their own error strings. This
12284 makes them more flexible to be built both as statically-linked ENGINEs
12285 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12286 Also, add stub code to each that makes building them as self-contained
12287 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12291 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12292 implementations into applications that are completely implemented in
12293 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12294 commands that can be used to configure what shared-library to load and
12295 to control aspects of the way it is handled. Also, made an update to
12296 the [README-Engine.md](README-Engine.md) file
12297 that brings its information up-to-date and
12298 provides some information and instructions on the "dynamic" ENGINE
12299 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12303 * Make it possible to unload ranges of ERR strings with a new
12304 "ERR_unload_strings" function.
12308 * Add a copy() function to EVP_MD.
12312 * Make EVP_MD routines take a context pointer instead of just the
12313 md_data void pointer.
12317 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12318 that the digest can only process a single chunk of data
12319 (typically because it is provided by a piece of
12320 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12321 is only going to provide a single chunk of data, and hence the
12322 framework needn't accumulate the data for oneshot drivers.
12326 * As with "ERR", make it possible to replace the underlying "ex_data"
12327 functions. This change also alters the storage and management of global
12328 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12329 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12330 index counters. The API functions that use this state have been changed
12331 to take a "class_index" rather than pointers to the class's local STACK
12332 and counter, and there is now an API function to dynamically create new
12333 classes. This centralisation allows us to (a) plug a lot of the
12334 thread-safety problems that existed, and (b) makes it possible to clean
12335 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12336 such data would previously have always leaked in application code and
12337 workarounds were in place to make the memory debugging turn a blind eye
12338 to it. Application code that doesn't use this new function will still
12339 leak as before, but their memory debugging output will announce it now
12340 rather than letting it slide.
12342 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12343 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12344 has a return value to indicate success or failure.
12348 * Make it possible to replace the underlying "ERR" functions such that the
12349 global state (2 LHASH tables and 2 locks) is only used by the "default"
12350 implementation. This change also adds two functions to "get" and "set"
12351 the implementation prior to it being automatically set the first time
12352 any other ERR function takes place. Ie. an application can call "get",
12353 pass the return value to a module it has just loaded, and that module
12354 can call its own "set" function using that value. This means the
12355 module's "ERR" operations will use (and modify) the error state in the
12356 application and not in its own statically linked copy of OpenSSL code.
12360 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12361 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12362 the operation, and provides a more encapsulated way for external code
12363 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12364 to use these functions rather than manually incrementing the counts.
12366 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12370 * Add EVP test program.
12374 * Add symmetric cipher support to ENGINE. Expect the API to change!
12378 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12379 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12380 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12381 These allow a CRL to be built without having to access X509_CRL fields
12382 directly. Modify 'ca' application to use new functions.
12386 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12387 bug workarounds. Rollback attack detection is a security feature.
12388 The problem will only arise on OpenSSL servers when TLSv1 is not
12389 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12390 Software authors not wanting to support TLSv1 will have special reasons
12391 for their choice and can explicitly enable this option.
12393 *Bodo Moeller, Lutz Jaenicke*
12395 * Rationalise EVP so it can be extended: don't include a union of
12396 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12397 (similar to those existing for EVP_CIPHER_CTX).
12402 EVP_MD_CTX_init(&md); /* new function call */
12403 EVP_DigestInit(&md, EVP_sha1());
12404 EVP_DigestUpdate(&md, in, len);
12405 EVP_DigestFinal(&md, out, NULL);
12406 EVP_MD_CTX_cleanup(&md); /* new function call */
12410 * Make DES key schedule conform to the usual scheme, as well as
12411 correcting its structure. This means that calls to DES functions
12412 now have to pass a pointer to a des_key_schedule instead of a
12413 plain des_key_schedule (which was actually always a pointer
12416 des_key_schedule ks;
12418 des_set_key_checked(..., &ks);
12419 des_ncbc_encrypt(..., &ks, ...);
12421 (Note that a later change renames 'des_...' into 'DES_...'.)
12425 * Initial reduction of linker bloat: the use of some functions, such as
12426 PEM causes large amounts of unused functions to be linked in due to
12427 poor organisation. For example pem_all.c contains every PEM function
12428 which has a knock on effect of linking in large amounts of (unused)
12429 ASN1 code. Grouping together similar functions and splitting unrelated
12430 functions prevents this.
12434 * Cleanup of EVP macros.
12438 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12439 correct `_ecb suffix`.
12443 * Add initial OCSP responder support to ocsp application. The
12444 revocation information is handled using the text based index
12445 use by the ca application. The responder can either handle
12446 requests generated internally, supplied in files (for example
12447 via a CGI script) or using an internal minimal server.
12451 * Add configuration choices to get zlib compression for TLS.
12455 * Changes to Kerberos SSL for RFC 2712 compliance:
12456 1. Implemented real KerberosWrapper, instead of just using
12457 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12458 2. Implemented optional authenticator field of KerberosWrapper.
12460 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12461 and authenticator structs; see crypto/krb5/.
12463 Generalized Kerberos calls to support multiple Kerberos libraries.
12464 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12465 via Richard Levitte*
12467 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12468 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12469 values for each of the key sizes rather than having just
12470 parameters (and 'speed' generating keys each time).
12474 * Speed up EVP routines.
12477 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12478 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12479 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12480 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12482 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12483 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12484 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12487 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12489 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12493 * Added the OS2-EMX target.
12495 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12497 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12498 New functions to support `NCONF` routines in extension code.
12499 New function `CONF_set_nconf()`
12500 to allow functions which take an `NCONF` to also handle the old `LHASH`
12501 structure: this means that the old `CONF` compatible routines can be
12502 retained (in particular w.rt. extensions) without having to duplicate the
12503 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12507 * Enhance the general user interface with mechanisms for inner control
12508 and with possibilities to have yes/no kind of prompts.
12512 * Change all calls to low-level digest routines in the library and
12513 applications to use EVP. Add missing calls to HMAC_cleanup() and
12514 don't assume HMAC_CTX can be copied using memcpy().
12516 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12518 * Add the possibility to control engines through control names but with
12519 arbitrary arguments instead of just a string.
12520 Change the key loaders to take a UI_METHOD instead of a callback
12521 function pointer. NOTE: this breaks binary compatibility with earlier
12522 versions of OpenSSL [engine].
12523 Adapt the nCipher code for these new conditions and add a card insertion
12528 * Enhance the general user interface with mechanisms to better support
12529 dialog box interfaces, application-defined prompts, the possibility
12530 to use defaults (for example default passwords from somewhere else)
12531 and interrupts/cancellations.
12535 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12536 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12540 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12541 tidy up some unnecessarily weird code in 'sk_new()').
12543 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12545 * Change the key loading routines for ENGINEs to use the same kind
12546 callback (pem_password_cb) as all other routines that need this
12551 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12552 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12553 than this minimum value is recommended.
12557 * New random seeder for OpenVMS, using the system process statistics
12558 that are easily reachable.
12562 * Windows apparently can't transparently handle global
12563 variables defined in DLLs. Initialisations such as:
12565 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12567 won't compile. This is used by the any applications that need to
12568 declare their own ASN1 modules. This was fixed by adding the option
12569 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12570 needed for static libraries under Win32.
12574 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12575 setting of purpose and trust fields. New X509_STORE trust and
12576 purpose functions and tidy up setting in other SSL functions.
12580 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12581 structure. These are inherited by X509_STORE_CTX when it is
12582 initialised. This allows various defaults to be set in the
12583 X509_STORE structure (such as flags for CRL checking and custom
12584 purpose or trust settings) for functions which only use X509_STORE_CTX
12585 internally such as S/MIME.
12587 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12588 trust settings if they are not set in X509_STORE. This allows X509_STORE
12589 purposes and trust (in S/MIME for example) to override any set by default.
12591 Add command line options for CRL checking to smime, s_client and s_server
12596 * Initial CRL based revocation checking. If the CRL checking flag(s)
12597 are set then the CRL is looked up in the X509_STORE structure and
12598 its validity and signature checked, then if the certificate is found
12599 in the CRL the verify fails with a revoked error.
12601 Various new CRL related callbacks added to X509_STORE_CTX structure.
12603 Command line options added to 'verify' application to support this.
12605 This needs some additional work, such as being able to handle multiple
12606 CRLs with different times, extension based lookup (rather than just
12607 by subject name) and ultimately more complete V2 CRL extension
12612 * Add a general user interface API (crypto/ui/). This is designed
12613 to replace things like des_read_password and friends (backward
12614 compatibility functions using this new API are provided).
12615 The purpose is to remove prompting functions from the DES code
12616 section as well as provide for prompting through dialog boxes in
12617 a window system and the like.
12621 * Add "ex_data" support to ENGINE so implementations can add state at a
12622 per-structure level rather than having to store it globally.
12626 * Make it possible for ENGINE structures to be copied when retrieved by
12627 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12628 This causes the "original" ENGINE structure to act like a template,
12629 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12630 operational state can be localised to each ENGINE structure, despite the
12631 fact they all share the same "methods". New ENGINE structures returned in
12632 this case have no functional references and the return value is the single
12633 structural reference. This matches the single structural reference returned
12634 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12639 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12640 needs to match any other type at all we need to manually clear the
12645 * Changes to the "openssl engine" utility to include;
12646 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12647 about an ENGINE's available control commands.
12648 - executing control commands from command line arguments using the
12649 '-pre' and '-post' switches. '-post' is only used if '-t' is
12650 specified and the ENGINE is successfully initialised. The syntax for
12651 the individual commands are colon-separated, for example;
12652 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12656 * New dynamic control command support for ENGINEs. ENGINEs can now
12657 declare their own commands (numbers), names (strings), descriptions,
12658 and input types for run-time discovery by calling applications. A
12659 subset of these commands are implicitly classed as "executable"
12660 depending on their input type, and only these can be invoked through
12661 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12662 can be based on user input, config files, etc). The distinction is
12663 that "executable" commands cannot return anything other than a boolean
12664 result and can only support numeric or string input, whereas some
12665 discoverable commands may only be for direct use through
12666 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12667 pointers, or other custom uses. The "executable" commands are to
12668 support parameterisations of ENGINE behaviour that can be
12669 unambiguously defined by ENGINEs and used consistently across any
12670 OpenSSL-based application. Commands have been added to all the
12671 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12672 control over shared-library paths without source code alterations.
12676 * Changed all ENGINE implementations to dynamically allocate their
12677 ENGINEs rather than declaring them statically. Apart from this being
12678 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12679 this also allows the implementations to compile without using the
12680 internal engine_int.h header.
12684 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12685 'const' value. Any code that should be able to modify a RAND_METHOD
12686 should already have non-const pointers to it (ie. they should only
12687 modify their own ones).
12691 * Made a variety of little tweaks to the ENGINE code.
12692 - "atalla" and "ubsec" string definitions were moved from header files
12693 to C code. "nuron" string definitions were placed in variables
12694 rather than hard-coded - allowing parameterisation of these values
12695 later on via ctrl() commands.
12696 - Removed unused "#if 0"'d code.
12697 - Fixed engine list iteration code so it uses ENGINE_free() to release
12698 structural references.
12699 - Constified the RAND_METHOD element of ENGINE structures.
12700 - Constified various get/set functions as appropriate and added
12701 missing functions (including a catch-all ENGINE_cpy that duplicates
12702 all ENGINE values onto a new ENGINE except reference counts/state).
12703 - Removed NULL parameter checks in get/set functions. Setting a method
12704 or function to NULL is a way of cancelling out a previously set
12705 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12706 and doesn't justify the extra error symbols and code.
12707 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12708 flags from engine_int.h to engine.h.
12709 - Changed prototypes for ENGINE handler functions (init(), finish(),
12710 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12714 * Implement binary inversion algorithm for BN_mod_inverse in addition
12715 to the algorithm using long division. The binary algorithm can be
12716 used only if the modulus is odd. On 32-bit systems, it is faster
12717 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12718 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12719 up to 450 bits. In 64-bit environments, the binary algorithm
12720 appears to be advantageous for much longer moduli; here we use it
12721 for moduli up to 2048 bits.
12725 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12726 could not support the combine flag in choice fields.
12730 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12731 extensions from a certificate request to the certificate.
12735 * Allow multiple 'certopt' and 'nameopt' options to be separated
12736 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12737 file: this allows the display of the certificate about to be
12738 signed to be customised, to allow certain fields to be included
12739 or excluded and extension details. The old system didn't display
12740 multicharacter strings properly, omitted fields not in the policy
12741 and couldn't display additional details such as extensions.
12745 * Function EC_POINTs_mul for multiple scalar multiplication
12746 of an arbitrary number of elliptic curve points
12747 \sum scalars[i]*points[i],
12748 optionally including the generator defined for the EC_GROUP:
12749 scalar*generator + \sum scalars[i]*points[i].
12751 EC_POINT_mul is a simple wrapper function for the typical case
12752 that the point list has just one item (besides the optional
12757 * First EC_METHODs for curves over GF(p):
12759 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12760 operations and provides various method functions that can also
12761 operate with faster implementations of modular arithmetic.
12763 EC_GFp_mont_method() reuses most functions that are part of
12764 EC_GFp_simple_method, but uses Montgomery arithmetic.
12766 *Bodo Moeller; point addition and point doubling
12767 implementation directly derived from source code provided by
12768 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12770 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12771 crypto/ec/ec_lib.c):
12773 Curves are EC_GROUP objects (with an optional group generator)
12774 based on EC_METHODs that are built into the library.
12776 Points are EC_POINT objects based on EC_GROUP objects.
12778 Most of the framework would be able to handle curves over arbitrary
12779 finite fields, but as there are no obvious types for fields other
12780 than GF(p), some functions are limited to that for now.
12784 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12785 that the file contains a complete HTTP response.
12789 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12790 change the def and num file printf format specifier from "%-40sXXX"
12791 to "%-39s XXX". The latter will always guarantee a space after the
12792 field while the former will cause them to run together if the field
12793 is 40 of more characters long.
12797 * Constify the cipher and digest 'method' functions and structures
12798 and modify related functions to take constant EVP_MD and EVP_CIPHER
12803 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12804 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12808 * Modify `EVP_Digest*()` routines so they now return values. Although the
12809 internal software routines can never fail additional hardware versions
12814 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12816 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12817 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12820 ERR_R_NESTED_ASN1_ERROR
12822 ERR_R_MISSING_ASN1_EOS
12823 were 4 .. 9, conflicting with
12824 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12826 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12827 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12829 Add new error code 'ERR_R_INTERNAL_ERROR'.
12833 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12838 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12839 sets the subject name for a new request or supersedes the
12840 subject name in a given request. Formats that can be parsed are
12841 'CN=Some Name, OU=myOU, C=IT'
12843 'CN=Some Name/OU=myOU/C=IT'.
12845 Add options '-batch' and '-verbose' to 'openssl req'.
12847 *Massimiliano Pala <madwolf@hackmasters.net>*
12849 * Introduce the possibility to access global variables through
12850 functions on platform were that's the best way to handle exporting
12851 global variables in shared libraries. To enable this functionality,
12852 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12853 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12854 is normally done by Configure or something similar).
12856 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12857 in the source file (foo.c) like this:
12859 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12860 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12862 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12863 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12865 OPENSSL_DECLARE_GLOBAL(int,foo);
12866 #define foo OPENSSL_GLOBAL_REF(foo)
12867 OPENSSL_DECLARE_GLOBAL(double,bar);
12868 #define bar OPENSSL_GLOBAL_REF(bar)
12870 The #defines are very important, and therefore so is including the
12871 header file everywhere where the defined globals are used.
12873 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12874 of ASN.1 items, but that structure is a bit different.
12876 The largest change is in util/mkdef.pl which has been enhanced with
12877 better and easier to understand logic to choose which symbols should
12878 go into the Windows .def files as well as a number of fixes and code
12879 cleanup (among others, algorithm keywords are now sorted
12880 lexicographically to avoid constant rewrites).
12884 * In BN_div() keep a copy of the sign of 'num' before writing the
12885 result to 'rm' because if rm==num the value will be overwritten
12886 and produce the wrong result if 'num' is negative: this caused
12887 problems with BN_mod() and BN_nnmod().
12891 * Function OCSP_request_verify(). This checks the signature on an
12892 OCSP request and verifies the signer certificate. The signer
12893 certificate is just checked for a generic purpose and OCSP request
12898 * Add OCSP_check_validity() function to check the validity of OCSP
12899 responses. OCSP responses are prepared in real time and may only
12900 be a few seconds old. Simply checking that the current time lies
12901 between thisUpdate and nextUpdate max reject otherwise valid responses
12902 caused by either OCSP responder or client clock inaccuracy. Instead
12903 we allow thisUpdate and nextUpdate to fall within a certain period of
12904 the current time. The age of the response can also optionally be
12905 checked. Two new options -validity_period and -status_age added to
12910 * If signature or public key algorithm is unrecognized print out its
12911 OID rather that just UNKNOWN.
12915 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12916 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12917 ID to be generated from the issuer certificate alone which can then be
12918 passed to OCSP_id_issuer_cmp().
12922 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12923 ASN1 modules to export functions returning ASN1_ITEM pointers
12924 instead of the ASN1_ITEM structures themselves. This adds several
12925 new macros which allow the underlying ASN1 function/structure to
12926 be accessed transparently. As a result code should not use ASN1_ITEM
12927 references directly (such as &X509_it) but instead use the relevant
12928 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12929 use of the new ASN1 code on platforms where exporting structures
12930 is problematical (for example in shared libraries) but exporting
12931 functions returning pointers to structures is not.
12935 * Add support for overriding the generation of SSL/TLS session IDs.
12936 These callbacks can be registered either in an SSL_CTX or per SSL.
12937 The purpose of this is to allow applications to control, if they wish,
12938 the arbitrary values chosen for use as session IDs, particularly as it
12939 can be useful for session caching in multiple-server environments. A
12940 command-line switch for testing this (and any client code that wishes
12941 to use such a feature) has been added to "s_server".
12943 *Geoff Thorpe, Lutz Jaenicke*
12945 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12946 of the form `#if defined(...) || defined(...) || ...` and
12947 `#if !defined(...) && !defined(...) && ...`. This also avoids
12948 the growing number of special cases it was previously handling.
12952 * Make all configuration macros available for application by making
12953 sure they are available in opensslconf.h, by giving them names starting
12954 with `OPENSSL_` to avoid conflicts with other packages and by making
12955 sure e_os2.h will cover all platform-specific cases together with
12957 Additionally, it is now possible to define configuration/platform-
12958 specific names (called "system identities"). In the C code, these
12959 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12960 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12961 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12966 * New option -set_serial to 'req' and 'x509' this allows the serial
12967 number to use to be specified on the command line. Previously self
12968 signed certificates were hard coded with serial number 0 and the
12969 CA options of 'x509' had to use a serial number in a file which was
12974 * New options to 'ca' utility to support V2 CRL entry extensions.
12975 Currently CRL reason, invalidity date and hold instruction are
12976 supported. Add new CRL extensions to V3 code and some new objects.
12980 * New function EVP_CIPHER_CTX_set_padding() this is used to
12981 disable standard block padding (aka PKCS#5 padding) in the EVP
12982 API, which was previously mandatory. This means that the data is
12983 not padded in any way and so the total length much be a multiple
12984 of the block size, otherwise an error occurs.
12988 * Initial (incomplete) OCSP SSL support.
12992 * New function OCSP_parse_url(). This splits up a URL into its host,
12993 port and path components: primarily to parse OCSP URLs. New -url
12994 option to ocsp utility.
12998 * New nonce behavior. The return value of OCSP_check_nonce() now
12999 reflects the various checks performed. Applications can decide
13000 whether to tolerate certain situations such as an absent nonce
13001 in a response when one was present in a request: the ocsp application
13002 just prints out a warning. New function OCSP_add1_basic_nonce()
13003 this is to allow responders to include a nonce in a response even if
13004 the request is nonce-less.
13008 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13009 skipped when using openssl x509 multiple times on a single input file,
13010 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13014 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13015 set string type: to handle setting ASN1_TIME structures. Fix ca
13016 utility to correctly initialize revocation date of CRLs.
13020 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13021 the clients preferred ciphersuites and rather use its own preferences.
13022 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13023 Internet Explorer by ensuring unchanged hash method during stepup.
13024 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13028 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13029 to aes and add a new 'exist' option to print out symbols that don't
13034 * Additional options to ocsp utility to allow flags to be set and
13035 additional certificates supplied.
13039 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13040 OCSP client a number of certificate to only verify the response
13045 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13046 handle the new API. Currently only ECB, CBC modes supported. Add new
13049 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13050 Encryption Standard (AES) Ciphersuites for Transport Layer
13051 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13052 not enabled by default and were not part of the "ALL" ciphersuite
13053 alias because they were not yet official; they could be
13054 explicitly requested by specifying the "AESdraft" ciphersuite
13055 group alias. In the final release of OpenSSL 0.9.7, the group
13056 alias is called "AES" and is part of "ALL".)
13058 *Ben Laurie, Steve Henson, Bodo Moeller*
13060 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13061 request to response.
13065 * Functions for OCSP responders. OCSP_request_onereq_count(),
13066 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13067 extract information from a certificate request. OCSP_response_create()
13068 creates a response and optionally adds a basic response structure.
13069 OCSP_basic_add1_status() adds a complete single response to a basic
13070 response and returns the OCSP_SINGLERESP structure just added (to allow
13071 extensions to be included for example). OCSP_basic_add1_cert() adds a
13072 certificate to a basic response and OCSP_basic_sign() signs a basic
13073 response with various flags. New helper functions ASN1_TIME_check()
13074 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13075 (converts ASN1_TIME to GeneralizedTime).
13079 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13080 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13081 structure from a certificate. X509_pubkey_digest() digests the public_key
13082 contents: this is used in various key identifiers.
13086 * Make sk_sort() tolerate a NULL argument.
13088 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13090 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13091 passed by the function are trusted implicitly. If any of them signed the
13092 response then it is assumed to be valid and is not verified.
13096 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13097 to data. This was previously part of the PKCS7 ASN1 code. This
13098 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13099 *Steve Henson, reported by Kenneth R. Robinette
13100 <support@securenetterm.com>*
13102 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13103 routines: without these tracing memory leaks is very painful.
13104 Fix leaks in PKCS12 and PKCS7 routines.
13108 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13109 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13110 effectively meant GeneralizedTime would never be used. Now it
13111 is initialised to -1 but X509_time_adj() now has to check the value
13112 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13113 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13114 *Steve Henson, reported by Kenneth R. Robinette
13115 <support@securenetterm.com>*
13117 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13118 result in a zero length in the ASN1_INTEGER structure which was
13119 not consistent with the structure when d2i_ASN1_INTEGER() was used
13120 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13121 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13122 where it did not print out a minus for negative ASN1_INTEGER.
13126 * Add summary printout to ocsp utility. The various functions which
13127 convert status values to strings have been renamed to:
13128 OCSP_response_status_str(), OCSP_cert_status_str() and
13129 OCSP_crl_reason_str() and are no longer static. New options
13130 to verify nonce values and to disable verification. OCSP response
13131 printout format cleaned up.
13135 * Add additional OCSP certificate checks. These are those specified
13136 in RFC2560. This consists of two separate checks: the CA of the
13137 certificate being checked must either be the OCSP signer certificate
13138 or the issuer of the OCSP signer certificate. In the latter case the
13139 OCSP signer certificate must contain the OCSP signing extended key
13140 usage. This check is performed by attempting to match the OCSP
13141 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13142 in the OCSP_CERTID structures of the response.
13146 * Initial OCSP certificate verification added to OCSP_basic_verify()
13147 and related routines. This uses the standard OpenSSL certificate
13148 verify routines to perform initial checks (just CA validity) and
13149 to obtain the certificate chain. Then additional checks will be
13150 performed on the chain. Currently the root CA is checked to see
13151 if it is explicitly trusted for OCSP signing. This is used to set
13152 a root CA as a global signing root: that is any certificate that
13153 chains to that CA is an acceptable OCSP signing certificate.
13157 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13158 extensions from a separate configuration file.
13159 As when reading extensions from the main configuration file,
13160 the '-extensions ...' option may be used for specifying the
13163 *Massimiliano Pala <madwolf@comune.modena.it>*
13165 * New OCSP utility. Allows OCSP requests to be generated or
13166 read. The request can be sent to a responder and the output
13167 parsed, outputted or printed in text form. Not complete yet:
13168 still needs to check the OCSP response validity.
13172 * New subcommands for 'openssl ca':
13173 `openssl ca -status <serial>` prints the status of the cert with
13174 the given serial number (according to the index file).
13175 `openssl ca -updatedb` updates the expiry status of certificates
13178 *Massimiliano Pala <madwolf@comune.modena.it>*
13180 * New '-newreq-nodes' command option to CA.pl. This is like
13181 '-newreq', but calls 'openssl req' with the '-nodes' option
13182 so that the resulting key is not encrypted.
13184 *Damien Miller <djm@mindrot.org>*
13186 * New configuration for the GNU Hurd.
13188 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13190 * Initial code to implement OCSP basic response verify. This
13191 is currently incomplete. Currently just finds the signer's
13192 certificate and verifies the signature on the response.
13196 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13197 value of OPENSSLDIR. This is available via the new '-d' option
13198 to 'openssl version', and is also included in 'openssl version -a'.
13202 * Allowing defining memory allocation callbacks that will be given
13203 file name and line number information in additional arguments
13204 (a `const char*` and an int). The basic functionality remains, as
13205 well as the original possibility to just replace malloc(),
13206 realloc() and free() by functions that do not know about these
13207 additional arguments. To register and find out the current
13208 settings for extended allocation functions, the following
13209 functions are provided:
13211 CRYPTO_set_mem_ex_functions
13212 CRYPTO_set_locked_mem_ex_functions
13213 CRYPTO_get_mem_ex_functions
13214 CRYPTO_get_locked_mem_ex_functions
13216 These work the same way as CRYPTO_set_mem_functions and friends.
13217 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13218 extended allocation function is enabled.
13219 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13220 a conventional allocation function is enabled.
13222 *Richard Levitte, Bodo Moeller*
13224 * Finish off removing the remaining LHASH function pointer casts.
13225 There should no longer be any prototype-casting required when using
13226 the LHASH abstraction, and any casts that remain are "bugs". See
13227 the callback types and macros at the head of lhash.h for details
13228 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13232 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13233 If /dev/[u]random devices are not available or do not return enough
13234 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13236 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13237 /etc/entropy will be queried once each in this sequence, querying stops
13238 when enough entropy was collected without querying more sockets.
13242 * Change the Unix RAND_poll() variant to be able to poll several
13243 random devices, as specified by DEVRANDOM, until a sufficient amount
13244 of data has been collected. We spend at most 10 ms on each file
13245 (select timeout) and read in non-blocking mode. DEVRANDOM now
13246 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13247 (previously it was just the string "/dev/urandom"), so on typical
13248 platforms the 10 ms delay will never occur.
13249 Also separate out the Unix variant to its own file, rand_unix.c.
13250 For VMS, there's a currently-empty rand_vms.c.
13254 * Move OCSP client related routines to ocsp_cl.c. These
13255 provide utility functions which an application needing
13256 to issue a request to an OCSP responder and analyse the
13257 response will typically need: as opposed to those which an
13258 OCSP responder itself would need which will be added later.
13260 OCSP_request_sign() signs an OCSP request with an API similar
13261 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13262 response. OCSP_response_get1_basic() extracts basic response
13263 from response. OCSP_resp_find_status(): finds and extracts status
13264 information from an OCSP_CERTID structure (which will be created
13265 when the request structure is built). These are built from lower
13266 level functions which work on OCSP_SINGLERESP structures but
13267 won't normally be used unless the application wishes to examine
13268 extensions in the OCSP response for example.
13270 Replace nonce routines with a pair of functions.
13271 OCSP_request_add1_nonce() adds a nonce value and optionally
13272 generates a random value. OCSP_check_nonce() checks the
13273 validity of the nonce in an OCSP response.
13277 * Change function OCSP_request_add() to OCSP_request_add0_id().
13278 This doesn't copy the supplied OCSP_CERTID and avoids the
13279 need to free up the newly created id. Change return type
13280 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13281 This can then be used to add extensions to the request.
13282 Deleted OCSP_request_new(), since most of its functionality
13283 is now in OCSP_REQUEST_new() (and the case insensitive name
13284 clash) apart from the ability to set the request name which
13285 will be added elsewhere.
13289 * Update OCSP API. Remove obsolete extensions argument from
13290 various functions. Extensions are now handled using the new
13291 OCSP extension code. New simple OCSP HTTP function which
13292 can be used to send requests and parse the response.
13296 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13297 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13298 uses the special reorder version of SET OF to sort the attributes
13299 and reorder them to match the encoded order. This resolves a long
13300 standing problem: a verify on a PKCS7 structure just after signing
13301 it used to fail because the attribute order did not match the
13302 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13303 it uses the received order. This is necessary to tolerate some broken
13304 software that does not order SET OF. This is handled by encoding
13305 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13306 to produce the required SET OF.
13310 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13311 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13312 files to get correct declarations of the ASN.1 item variables.
13316 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13317 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13318 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13319 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13320 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13321 ASN1_ITEM and no wrapper functions.
13325 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13326 replace the old function pointer based I/O routines. Change most of
13327 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13331 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13332 lines, recognize more "algorithms" that can be deselected, and make
13333 it complain about algorithm deselection that isn't recognised.
13337 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13338 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13339 to use new functions. Add NO_ASN1_OLD which can be set to remove
13340 some old style ASN1 functions: this can be used to determine if old
13341 code will still work when these eventually go away.
13345 * New extension functions for OCSP structures, these follow the
13346 same conventions as certificates and CRLs.
13350 * New function X509V3_add1_i2d(). This automatically encodes and
13351 adds an extension. Its behaviour can be customised with various
13352 flags to append, replace or delete. Various wrappers added for
13353 certificates and CRLs.
13357 * Fix to avoid calling the underlying ASN1 print routine when
13358 an extension cannot be parsed. Correct a typo in the
13359 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13363 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13364 entries for variables.
13368 * Add functionality to `apps/openssl.c` for detecting locking
13369 problems: As the program is single-threaded, all we have
13370 to do is register a locking callback using an array for
13371 storing which locks are currently held by the program.
13375 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13376 SSL_get_ex_data_X509_STORE_idx(), which is used in
13377 ssl_verify_cert_chain() and thus can be called at any time
13378 during TLS/SSL handshakes so that thread-safety is essential.
13379 Unfortunately, the ex_data design is not at all suited
13380 for multi-threaded use, so it probably should be abolished.
13384 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13386 *Broadcom, tweaked and integrated by Geoff Thorpe*
13388 * Move common extension printing code to new function
13389 X509V3_print_extensions(). Reorganise OCSP print routines and
13390 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13394 * New function X509_signature_print() to remove duplication in some
13399 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13400 set (this was treated exactly the same as SET OF previously). This
13401 is used to reorder the STACK representing the structure to match the
13402 encoding. This will be used to get round a problem where a PKCS7
13403 structure which was signed could not be verified because the STACK
13404 order did not reflect the encoded order.
13408 * Reimplement the OCSP ASN1 module using the new code.
13412 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13413 for its ASN1 operations. The old style function pointers still exist
13414 for now but they will eventually go away.
13418 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13419 completely replaces the old ASN1 functionality with a table driven
13420 encoder and decoder which interprets an ASN1_ITEM structure describing
13421 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13422 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13423 has also been converted to the new form.
13427 * Change BN_mod_exp_recp so that negative moduli are tolerated
13428 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13429 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13430 for negative moduli.
13434 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13435 of not touching the result's sign bit.
13439 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13444 * Changed the LHASH code to use prototypes for callbacks, and created
13445 macros to declare and implement thin (optionally static) functions
13446 that provide type-safety and avoid function pointer casting for the
13447 type-specific callbacks.
13451 * Added Kerberos Cipher Suites to be used with TLS, as written in
13453 *Veers Staats <staatsvr@asc.hpc.mil>,
13454 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13456 * Reformat the FAQ so the different questions and answers can be divided
13457 in sections depending on the subject.
13461 * Have the zlib compression code load ZLIB.DLL dynamically under
13466 * New function BN_mod_sqrt for computing square roots modulo a prime
13467 (using the probabilistic Tonelli-Shanks algorithm unless
13468 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13469 be handled deterministically).
13471 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13473 * Make BN_mod_inverse faster by explicitly handling small quotients
13474 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13475 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13479 * New function BN_kronecker.
13483 * Fix BN_gcd so that it works on negative inputs; the result is
13484 positive unless both parameters are zero.
13485 Previously something reasonably close to an infinite loop was
13486 possible because numbers could be growing instead of shrinking
13487 in the implementation of Euclid's algorithm.
13491 * Fix BN_is_word() and BN_is_one() macros to take into account the
13492 sign of the number in question.
13494 Fix BN_is_word(a,w) to work correctly for w == 0.
13496 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13497 because its test if the absolute value of 'a' equals 'w'.
13498 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13499 it exists mostly for use in the implementations of BN_is_zero(),
13500 BN_is_one(), and BN_is_word().
13504 * New function BN_swap.
13508 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13509 the exponentiation functions are more likely to produce reasonable
13510 results on negative inputs.
13514 * Change BN_mod_mul so that the result is always non-negative.
13515 Previously, it could be negative if one of the factors was negative;
13516 I don't think anyone really wanted that behaviour.
13520 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13521 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13522 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13523 and add new functions:
13532 BN_mod_lshift1_quick
13534 BN_mod_lshift_quick
13536 These functions always generate non-negative results.
13538 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13539 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13541 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13542 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13543 be reduced modulo `m`.
13545 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13548 The following entry accidentally appeared in the CHANGES file
13549 distributed with OpenSSL 0.9.7. The modifications described in
13550 it do *not* apply to OpenSSL 0.9.7.
13552 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13553 was actually never needed) and in BN_mul(). The removal in BN_mul()
13554 required a small change in bn_mul_part_recursive() and the addition
13555 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13556 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13557 bn_sub_words() and bn_add_words() except they take arrays with
13563 * In 'openssl passwd', verify passwords read from the terminal
13564 unless the '-salt' option is used (which usually means that
13565 verification would just waste user's time since the resulting
13566 hash is going to be compared with some given password hash)
13567 or the new '-noverify' option is used.
13569 This is an incompatible change, but it does not affect
13570 non-interactive use of 'openssl passwd' (passwords on the command
13571 line, '-stdin' option, '-in ...' option) and thus should not
13572 cause any problems.
13576 * Remove all references to RSAref, since there's no more need for it.
13580 * Make DSO load along a path given through an environment variable
13581 (SHLIB_PATH) with shl_load().
13585 * Constify the ENGINE code as a result of BIGNUM constification.
13586 Also constify the RSA code and most things related to it. In a
13587 few places, most notable in the depth of the ASN.1 code, ugly
13588 casts back to non-const were required (to be solved at a later
13593 * Make it so the openssl application has all engines loaded by default.
13597 * Constify the BIGNUM routines a little more.
13601 * Add the following functions:
13603 ENGINE_load_cswift()
13605 ENGINE_load_atalla()
13606 ENGINE_load_nuron()
13607 ENGINE_load_builtin_engines()
13609 That way, an application can itself choose if external engines that
13610 are built-in in OpenSSL shall ever be used or not. The benefit is
13611 that applications won't have to be linked with libdl or other dso
13612 libraries unless it's really needed.
13614 Changed 'openssl engine' to load all engines on demand.
13615 Changed the engine header files to avoid the duplication of some
13616 declarations (they differed!).
13620 * 'openssl engine' can now list capabilities.
13624 * Better error reporting in 'openssl engine'.
13628 * Never call load_dh_param(NULL) in s_server.
13632 * Add engine application. It can currently list engines by name and
13633 identity, and test if they are actually available.
13637 * Improve RPM specification file by forcing symbolic linking and making
13638 sure the installed documentation is also owned by root.root.
13640 *Damien Miller <djm@mindrot.org>*
13642 * Give the OpenSSL applications more possibilities to make use of
13643 keys (public as well as private) handled by engines.
13647 * Add OCSP code that comes from CertCo.
13651 * Add VMS support for the Rijndael code.
13655 * Added untested support for Nuron crypto accelerator.
13659 * Add support for external cryptographic devices. This code was
13660 previously distributed separately as the "engine" branch.
13662 *Geoff Thorpe, Richard Levitte*
13664 * Rework the filename-translation in the DSO code. It is now possible to
13665 have far greater control over how a "name" is turned into a filename
13666 depending on the operating environment and any oddities about the
13667 different shared library filenames on each system.
13671 * Support threads on FreeBSD-elf in Configure.
13675 * Fix for SHA1 assembly problem with MASM: it produces
13676 warnings about corrupt line number information when assembling
13677 with debugging information. This is caused by the overlapping
13680 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13683 NCONF_get_number() has no error checking at all. As a replacement,
13684 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13685 promoted strongly. The old NCONF_get_number is kept around for
13686 binary backward compatibility.
13687 Make it possible for methods to load from something other than a BIO,
13688 by providing a function pointer that is given a name instead of a BIO.
13689 For example, this could be used to load configuration data from an
13694 * Fix for non blocking accept BIOs. Added new I/O special reason
13695 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13696 with non blocking I/O was not possible because no retry code was
13697 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13702 * Added the beginnings of Rijndael support.
13706 * Fix for bug in DirectoryString mask setting. Add support for
13707 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13708 to allow certificate printing to more controllable, additional
13709 'certopt' option to 'x509' to allow new printing options to be
13714 * Clean old EAY MD5 hack from e_os.h.
13718 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13720 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13721 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13723 *Joe Orton, Steve Henson*
13725 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13727 * Fix additional bug revealed by the NISCC test suite:
13729 Stop bug triggering large recursion when presented with
13730 certain ASN.1 tags ([CVE-2003-0851])
13734 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13736 * Fix various bugs revealed by running the NISCC test suite:
13738 Stop out of bounds reads in the ASN1 code when presented with
13739 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13741 If verify callback ignores invalid public key errors don't try to check
13742 certificate signature with the NULL public key.
13746 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13747 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13752 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13753 extra data after the compression methods not only for TLS 1.0
13754 but also for SSL 3.0 (as required by the specification).
13756 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13758 * Change X509_certificate_type() to mark the key as exported/exportable
13759 when it's 512 *bits* long, not 512 bytes.
13763 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13765 * Countermeasure against the Klima-Pokorny-Rosa extension of
13766 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13767 a protocol version number mismatch like a decryption error
13768 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13772 * Turn on RSA blinding by default in the default implementation
13773 to avoid a timing attack. Applications that don't want it can call
13774 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13775 They would be ill-advised to do so in most cases.
13777 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13779 * Change RSA blinding code so that it works when the PRNG is not
13780 seeded (in this case, the secret RSA exponent is abused as
13781 an unpredictable seed -- if it is not unpredictable, there
13782 is no point in blinding anyway). Make RSA blinding thread-safe
13783 by remembering the creator's thread ID in rsa->blinding and
13784 having all other threads use local one-time blinding factors
13785 (this requires more computation than sharing rsa->blinding, but
13786 avoids excessive locking; and if an RSA object is not shared
13787 between threads, blinding will still be very fast).
13791 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13793 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13794 via timing by performing a MAC computation even if incorrect
13795 block cipher padding has been found. This is a countermeasure
13796 against active attacks where the attacker has to distinguish
13797 between bad padding and a MAC verification error. ([CVE-2003-0078])
13799 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13800 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13801 Martin Vuagnoux (EPFL, Ilion)*
13803 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13805 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13806 memory from its contents. This is done with a counter that will
13807 place alternating values in each byte. This can be used to solve
13808 two issues: 1) the removal of calls to memset() by highly optimizing
13809 compilers, and 2) cleansing with other values than 0, since those can
13810 be read through on certain media, for example a swap space on disk.
13814 * Bugfix: client side session caching did not work with external caching,
13815 because the session->cipher setting was not restored when reloading
13816 from the external cache. This problem was masked, when
13817 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13818 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13822 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13823 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13825 *Zeev Lieber <zeev-l@yahoo.com>*
13827 * Undo an undocumented change introduced in 0.9.6e which caused
13828 repeated calls to OpenSSL_add_all_ciphers() and
13829 OpenSSL_add_all_digests() to be ignored, even after calling
13834 * Change the default configuration reader to deal with last line not
13835 being properly terminated.
13839 * Change X509_NAME_cmp() so it applies the special rules on handling
13840 DN values that are of type PrintableString, as well as RDNs of type
13841 emailAddress where the value has the type ia5String.
13843 *stefank@valicert.com via Richard Levitte*
13845 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13846 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13847 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13848 the bitwise-OR of the two for use by the majority of applications
13849 wanting this behaviour, and update the docs. The documented
13850 behaviour and actual behaviour were inconsistent and had been
13851 changing anyway, so this is more a bug-fix than a behavioural
13854 *Geoff Thorpe, diagnosed by Nadav Har'El*
13856 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13857 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13861 * Fix initialization code race conditions in
13862 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13863 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13864 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13865 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13866 ssl2_get_cipher_by_char(),
13867 ssl3_get_cipher_by_char().
13869 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13871 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13872 the cached sessions are flushed, as the remove_cb() might use ex_data
13873 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13874 (see [openssl.org #212]).
13876 *Geoff Thorpe, Lutz Jaenicke*
13878 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13879 length, instead of the encoding length to d2i_ASN1_OBJECT.
13883 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13885 * [In 0.9.6g-engine release:]
13886 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13888 *Lynn Gazis <lgazis@rainbow.com>*
13890 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13892 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13893 and get fix the header length calculation.
13894 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13895 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13897 * Use proper error handling instead of 'assertions' in buffer
13898 overflow checks added in 0.9.6e. This prevents DoS (the
13899 assertions could call abort()).
13901 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13903 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13905 * Add various sanity checks to asn1_get_length() to reject
13906 the ASN1 length bytes if they exceed sizeof(long), will appear
13907 negative or the content length exceeds the length of the
13910 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13912 * Fix cipher selection routines: ciphers without encryption had no flags
13913 for the cipher strength set and where therefore not handled correctly
13914 by the selection routines (PR #130).
13918 * Fix EVP_dsa_sha macro.
13923 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13924 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13925 that was added in OpenSSL 0.9.6d.
13927 As the countermeasure turned out to be incompatible with some
13928 broken SSL implementations, the new option is part of SSL_OP_ALL.
13929 SSL_OP_ALL is usually employed when compatibility with weird SSL
13930 implementations is desired (e.g. '-bugs' option to 's_client' and
13931 's_server'), so the new option is automatically set in many
13936 * Changes in security patch:
13938 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13939 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13940 Air Force Materiel Command, USAF, under agreement number
13943 * Add various sanity checks to asn1_get_length() to reject
13944 the ASN1 length bytes if they exceed sizeof(long), will appear
13945 negative or the content length exceeds the length of the
13946 supplied buffer. ([CVE-2002-0659])
13948 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13950 * Assertions for various potential buffer overflows, not known to
13951 happen in practice.
13953 *Ben Laurie (CHATS)*
13955 * Various temporary buffers to hold ASCII versions of integers were
13956 too small for 64 bit platforms. ([CVE-2002-0655])
13957 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13959 * Remote buffer overflow in SSL3 protocol - an attacker could
13960 supply an oversized session ID to a client. ([CVE-2002-0656])
13962 *Ben Laurie (CHATS)*
13964 * Remote buffer overflow in SSL2 protocol - an attacker could
13965 supply an oversized client master key. ([CVE-2002-0656])
13967 *Ben Laurie (CHATS)*
13969 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13971 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13972 encoded as NULL) with id-dsa-with-sha1.
13974 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13976 * Check various `X509_...()` return values in `apps/req.c`.
13978 *Nils Larsch <nla@trustcenter.de>*
13980 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13981 an end-of-file condition would erroneously be flagged, when the CRLF
13982 was just at the end of a processed block. The bug was discovered when
13983 processing data through a buffering memory BIO handing the data to a
13984 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13985 <ptsekov@syntrex.com> and Nedelcho Stanev.
13989 * Implement a countermeasure against a vulnerability recently found
13990 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13991 before application data chunks to avoid the use of known IVs
13992 with data potentially chosen by the attacker.
13996 * Fix length checks in ssl3_get_client_hello().
14000 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14001 to prevent ssl3_read_internal() from incorrectly assuming that
14002 ssl3_read_bytes() found application data while handshake
14003 processing was enabled when in fact s->s3->in_read_app_data was
14004 merely automatically cleared during the initial handshake.
14006 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14008 * Fix object definitions for Private and Enterprise: they were not
14009 recognized in their shortname (=lowercase) representation. Extend
14010 obj_dat.pl to issue an error when using undefined keywords instead
14011 of silently ignoring the problem (Svenning Sorensen
14012 <sss@sss.dnsalias.net>).
14016 * Fix DH_generate_parameters() so that it works for 'non-standard'
14017 generators, i.e. generators other than 2 and 5. (Previously, the
14018 code did not properly initialise the 'add' and 'rem' values to
14019 BN_generate_prime().)
14021 In the new general case, we do not insist that 'generator' is
14022 actually a primitive root: This requirement is rather pointless;
14023 a generator of the order-q subgroup is just as good, if not
14028 * Map new X509 verification errors to alerts. Discovered and submitted by
14029 Tom Wu <tom@arcot.com>.
14033 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14034 returning non-zero before the data has been completely received
14035 when using non-blocking I/O.
14037 *Bodo Moeller; problem pointed out by John Hughes*
14039 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14041 *Ben Laurie, Lutz Jaenicke*
14043 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14044 Yoram Zahavi <YoramZ@gilian.com>).
14048 * Add information about CygWin 1.3 and on, and preserve proper
14049 configuration for the versions before that.
14051 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14053 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14054 check whether we deal with a copy of a session and do not delete from
14055 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14056 <izhar@checkpoint.com>.
14060 * Do not store session data into the internal session cache, if it
14061 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14062 flag is set). Proposed by Aslam <aslam@funk.com>.
14066 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14071 * [In 0.9.6d-engine release:]
14072 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14074 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14076 * Add the configuration target linux-s390x.
14078 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14080 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14081 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14082 variable as an indication that a ClientHello message has been
14083 received. As the flag value will be lost between multiple
14084 invocations of ssl3_accept when using non-blocking I/O, the
14085 function may not be aware that a handshake has actually taken
14086 place, thus preventing a new session from being added to the
14089 To avoid this problem, we now set s->new_session to 2 instead of
14090 using a local variable.
14092 *Lutz Jaenicke, Bodo Moeller*
14094 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14095 if the SSL_R_LENGTH_MISMATCH error is detected.
14097 *Geoff Thorpe, Bodo Moeller*
14099 * New 'shared_ldflag' column in Configure platform table.
14103 * Fix EVP_CIPHER_mode macro.
14105 *"Dan S. Camper" <dan@bti.net>*
14107 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14108 type, we must throw them away by setting rr->length to 0.
14110 *D P Chang <dpc@qualys.com>*
14112 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14114 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14115 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14116 worked incorrectly for those cases where range = `10..._2` and
14117 `3*range` is two bits longer than range.)
14121 * Only add signing time to PKCS7 structures if it is not already
14126 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14127 OBJ_ld_ce should be OBJ_id_ce.
14128 Also some ip-pda OIDs in crypto/objects/objects.txt were
14129 incorrect (cf. RFC 3039).
14131 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14133 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14134 returns early because it has nothing to do.
14136 *Andy Schneider <andy.schneider@bjss.co.uk>*
14138 * [In 0.9.6c-engine release:]
14139 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14141 *Andy Schneider <andy.schneider@bjss.co.uk>*
14143 * [In 0.9.6c-engine release:]
14144 Add support for Cryptographic Appliance's keyserver technology.
14145 (Use engine 'keyclient')
14147 *Cryptographic Appliances and Geoff Thorpe*
14149 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14150 is called via tools/c89.sh because arguments have to be
14151 rearranged (all '-L' options must appear before the first object
14154 *Richard Shapiro <rshapiro@abinitio.com>*
14156 * [In 0.9.6c-engine release:]
14157 Add support for Broadcom crypto accelerator cards, backported
14160 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14162 * [In 0.9.6c-engine release:]
14163 Add support for SureWare crypto accelerator cards from
14164 Baltimore Technologies. (Use engine 'sureware')
14166 *Baltimore Technologies and Mark Cox*
14168 * [In 0.9.6c-engine release:]
14169 Add support for crypto accelerator cards from Accelerated
14170 Encryption Processing, www.aep.ie. (Use engine 'aep')
14172 *AEP Inc. and Mark Cox*
14174 * Add a configuration entry for gcc on UnixWare.
14176 *Gary Benson <gbenson@redhat.com>*
14178 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14179 messages are stored in a single piece (fixed-length part and
14180 variable-length part combined) and fix various bugs found on the way.
14184 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14185 instead. BIO_gethostbyname() does not know what timeouts are
14186 appropriate, so entries would stay in cache even when they have
14188 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14190 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14191 faced with a pathologically small ClientHello fragment that does
14192 not contain client_version: Instead of aborting with an error,
14193 simply choose the highest available protocol version (i.e.,
14194 TLS 1.0 unless it is disabled). In practice, ClientHello
14195 messages are never sent like this, but this change gives us
14196 strictly correct behaviour at least for TLS.
14200 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14201 never resets s->method to s->ctx->method when called from within
14202 one of the SSL handshake functions.
14204 *Bodo Moeller; problem pointed out by Niko Baric*
14206 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14207 (sent using the client's version number) if client_version is
14208 smaller than the protocol version in use. Also change
14209 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14210 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14211 the client will at least see that alert.
14215 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14220 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14221 client receives HelloRequest while in a handshake.
14223 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14225 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14226 should end in 'break', not 'goto end' which circumvents various
14227 cleanups done in state SSL_ST_OK. But session related stuff
14228 must be disabled for SSL_ST_OK in the case that we just sent a
14231 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14232 before just sending a HelloRequest.
14234 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14236 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14237 reveal whether illegal block cipher padding was found or a MAC
14238 verification error occurred. (Neither SSLerr() codes nor alerts
14239 are directly visible to potential attackers, but the information
14240 may leak via logfiles.)
14242 Similar changes are not required for the SSL 2.0 implementation
14243 because the number of padding bytes is sent in clear for SSL 2.0,
14244 and the extra bytes are just ignored. However ssl/s2_pkt.c
14245 failed to verify that the purported number of padding bytes is in
14250 * Add OpenUNIX-8 support including shared libraries
14251 (Boyd Lynn Gerber <gerberb@zenez.com>).
14255 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14256 'wristwatch attack' using huge encoding parameters (cf.
14257 James H. Manger's CRYPTO 2001 paper). Note that the
14258 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14259 encoding parameters and hence was not vulnerable.
14263 * BN_sqr() bug fix.
14265 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14267 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14268 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14269 followed by modular reduction.
14271 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14273 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14274 equivalent based on BN_pseudo_rand() instead of BN_rand().
14278 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14279 This function was broken, as the check for a new client hello message
14280 to handle SGC did not allow these large messages.
14281 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14285 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14289 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14290 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14294 * Rework the configuration and shared library support for Tru64 Unix.
14295 The configuration part makes use of modern compiler features and
14296 still retains old compiler behavior for those that run older versions
14297 of the OS. The shared library support part includes a variant that
14298 uses the RPATH feature, and is available through the special
14299 configuration target "alpha-cc-rpath", which will never be selected
14302 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14304 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14305 with the same message size as in ssl3_get_certificate_request().
14306 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14307 messages might inadvertently be reject as too long.
14309 *Petr Lampa <lampa@fee.vutbr.cz>*
14311 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14315 * Modified SSL library such that the verify_callback that has been set
14316 specifically for an SSL object with SSL_set_verify() is actually being
14317 used. Before the change, a verify_callback set with this function was
14318 ignored and the verify_callback() set in the SSL_CTX at the time of
14319 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14320 to allow the necessary settings.
14324 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14325 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14326 done automatically (in contradiction to the requirements of the C
14327 standard). This made problems when used from OpenSSH.
14331 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14332 dh->length and always used
14334 BN_rand_range(priv_key, dh->p).
14336 BN_rand_range() is not necessary for Diffie-Hellman, and this
14337 specific range makes Diffie-Hellman unnecessarily inefficient if
14338 dh->length (recommended exponent length) is much smaller than the
14339 length of dh->p. We could use BN_rand_range() if the order of
14340 the subgroup was stored in the DH structure, but we only have
14345 BN_rand(priv_key, l, ...)
14347 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14354 RSA_eay_public_encrypt
14355 RSA_eay_private_decrypt
14356 RSA_eay_private_encrypt (signing)
14357 RSA_eay_public_decrypt (signature verification)
14359 (default implementations for RSA_public_encrypt,
14360 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14361 always reject numbers >= n.
14365 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14366 to synchronize access to 'locking_thread'. This is necessary on
14367 systems where access to 'locking_thread' (an 'unsigned long'
14368 variable) is not atomic.
14372 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14373 *before* setting the 'crypto_lock_rand' flag. The previous code had
14374 a race condition if 0 is a valid thread ID.
14376 *Travis Vitek <vitek@roguewave.com>*
14378 * Add support for shared libraries under Irix.
14380 *Albert Chin-A-Young <china@thewrittenword.com>*
14382 * Add configuration option to build on Linux on both big-endian and
14383 little-endian MIPS.
14385 *Ralf Baechle <ralf@uni-koblenz.de>*
14387 * Add the possibility to create shared libraries on HP-UX.
14391 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14393 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14394 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14395 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14396 PRNG state recovery was possible based on the output of
14397 one PRNG request appropriately sized to gain knowledge on
14398 'md' followed by enough consecutive 1-byte PRNG requests
14399 to traverse all of 'state'.
14401 1. When updating 'md_local' (the current thread's copy of 'md')
14402 during PRNG output generation, hash all of the previous
14403 'md_local' value, not just the half used for PRNG output.
14405 2. Make the number of bytes from 'state' included into the hash
14406 independent from the number of PRNG bytes requested.
14408 The first measure alone would be sufficient to avoid
14409 Markku-Juhani's attack. (Actually it had never occurred
14410 to me that the half of 'md_local' used for chaining was the
14411 half from which PRNG output bytes were taken -- I had always
14412 assumed that the secret half would be used.) The second
14413 measure makes sure that additional data from 'state' is never
14414 mixed into 'md_local' in small portions; this heuristically
14415 further strengthens the PRNG.
14419 * Fix crypto/bn/asm/mips3.s.
14423 * When only the key is given to "enc", the IV is undefined. Print out
14424 an error message in this case.
14428 * Handle special case when X509_NAME is empty in X509 printing routines.
14432 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14433 positive and less than q.
14437 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14438 used: it isn't thread safe and the add_lock_callback should handle
14441 *Paul Rose <Paul.Rose@bridge.com>*
14443 * Verify that incoming data obeys the block size in
14444 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14450 *Ulf Möller, Bodo Möller*
14452 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14453 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14454 when fixing the server behaviour for backwards-compatible 'client
14455 hello' messages. (Note that the attack is impractical against
14456 SSL 3.0 and TLS 1.0 anyway because length and version checking
14457 means that the probability of guessing a valid ciphertext is
14458 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14461 Before 0.9.5, the countermeasure (hide the error by generating a
14462 random 'decryption result') did not work properly because
14463 ERR_clear_error() was missing, meaning that SSL_get_error() would
14464 detect the supposedly ignored error.
14466 Both problems are now fixed.
14470 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14471 (previously it was 1024).
14475 * Fix for compatibility mode trust settings: ignore trust settings
14476 unless some valid trust or reject settings are present.
14480 * Fix for blowfish EVP: its a variable length cipher.
14484 * Fix various bugs related to DSA S/MIME verification. Handle missing
14485 parameters in DSA public key structures and return an error in the
14486 DSA routines if parameters are absent.
14490 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14491 in the current directory if neither $RANDFILE nor $HOME was set.
14492 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14493 caused some confusion to Windows users who haven't defined $HOME.
14494 Thus RAND_file_name() is changed again: e_os.h can define a
14495 DEFAULT_HOME, which will be used if $HOME is not set.
14496 For Windows, we use "C:"; on other platforms, we still require
14497 environment variables.
14499 * Move 'if (!initialized) RAND_poll()' into regions protected by
14500 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14501 having multiple threads call RAND_poll() concurrently.
14505 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14506 combination of a flag and a thread ID variable.
14507 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14508 flag), *other* threads can enter ssleay_add_bytes without obeying
14509 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14510 that they do not hold after the first thread unsets add_do_not_lock).
14514 * Change bctest again: '-x' expressions are not available in all
14515 versions of 'test'.
14519 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14521 * Fix a couple of memory leaks in PKCS7_dataDecode()
14523 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14525 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14526 the default extension for executables, if any. Also, make the perl
14527 scripts that use symlink() to test if it really exists and use "cp"
14528 if it doesn't. All this made OpenSSL compilable and installable in
14533 * Fix for asn1_GetSequence() for indefinite length constructed data.
14534 If SEQUENCE is length is indefinite just set c->slen to the total
14535 amount of data available.
14537 *Steve Henson, reported by shige@FreeBSD.org*
14539 *This change does not apply to 0.9.7.*
14541 * Change bctest to avoid here-documents inside command substitution
14542 (workaround for FreeBSD /bin/sh bug).
14543 For compatibility with Ultrix, avoid shell functions (introduced
14544 in the bctest version that searches along $PATH).
14548 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14549 with des_encrypt() defined on some operating systems, like Solaris
14554 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14555 On the Importance of Eliminating Errors in Cryptographic
14556 Computations, J. Cryptology 14 (2001) 2, 101-119,
14557 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14561 * MIPS assembler BIGNUM division bug fix.
14565 * Disabled incorrect Alpha assembler code.
14569 * Fix PKCS#7 decode routines so they correctly update the length
14570 after reading an EOC for the EXPLICIT tag.
14574 *This change does not apply to 0.9.7.*
14576 * Fix bug in PKCS#12 key generation routines. This was triggered
14577 if a 3DES key was generated with a 0 initial byte. Include
14578 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14579 (but broken) behaviour.
14583 * Enhance bctest to search for a working bc along $PATH and print
14586 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14588 * Fix memory leaks in err.c: free err_data string if necessary;
14589 don't write to the wrong index in ERR_set_error_data.
14593 * Implement ssl23_peek (analogous to ssl23_read), which previously
14598 * Replace rdtsc with `_emit` statements for VC++ version 5.
14600 *Jeremy Cooper <jeremy@baymoo.org>*
14602 * Make it possible to reuse SSLv2 sessions.
14606 * In copy_email() check for >= 0 as a return value for
14607 X509_NAME_get_index_by_NID() since 0 is a valid index.
14609 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14611 * Avoid coredump with unsupported or invalid public keys by checking if
14612 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14613 PKCS7_verify() fails with non detached data.
14617 * Don't use getenv in library functions when run as setuid/setgid.
14618 New function OPENSSL_issetugid().
14622 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14623 due to incorrect handling of multi-threading:
14625 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14627 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14629 3. Count how many times MemCheck_off() has been called so that
14630 nested use can be treated correctly. This also avoids
14631 inband-signalling in the previous code (which relied on the
14632 assumption that thread ID 0 is impossible).
14636 * Add "-rand" option also to s_client and s_server.
14640 * Fix CPU detection on Irix 6.x.
14641 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14642 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14644 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14649 *This change does not apply to 0.9.7.*
14651 * Use the cached encoding of an X509_NAME structure rather than
14652 copying it. This is apparently the reason for the libsafe "errors"
14653 but the code is actually correct.
14657 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14658 Bleichenbacher's DSA attack.
14659 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14660 to be set and top=0 forces the highest bit to be set; top=-1 is new
14661 and leaves the highest bit random.
14663 *Ulf Moeller, Bodo Moeller*
14665 * In the `NCONF_...`-based implementations for `CONF_...` queries
14666 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14667 a temporary CONF structure with the data component set to NULL
14668 (which gives segmentation faults in lh_retrieve).
14669 Instead, use NULL for the CONF pointer in CONF_get_string and
14670 CONF_get_number (which may use environment variables) and directly
14671 return NULL from CONF_get_section.
14675 * Fix potential buffer overrun for EBCDIC.
14679 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14680 keyUsage if basicConstraints absent for a CA.
14684 * Make SMIME_write_PKCS7() write mail header values with a format that
14685 is more generally accepted (no spaces before the semicolon), since
14686 some programs can't parse those values properly otherwise. Also make
14687 sure BIO's that break lines after each write do not create invalid
14692 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14693 macros previously used would not encode an empty SEQUENCE OF
14694 and break the signature.
14698 *This change does not apply to 0.9.7.*
14700 * Zero the premaster secret after deriving the master secret in
14705 * Add some EVP_add_digest_alias registrations (as found in
14706 OpenSSL_add_all_digests()) to SSL_library_init()
14707 aka OpenSSL_add_ssl_algorithms(). This provides improved
14708 compatibility with peers using X.509 certificates
14709 with unconventional AlgorithmIdentifier OIDs.
14713 * Fix for Irix with NO_ASM.
14715 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14717 * ./config script fixes.
14719 *Ulf Moeller, Richard Levitte*
14721 * Fix 'openssl passwd -1'.
14725 * Change PKCS12_key_gen_asc() so it can cope with non null
14726 terminated strings whose length is passed in the passlen
14727 parameter, for example from PEM callbacks. This was done
14728 by adding an extra length parameter to asc2uni().
14730 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14732 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14733 call failed, free the DSA structure.
14737 * Fix to uni2asc() to cope with zero length Unicode strings.
14738 These are present in some PKCS#12 files.
14742 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14743 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14744 when writing a 32767 byte record.
14746 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14748 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14749 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14751 (RSA objects have a reference count access to which is protected
14752 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14753 so they are meant to be shared between threads.)
14754 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14755 "Reddie, Steven" <Steven.Reddie@ca.com>*
14757 * Fix a deadlock in CRYPTO_mem_leaks().
14761 * Use better test patterns in bntest.
14765 * rand_win.c fix for Borland C.
14769 * BN_rshift bugfix for n == 0.
14773 * Add a 'bctest' script that checks for some known 'bc' bugs
14774 so that 'make test' does not abort just because 'bc' is broken.
14778 * Store verify_result within SSL_SESSION also for client side to
14779 avoid potential security hole. (Re-used sessions on the client side
14780 always resulted in verify_result==X509_V_OK, not using the original
14781 result of the server certificate verification.)
14785 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14786 SSL3_RT_APPLICATION_DATA, return 0.
14787 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14792 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14793 releases, have been re-implemented by renaming the previous
14794 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14795 and ssl3_read_internal, respectively, and adding 'peek' parameters
14796 to them. The new ssl[23]_{read,peek} functions are calls to
14797 ssl[23]_read_internal with the 'peek' flag set appropriately.
14798 A 'peek' parameter has also been added to ssl3_read_bytes, which
14799 does the actual work for ssl3_read_internal.
14803 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14804 the method-specific "init()" handler. Also clean up ex_data after
14805 calling the method-specific "finish()" handler. Previously, this was
14806 happening the other way round.
14810 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14811 The previous value, 12, was not always sufficient for BN_mod_exp().
14815 * Make sure that shared libraries get the internal name engine with
14816 the full version number and not just 0. This should mark the
14817 shared libraries as not backward compatible. Of course, this should
14818 be changed again when we can guarantee backward binary compatibility.
14822 * Fix typo in get_cert_by_subject() in by_dir.c
14824 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14826 * Rework the system to generate shared libraries:
14828 - Make note of the expected extension for the shared libraries and
14829 if there is a need for symbolic links from for example libcrypto.so.0
14830 to libcrypto.so.0.9.7. There is extended info in Configure for
14833 - Make as few rebuilds of the shared libraries as possible.
14835 - Still avoid linking the OpenSSL programs with the shared libraries.
14837 - When installing, install the shared libraries separately from the
14842 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14844 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14845 and not in SSL_clear because the latter is also used by the
14846 accept/connect functions; previously, the settings made by
14847 SSL_set_read_ahead would be lost during the handshake.
14849 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14851 * Correct util/mkdef.pl to be selective about disabled algorithms.
14852 Previously, it would create entries for disabled algorithms no
14857 * Added several new manual pages for SSL_* function.
14861 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14863 * In ssl23_get_client_hello, generate an error message when faced
14864 with an initial SSL 3.0/TLS record that is too small to contain the
14865 first two bytes of the ClientHello message, i.e. client_version.
14866 (Note that this is a pathologic case that probably has never happened
14867 in real life.) The previous approach was to use the version number
14868 from the record header as a substitute; but our protocol choice
14869 should not depend on that one because it is not authenticated
14870 by the Finished messages.
14874 * More robust randomness gathering functions for Windows.
14876 *Jeffrey Altman <jaltman@columbia.edu>*
14878 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14879 not set then we don't setup the error code for issuer check errors
14880 to avoid possibly overwriting other errors which the callback does
14881 handle. If an application does set the flag then we assume it knows
14882 what it is doing and can handle the new informational codes
14887 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14888 a general "ANY" type, as such it should be able to decode anything
14889 including tagged types. However it didn't check the class so it would
14890 wrongly interpret tagged types in the same way as their universal
14891 counterpart and unknown types were just rejected. Changed so that the
14892 tagged and unknown types are handled in the same way as a SEQUENCE:
14893 that is the encoding is stored intact. There is also a new type
14894 "V_ASN1_OTHER" which is used when the class is not universal, in this
14895 case we have no idea what the actual type is so we just lump them all
14900 * On VMS, stdout may very well lead to a file that is written to
14901 in a record-oriented fashion. That means that every write() will
14902 write a separate record, which will be read separately by the
14903 programs trying to read from it. This can be very confusing.
14905 The solution is to put a BIO filter in the way that will buffer
14906 text until a linefeed is reached, and then write everything a
14907 line at a time, so every record written will be an actual line,
14908 not chunks of lines and not (usually doesn't happen, but I've
14909 seen it once) several lines in one record. BIO_f_linebuffer() is
14912 Currently, it's a VMS-only method, because that's where it has
14913 been tested well enough.
14917 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14918 it can return incorrect results.
14919 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14920 but it was in 0.9.6-beta[12].)
14924 * Disable the check for content being present when verifying detached
14925 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14926 include zero length content when signing messages.
14930 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14931 BIO_ctrl (for BIO pairs).
14935 * Add DSO method for VMS.
14939 * Bug fix: Montgomery multiplication could produce results with the
14944 * Add RPM specification openssl.spec and modify it to build three
14945 packages. The default package contains applications, application
14946 documentation and run-time libraries. The devel package contains
14947 include files, static libraries and function documentation. The
14948 doc package contains the contents of the doc directory. The original
14949 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14953 * Add a large number of documentation files for many SSL routines.
14955 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14957 * Add a configuration entry for Sony News 4.
14959 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14961 * Don't set the two most significant bits to one when generating a
14962 random number < q in the DSA library.
14966 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14967 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14968 the underlying transport is blocking) if a handshake took place.
14969 (The default behaviour is needed by applications such as s_client
14970 and s_server that use select() to determine when to use SSL_read;
14971 but for applications that know in advance when to expect data, it
14972 just makes things more complicated.)
14976 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14981 * Add a few more EBCDIC conditionals that make `req` and `x509`
14982 work better on such systems.
14984 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14986 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14987 Update PKCS12_parse() so it copies the friendlyName and the
14988 keyid to the certificates aux info.
14992 * Fix bug in PKCS7_verify() which caused an infinite loop
14993 if there was more than one signature.
14995 *Sven Uszpelkat <su@celocom.de>*
14997 * Major change in util/mkdef.pl to include extra information
14998 about each symbol, as well as presenting variables as well
14999 as functions. This change means that there's n more need
15000 to rebuild the .num files when some algorithms are excluded.
15004 * Allow the verify time to be set by an application,
15005 rather than always using the current time.
15009 * Phase 2 verify code reorganisation. The certificate
15010 verify code now looks up an issuer certificate by a
15011 number of criteria: subject name, authority key id
15012 and key usage. It also verifies self signed certificates
15013 by the same criteria. The main comparison function is
15014 X509_check_issued() which performs these checks.
15016 Lot of changes were necessary in order to support this
15017 without completely rewriting the lookup code.
15019 Authority and subject key identifier are now cached.
15021 The LHASH 'certs' is X509_STORE has now been replaced
15022 by a STACK_OF(X509_OBJECT). This is mainly because an
15023 LHASH can't store or retrieve multiple objects with
15024 the same hash value.
15026 As a result various functions (which were all internal
15027 use only) have changed to handle the new X509_STORE
15028 structure. This will break anything that messed round
15029 with X509_STORE internally.
15031 The functions X509_STORE_add_cert() now checks for an
15032 exact match, rather than just subject name.
15034 The X509_STORE API doesn't directly support the retrieval
15035 of multiple certificates matching a given criteria, however
15036 this can be worked round by performing a lookup first
15037 (which will fill the cache with candidate certificates)
15038 and then examining the cache for matches. This is probably
15039 the best we can do without throwing out X509_LOOKUP
15040 entirely (maybe later...).
15042 The X509_VERIFY_CTX structure has been enhanced considerably.
15044 All certificate lookup operations now go via a get_issuer()
15045 callback. Although this currently uses an X509_STORE it
15046 can be replaced by custom lookups. This is a simple way
15047 to bypass the X509_STORE hackery necessary to make this
15048 work and makes it possible to use more efficient techniques
15049 in future. A very simple version which uses a simple
15050 STACK for its trusted certificate store is also provided
15051 using X509_STORE_CTX_trusted_stack().
15053 The verify_cb() and verify() callbacks now have equivalents
15054 in the X509_STORE_CTX structure.
15056 X509_STORE_CTX also has a 'flags' field which can be used
15057 to customise the verify behaviour.
15061 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15062 excludes S/MIME capabilities.
15066 * When a certificate request is read in keep a copy of the
15067 original encoding of the signed data and use it when outputting
15068 again. Signatures then use the original encoding rather than
15069 a decoded, encoded version which may cause problems if the
15070 request is improperly encoded.
15074 * For consistency with other BIO_puts implementations, call
15075 buffer_write(b, ...) directly in buffer_puts instead of calling
15078 In BIO_puts, increment b->num_write as in BIO_write.
15080 *Peter.Sylvester@EdelWeb.fr*
15082 * Fix BN_mul_word for the case where the word is 0. (We have to use
15083 BN_zero, we may not return a BIGNUM with an array consisting of
15084 words set to zero.)
15088 * Avoid calling abort() from within the library when problems are
15089 detected, except if preprocessor symbols have been defined
15090 (such as REF_CHECK, BN_DEBUG etc.).
15094 * New openssl application 'rsautl'. This utility can be
15095 used for low-level RSA operations. DER public key
15096 BIO/fp routines also added.
15100 * New Configure entry and patches for compiling on QNX 4.
15102 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15104 * A demo state-machine implementation was sponsored by
15105 Nuron (<http://www.nuron.com/>) and is now available in
15106 demos/state_machine.
15110 * New options added to the 'dgst' utility for signature
15111 generation and verification.
15115 * Unrecognized PKCS#7 content types are now handled via a
15116 catch all ASN1_TYPE structure. This allows unsupported
15117 types to be stored as a "blob" and an application can
15118 encode and decode it manually.
15122 * Fix various signed/unsigned issues to make a_strex.c
15123 compile under VC++.
15125 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15127 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15128 length if passed a buffer. ASN1_INTEGER_to_BN failed
15129 if passed a NULL BN and its argument was negative.
15131 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15133 * Modification to PKCS#7 encoding routines to output definite
15134 length encoding. Since currently the whole structures are in
15135 memory there's not real point in using indefinite length
15136 constructed encoding. However if OpenSSL is compiled with
15137 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15141 * Added BIO_vprintf() and BIO_vsnprintf().
15145 * Added more prefixes to parse for in the strings written
15146 through a logging bio, to cover all the levels that are available
15147 through syslog. The prefixes are now:
15149 PANIC, EMERG, EMR => LOG_EMERG
15150 ALERT, ALR => LOG_ALERT
15151 CRIT, CRI => LOG_CRIT
15152 ERROR, ERR => LOG_ERR
15153 WARNING, WARN, WAR => LOG_WARNING
15154 NOTICE, NOTE, NOT => LOG_NOTICE
15155 INFO, INF => LOG_INFO
15156 DEBUG, DBG => LOG_DEBUG
15158 and as before, if none of those prefixes are present at the
15159 beginning of the string, LOG_ERR is chosen.
15161 On Win32, the `LOG_*` levels are mapped according to this:
15163 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15164 LOG_WARNING => EVENTLOG_WARNING_TYPE
15165 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15169 * Made it possible to reconfigure with just the configuration
15170 argument "reconf" or "reconfigure". The command line arguments
15171 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15172 and are retrieved from there when reconfiguring.
15178 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15180 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15184 * The obj_dat.pl script was messing up the sorting of object
15185 names. The reason was that it compared the quoted version
15186 of strings as a result "OCSP" > "OCSP Signing" because
15187 " > SPACE. Changed script to store unquoted versions of
15188 names and add quotes on output. It was also omitting some
15189 names from the lookup table if they were given a default
15190 value (that is if SN is missing it is given the same
15191 value as LN and vice versa), these are now added on the
15192 grounds that if an object has a name we should be able to
15193 look it up. Finally added warning output when duplicate
15194 short or long names are found.
15198 * Changes needed for Tandem NSK.
15200 *Scott Uroff <scott@xypro.com>*
15202 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15203 RSA_padding_check_SSLv23(), special padding was never detected
15204 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15205 version rollback attacks was not effective.
15207 In s23_clnt.c, don't use special rollback-attack detection padding
15208 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15209 client; similarly, in s23_srvr.c, don't do the rollback check if
15210 SSL 2.0 is the only protocol enabled in the server.
15214 * Make it possible to get hexdumps of unprintable data with 'openssl
15215 asn1parse'. By implication, the functions ASN1_parse_dump() and
15216 BIO_dump_indent() are added.
15220 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15221 these print out strings and name structures based on various
15222 flags including RFC2253 support and proper handling of
15223 multibyte characters. Added options to the 'x509' utility
15224 to allow the various flags to be set.
15228 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15229 Also change the functions X509_cmp_current_time() and
15230 X509_gmtime_adj() work with an ASN1_TIME structure,
15231 this will enable certificates using GeneralizedTime in validity
15232 dates to be checked.
15236 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15237 negative public key encodings) on by default,
15238 NO_NEG_PUBKEY_BUG can be set to disable it.
15242 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15243 content octets. An i2c_ASN1_OBJECT is unnecessary because
15244 the encoding can be trivially obtained from the structure.
15248 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15249 not read locks (`CRYPTO_r_[un]lock`).
15253 * A first attempt at creating official support for shared
15254 libraries through configuration. I've kept it so the
15255 default is static libraries only, and the OpenSSL programs
15256 are always statically linked for now, but there are
15257 preparations for dynamic linking in place.
15258 This has been tested on Linux and Tru64.
15262 * Randomness polling function for Win9x, as described in:
15263 Peter Gutmann, Software Generation of Practically Strong
15268 * Fix so PRNG is seeded in req if using an already existing
15273 * New options to smime application. -inform and -outform
15274 allow alternative formats for the S/MIME message including
15275 PEM and DER. The -content option allows the content to be
15276 specified separately. This should allow things like Netscape
15277 form signing output easier to verify.
15281 * Fix the ASN1 encoding of tags using the 'long form'.
15285 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15286 STRING types. These convert content octets to and from the
15287 underlying type. The actual tag and length octets are
15288 already assumed to have been read in and checked. These
15289 are needed because all other string types have virtually
15290 identical handling apart from the tag. By having versions
15291 of the ASN1 functions that just operate on content octets
15292 IMPLICIT tagging can be handled properly. It also allows
15293 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15294 and ASN1_INTEGER are identical apart from the tag.
15298 * Change the handling of OID objects as follows:
15300 - New object identifiers are inserted in objects.txt, following
15301 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15302 - objects.pl is used to process obj_mac.num and create a new
15304 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15307 This is currently kind of a hack, and the perl code in objects.pl
15308 isn't very elegant, but it works as I intended. The simplest way
15309 to check that it worked correctly is to look in obj_dat.h and
15310 check the array nid_objs and make sure the objects haven't moved
15311 around (this is important!). Additions are OK, as well as
15312 consistent name changes.
15316 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15320 * Addition of the command line parameter '-rand file' to 'openssl req'.
15321 The given file adds to whatever has already been seeded into the
15322 random pool through the RANDFILE configuration file option or
15323 environment variable, or the default random state file.
15327 * mkstack.pl now sorts each macro group into lexical order.
15328 Previously the output order depended on the order the files
15329 appeared in the directory, resulting in needless rewriting
15334 * Patches to make OpenSSL compile under Win32 again. Mostly
15335 work arounds for the VC++ problem that it treats func() as
15336 func(void). Also stripped out the parts of mkdef.pl that
15337 added extra typesafe functions: these no longer exist.
15341 * Reorganisation of the stack code. The macros are now all
15342 collected in safestack.h . Each macro is defined in terms of
15343 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15344 DEBUG_SAFESTACK is now handled in terms of function casts,
15345 this has the advantage of retaining type safety without the
15346 use of additional functions. If DEBUG_SAFESTACK is not defined
15347 then the non typesafe macros are used instead. Also modified the
15348 mkstack.pl script to handle the new form. Needs testing to see
15349 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15350 the default if no major problems. Similar behaviour for ASN1_SET_OF
15351 and PKCS12_STACK_OF.
15355 * When some versions of IIS use the 'NET' form of private key the
15356 key derivation algorithm is different. Normally MD5(password) is
15357 used as a 128 bit RC4 key. In the modified case
15358 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15359 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15360 as the old Netscape_RSA functions except they have an additional
15361 'sgckey' parameter which uses the modified algorithm. Also added
15362 an -sgckey command line option to the rsa utility. Thanks to
15363 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15364 algorithm to openssl-dev.
15368 * The evp_local.h macros were using 'c.##kname' which resulted in
15369 invalid expansion on some systems (SCO 5.0.5 for example).
15370 Corrected to 'c.kname'.
15372 *Phillip Porch <root@theporch.com>*
15374 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15375 a STACK of email addresses from a certificate or request, these look
15376 in the subject name and the subject alternative name extensions and
15377 omit any duplicate addresses.
15381 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15382 This makes DSA verification about 2 % faster.
15386 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15387 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15388 plus overhead for 1024 bit moduli).
15389 This makes exponentiations about 0.5 % faster for 1024 bit
15390 exponents (as measured by "openssl speed rsa2048").
15394 * Rename memory handling macros to avoid conflicts with other
15396 Malloc => OPENSSL_malloc
15397 Malloc_locked => OPENSSL_malloc_locked
15398 Realloc => OPENSSL_realloc
15399 Free => OPENSSL_free
15403 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15404 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15408 * CygWin32 support.
15410 *John Jarvie <jjarvie@newsguy.com>*
15412 * The type-safe stack code has been rejigged. It is now only compiled
15413 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15414 by default all type-specific stack functions are "#define"d back to
15415 standard stack functions. This results in more streamlined output
15416 but retains the type-safety checking possibilities of the original
15421 * The STACK code has been cleaned up, and certain type declarations
15422 that didn't make a lot of sense have been brought in line. This has
15423 also involved a cleanup of sorts in safestack.h to more correctly
15424 map type-safe stack functions onto their plain stack counterparts.
15425 This work has also resulted in a variety of "const"ifications of
15426 lots of the code, especially `_cmp` operations which should normally
15427 be prototyped with "const" parameters anyway.
15431 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15432 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15433 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15434 where all of 'md' is used each time the PRNG is used, but 'state'
15435 is used only indexed by a cyclic counter. As entropy may not be
15436 well distributed from the beginning, 'md' is important as a
15437 chaining variable. However, the output function chains only half
15438 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15439 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15440 in all of 'state' being rewritten, with the new values depending
15441 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15445 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15446 the handshake is continued after ssl_verify_cert_chain();
15447 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15448 can lead to 'unexplainable' connection aborts later.
15450 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15452 * Major EVP API cipher revision.
15453 Add hooks for extra EVP features. This allows various cipher
15454 parameters to be set in the EVP interface. Support added for variable
15455 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15456 setting of RC2 and RC5 parameters.
15458 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15461 Remove lots of duplicated code from the EVP library. For example *every*
15462 cipher init() function handles the 'iv' in the same way according to the
15463 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15464 for CFB and OFB modes they zero ctx->num.
15466 New functionality allows removal of S/MIME code RC2 hack.
15468 Most of the routines have the same form and so can be declared in terms
15471 By shifting this to the top level EVP_CipherInit() it can be removed from
15472 all individual ciphers. If the cipher wants to handle IVs or keys
15473 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15476 Change lots of functions like EVP_EncryptUpdate() to now return a
15477 value: although software versions of the algorithms cannot fail
15478 any installed hardware versions can.
15482 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15483 this option is set, tolerate broken clients that send the negotiated
15484 protocol version number instead of the requested protocol version
15489 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15490 i.e. non-zero for export ciphersuites, zero otherwise.
15491 Previous versions had this flag inverted, inconsistent with
15492 rsa_tmp_cb (..._TMP_RSA_CB).
15494 *Bodo Moeller; problem reported by Amit Chopra*
15496 * Add missing DSA library text string. Work around for some IIS
15497 key files with invalid SEQUENCE encoding.
15501 * Add a document (doc/standards.txt) that list all kinds of standards
15502 and so on that are implemented in OpenSSL.
15506 * Enhance c_rehash script. Old version would mishandle certificates
15507 with the same subject name hash and wouldn't handle CRLs at all.
15508 Added -fingerprint option to crl utility, to support new c_rehash
15513 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15517 * Fix for SSL server purpose checking. Server checking was
15518 rejecting certificates which had extended key usage present
15519 but no ssl client purpose.
15521 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15523 * Make PKCS#12 code work with no password. The PKCS#12 spec
15524 is a little unclear about how a blank password is handled.
15525 Since the password in encoded as a BMPString with terminating
15526 double NULL a zero length password would end up as just the
15527 double NULL. However no password at all is different and is
15528 handled differently in the PKCS#12 key generation code. NS
15529 treats a blank password as zero length. MSIE treats it as no
15530 password on export: but it will try both on import. We now do
15531 the same: PKCS12_parse() tries zero length and no password if
15532 the password is set to "" or NULL (NULL is now a valid password:
15533 it wasn't before) as does the pkcs12 application.
15537 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15538 perror when PEM_read_bio_X509_REQ fails, the error message must
15539 be obtained from the error queue.
15543 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15544 it in ERR_remove_state if appropriate, and change ERR_get_state
15545 accordingly to avoid race conditions (this is necessary because
15546 thread_hash is no longer constant once set).
15550 * Bugfix for linux-elf makefile.one.
15554 * RSA_get_default_method() will now cause a default
15555 RSA_METHOD to be chosen if one doesn't exist already.
15556 Previously this was only set during a call to RSA_new()
15557 or RSA_new_method(NULL) meaning it was possible for
15558 RSA_get_default_method() to return NULL.
15562 * Added native name translation to the existing DSO code
15563 that will convert (if the flag to do so is set) filenames
15564 that are sufficiently small and have no path information
15565 into a canonical native form. Eg. "blah" converted to
15566 "libblah.so" or "blah.dll" etc.
15570 * New function ERR_error_string_n(e, buf, len) which is like
15571 ERR_error_string(e, buf), but writes at most 'len' bytes
15572 including the 0 terminator. For ERR_error_string_n, 'buf'
15575 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15577 * CONF library reworked to become more general. A new CONF
15578 configuration file reader "class" is implemented as well as a
15579 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15580 old `CONF_*` functions are still there, but are reimplemented to
15581 work in terms of the new functions. Also, a set of functions
15582 to handle the internal storage of the configuration data is
15583 provided to make it easier to write new configuration file
15584 reader "classes" (I can definitely see something reading a
15585 configuration file in XML format, for example), called `_CONF_*`,
15586 or "the configuration storage API"...
15588 The new configuration file reading functions are:
15590 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15591 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15593 NCONF_default, NCONF_WIN32
15595 NCONF_dump_fp, NCONF_dump_bio
15597 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15598 NCONF_new creates a new CONF object. This works in the same way
15599 as other interfaces in OpenSSL, like the BIO interface.
15600 `NCONF_dump_*` dump the internal storage of the configuration file,
15601 which is useful for debugging. All other functions take the same
15602 arguments as the old `CONF_*` functions with the exception of the
15603 first that must be a `CONF *` instead of a `LHASH *`.
15605 To make it easier to use the new classes with the old `CONF_*` functions,
15606 the function CONF_set_default_method is provided.
15610 * Add '-tls1' option to 'openssl ciphers', which was already
15611 mentioned in the documentation but had not been implemented.
15612 (This option is not yet really useful because even the additional
15613 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15617 * Initial DSO code added into libcrypto for letting OpenSSL (and
15618 OpenSSL-based applications) load shared libraries and bind to
15619 them in a portable way.
15621 *Geoff Thorpe, with contributions from Richard Levitte*
15623 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15625 * Make sure _lrotl and _lrotr are only used with MSVC.
15627 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15628 (the default implementation of RAND_status).
15630 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15631 to '-clrext' (= clear extensions), as intended and documented.
15632 *Bodo Moeller; inconsistency pointed out by Michael Attili
15633 <attili@amaxo.com>*
15635 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15636 was larger than the MD block size.
15638 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15640 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15641 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15642 using the passed key: if the passed key was a private key the result
15643 of X509_print(), for example, would be to print out all the private key
15648 * des_quad_cksum() byte order bug fix.
15649 *Ulf Möller, using the problem description in krb4-0.9.7, where
15650 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15652 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15655 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15657 * For easily testing in shell scripts whether some command
15658 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15659 returns with exit code 0 iff no command of the given name is available.
15660 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15661 the output goes to stdout and nothing is printed to stderr.
15662 Additional arguments are always ignored.
15664 Since for each cipher there is a command of the same name,
15665 the 'no-cipher' compilation switches can be tested this way.
15667 ('openssl no-XXX' is not able to detect pseudo-commands such
15668 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15672 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15676 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15677 is set; it will be thrown away anyway because each handshake creates
15679 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15680 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15681 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15682 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15686 * New s_client option -ign_eof: EOF at stdin is ignored, and
15687 'Q' and 'R' lose their special meanings (quit/renegotiate).
15688 This is part of what -quiet does; unlike -quiet, -ign_eof
15689 does not suppress any output.
15693 * Add compatibility options to the purpose and trust code. The
15694 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15695 accepts a certificate or CA, this was the previous behaviour,
15696 with all the associated security issues.
15698 X509_TRUST_COMPAT is the old trust behaviour: only and
15699 automatically trust self signed roots in certificate store. A
15700 new trust setting X509_TRUST_DEFAULT is used to specify that
15701 a purpose has no associated trust setting and it should instead
15702 use the value in the default purpose.
15706 * Fix the PKCS#8 DSA private key code so it decodes keys again
15707 and fix a memory leak.
15711 * In util/mkerr.pl (which implements 'make errors'), preserve
15712 reason strings from the previous version of the .c file, as
15713 the default to have only downcase letters (and digits) in
15714 automatically generated reasons codes is not always appropriate.
15718 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15719 using strerror. Previously, ERR_reason_error_string() returned
15720 library names as reason strings for SYSerr; but SYSerr is a special
15721 case where small numbers are errno values, not library numbers.
15725 * Add '-dsaparam' option to 'openssl dhparam' application. This
15726 converts DSA parameters into DH parameters. (When creating parameters,
15727 DSA_generate_parameters is used.)
15731 * Include 'length' (recommended exponent length) in C code generated
15732 by 'openssl dhparam -C'.
15736 * The second argument to set_label in perlasm was already being used
15737 so couldn't be used as a "file scope" flag. Moved to third argument
15742 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15743 instead of RAND_bytes for encryption IVs and salts.
15747 * Include RAND_status() into RAND_METHOD instead of implementing
15748 it only for md_rand.c Otherwise replacing the PRNG by calling
15749 RAND_set_rand_method would be impossible.
15753 * Don't let DSA_generate_key() enter an infinite loop if the random
15754 number generation fails.
15758 * New 'rand' application for creating pseudo-random output.
15762 * Added configuration support for Linux/IA64
15764 *Rolf Haberrecker <rolf@suse.de>*
15766 * Assembler module support for Mingw32.
15770 * Shared library support for HPUX (in shlib/).
15772 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15774 * Shared library support for Solaris gcc.
15776 *Lutz Behnke <behnke@trustcenter.de>*
15778 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15780 * PKCS7_encrypt() was adding text MIME headers twice because they
15781 were added manually and by SMIME_crlf_copy().
15785 * In bntest.c don't call BN_rand with zero bits argument.
15787 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15789 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15790 case was implemented. This caused BN_div_recp() to fail occasionally.
15794 * Add an optional second argument to the set_label() in the perl
15795 assembly language builder. If this argument exists and is set
15796 to 1 it signals that the assembler should use a symbol whose
15797 scope is the entire file, not just the current function. This
15798 is needed with MASM which uses the format label:: for this scope.
15800 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15802 * Change the ASN1 types so they are typedefs by default. Before
15803 almost all types were #define'd to ASN1_STRING which was causing
15804 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15809 * Change names of new functions to the new get1/get0 naming
15810 convention: After 'get1', the caller owns a reference count
15811 and has to call `..._free`; 'get0' returns a pointer to some
15812 data structure without incrementing reference counters.
15813 (Some of the existing 'get' functions increment a reference
15814 counter, some don't.)
15815 Similarly, 'set1' and 'add1' functions increase reference
15816 counters or duplicate objects.
15820 * Allow for the possibility of temp RSA key generation failure:
15821 the code used to assume it always worked and crashed on failure.
15825 * Fix potential buffer overrun problem in BIO_printf().
15826 *Ulf Möller, using public domain code by Patrick Powell; problem
15827 pointed out by David Sacerdote <das33@cornell.edu>*
15829 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15830 RAND_egd() and RAND_status(). In the command line application,
15831 the EGD socket can be specified like a seed file using RANDFILE
15836 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15837 Some CAs (e.g. Verisign) distribute certificates in this form.
15841 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15842 list to exclude them. This means that no special compilation option
15843 is needed to use anonymous DH: it just needs to be included in the
15848 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15849 EVP_MD_type. The old functionality is available in a new macro called
15850 EVP_MD_md(). Change code that uses it and update docs.
15854 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15855 where the `void *` argument is replaced by a function pointer argument.
15856 Previously `void *` was abused to point to functions, which works on
15857 many platforms, but is not correct. As these functions are usually
15858 called by macros defined in OpenSSL header files, most source code
15859 should work without changes.
15863 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15864 sections with information on -D... compiler switches used for
15865 compiling the library so that applications can see them. To enable
15866 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15867 must be defined. E.g.,
15868 #define OPENSSL_ALGORITHM_DEFINES
15869 #include <openssl/opensslconf.h>
15870 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15872 *Richard Levitte, Ulf and Bodo Möller*
15874 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15879 * Change the 'other' type in certificate aux info to a STACK_OF
15880 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15881 the required ASN1 format: arbitrary types determined by an OID.
15885 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15886 argument to 'req'. This is not because the function is newer or
15887 better than others it just uses the work 'NEW' in the certificate
15888 request header lines. Some software needs this.
15892 * Reorganise password command line arguments: now passwords can be
15893 obtained from various sources. Delete the PEM_cb function and make
15894 it the default behaviour: i.e. if the callback is NULL and the
15895 usrdata argument is not NULL interpret it as a null terminated pass
15896 phrase. If usrdata and the callback are NULL then the pass phrase
15897 is prompted for as usual.
15901 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15902 the support is automatically enabled. The resulting binaries will
15903 autodetect the card and use it if present.
15905 *Ben Laurie and Compaq Inc.*
15907 * Work around for Netscape hang bug. This sends certificate request
15908 and server done in one record. Since this is perfectly legal in the
15909 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15910 the bugs/SSLv3 entry for more info.
15914 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15918 * Add -rand argument to smime and pkcs12 applications and read/write
15923 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15927 * Add command line password options to the remaining applications.
15931 * Bug fix for BN_div_recp() for numerators with an even number of
15936 * More tests in bntest.c, and changed test_bn output.
15940 * ./config recognizes MacOS X now.
15944 * Bug fix for BN_div() when the first words of num and divisor are
15945 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15949 * Add support for various broken PKCS#8 formats, and command line
15950 options to produce them.
15954 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15955 get temporary BIGNUMs from a BN_CTX.
15959 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15964 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15965 include a #define from the old name to the new. The original intent
15966 was that statically linked binaries could for example just call
15967 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15968 link with digests. This never worked because SSLeay_add_all_digests()
15969 and SSLeay_add_all_ciphers() were in the same source file so calling
15970 one would link with the other. They are now in separate source files.
15974 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15978 * Use a less unusual form of the Miller-Rabin primality test (it used
15979 a binary algorithm for exponentiation integrated into the Miller-Rabin
15980 loop, our standard modexp algorithms are faster).
15984 * Support for the EBCDIC character set completed.
15986 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15988 * Source code cleanups: use const where appropriate, eliminate casts,
15989 use `void *` instead of `char *` in lhash.
15993 * Bugfix: ssl3_send_server_key_exchange was not restartable
15994 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15995 this the server could overwrite ephemeral keys that the client
16000 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16001 using 50 iterations of the Rabin-Miller test.
16003 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16004 iterations of the Rabin-Miller test as required by the appendix
16005 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16006 As BN_is_prime_fasttest includes trial division, DSA parameter
16007 generation becomes much faster.
16009 This implies a change for the callback functions in DSA_is_prime
16010 and DSA_generate_parameters: The callback function is called once
16011 for each positive witness in the Rabin-Miller test, not just
16012 occasionally in the inner loop; and the parameters to the
16013 callback function now provide an iteration count for the outer
16014 loop rather than for the current invocation of the inner loop.
16015 DSA_generate_parameters additionally can call the callback
16016 function with an 'iteration count' of -1, meaning that a
16017 candidate has passed the trial division test (when q is generated
16018 from an application-provided seed, trial division is skipped).
16022 * New function BN_is_prime_fasttest that optionally does trial
16023 division before starting the Rabin-Miller test and has
16024 an additional BN_CTX * argument (whereas BN_is_prime always
16025 has to allocate at least one BN_CTX).
16026 'callback(1, -1, cb_arg)' is called when a number has passed the
16027 trial division stage.
16031 * Fix for bug in CRL encoding. The validity dates weren't being handled
16036 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16040 * New function BN_pseudo_rand().
16044 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16045 bignum version of BN_from_montgomery() with the working code from
16046 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16051 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16052 made it impossible to use the same SSL_SESSION data structure in
16053 SSL2 clients in multiple threads.
16057 * The return value of RAND_load_file() no longer counts bytes obtained
16058 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16059 to seed the PRNG (previously an explicit byte count was required).
16061 *Ulf Möller, Bodo Möller*
16063 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16064 used `char *` instead of `void *` and had casts all over the place.
16068 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16072 * Retain source code compatibility for BN_prime_checks macro:
16073 BN_is_prime(..., BN_prime_checks, ...) now uses
16074 BN_prime_checks_for_size to determine the appropriate number of
16075 Rabin-Miller iterations.
16079 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16080 DH_CHECK_P_NOT_SAFE_PRIME.
16081 (Check if this is true? OpenPGP calls them "strong".)
16085 * Merge the functionality of "dh" and "gendh" programs into a new program
16086 "dhparam". The old programs are retained for now but will handle DH keys
16087 (instead of parameters) in future.
16091 * Make the ciphers, s_server and s_client programs check the return values
16092 when a new cipher list is set.
16096 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16097 ciphers. Before when the 56bit ciphers were enabled the sorting was
16100 The syntax for the cipher sorting has been extended to support sorting by
16101 cipher-strength (using the strength_bits hard coded in the tables).
16102 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16104 Fix a bug in the cipher-command parser: when supplying a cipher command
16105 string with an "undefined" symbol (neither command nor alphanumeric
16106 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16107 an error is flagged.
16109 Due to the strength-sorting extension, the code of the
16110 ssl_create_cipher_list() function was completely rearranged. I hope that
16111 the readability was also increased :-)
16113 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16115 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16116 for the first serial number and places 2 in the serial number file. This
16117 avoids problems when the root CA is created with serial number zero and
16118 the first user certificate has the same issuer name and serial number
16123 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16124 the new code. Add documentation for this stuff.
16128 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16129 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16130 structures and behave in an analogous way to the X509v3 functions:
16131 they shouldn't be called directly but wrapper functions should be used
16134 So we also now have some wrapper functions that call the X509at functions
16135 when passed certificate requests. (TO DO: similar things can be done with
16136 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16137 things. Some of these need some d2i or i2d and print functionality
16138 because they handle more complex structures.)
16142 * Add missing #ifndefs that caused missing symbols when building libssl
16143 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16144 NO_RSA in `ssl/s2*.c`.
16146 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16148 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16149 has a return value which indicates the quality of the random data
16150 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16151 error queue. New function RAND_pseudo_bytes() generates output that is
16152 guaranteed to be unique but not unpredictable. RAND_add is like
16153 RAND_seed, but takes an extra argument for an entropy estimate
16154 (RAND_seed always assumes full entropy).
16158 * Do more iterations of Rabin-Miller probable prime test (specifically,
16159 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16160 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16161 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16162 false-positive rate of at most 2^-80 for random input.
16166 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16170 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16171 in the 0.9.5 release), this returns the chain
16172 from an X509_CTX structure with a dup of the stack and all
16173 the X509 reference counts upped: so the stack will exist
16174 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16177 Also make SSL_SESSION_print() print out the verify return
16182 * Add manpage for the pkcs12 command. Also change the default
16183 behaviour so MAC iteration counts are used unless the new
16184 -nomaciter option is used. This improves file security and
16185 only older versions of MSIE (4.0 for example) need it.
16189 * Honor the no-xxx Configure options when creating .DEF files.
16193 * Add PKCS#10 attributes to field table: challengePassword,
16194 unstructuredName and unstructuredAddress. These are taken from
16195 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16196 international characters are used.
16198 More changes to X509_ATTRIBUTE code: allow the setting of types
16199 based on strings. Remove the 'loc' parameter when adding
16200 attributes because these will be a SET OF encoding which is sorted
16205 * Initial changes to the 'req' utility to allow request generation
16206 automation. This will allow an application to just generate a template
16207 file containing all the field values and have req construct the
16210 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16211 used all over the place including certificate requests and PKCS#7
16212 structures. They are currently handled manually where necessary with
16213 some primitive wrappers for PKCS#7. The new functions behave in a
16214 manner analogous to the X509 extension functions: they allow
16215 attributes to be looked up by NID and added.
16217 Later something similar to the X509V3 code would be desirable to
16218 automatically handle the encoding, decoding and printing of the
16219 more complex types. The string types like challengePassword can
16220 be handled by the string table functions.
16222 Also modified the multi byte string table handling. Now there is
16223 a 'global mask' which masks out certain types. The table itself
16224 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16225 is useful when for example there is only one permissible type
16226 (as in countryName) and using the mask might result in no valid
16231 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16232 SSL_get_peer_finished to allow applications to obtain the latest
16233 Finished messages sent to the peer or expected from the peer,
16234 respectively. (SSL_get_peer_finished is usually the Finished message
16235 actually received from the peer, otherwise the protocol will be aborted.)
16237 As the Finished message are message digests of the complete handshake
16238 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16239 be used for external authentication procedures when the authentication
16240 provided by SSL/TLS is not desired or is not enough.
16244 * Enhanced support for Alpha Linux is added. Now ./config checks if
16245 the host supports BWX extension and if Compaq C is present on the
16246 $PATH. Just exploiting of the BWX extension results in 20-30%
16247 performance kick for some algorithms, e.g. DES and RC4 to mention
16248 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16253 * Add support for MS "fast SGC". This is arguably a violation of the
16254 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16255 weak crypto and after checking the certificate is SGC a second one
16256 with strong crypto. MS SGC stops the first handshake after receiving
16257 the server certificate message and sends a second client hello. Since
16258 a server will typically do all the time consuming operations before
16259 expecting any further messages from the client (server key exchange
16260 is the most expensive) there is little difference between the two.
16262 To get OpenSSL to support MS SGC we have to permit a second client
16263 hello message after we have sent server done. In addition we have to
16264 reset the MAC if we do get this second client hello.
16268 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16269 if a DER encoded private key is RSA or DSA traditional format. Changed
16270 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16271 format DER encoded private key. Newer code should use PKCS#8 format which
16272 has the key type encoded in the ASN1 structure. Added DER private key
16273 support to pkcs8 application.
16277 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16278 ciphersuites has been selected (as required by the SSL 3/TLS 1
16279 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16280 is set, we interpret this as a request to violate the specification
16281 (the worst that can happen is a handshake failure, and 'correct'
16282 behaviour would result in a handshake failure anyway).
16286 * In SSL_CTX_add_session, take into account that there might be multiple
16287 SSL_SESSION structures with the same session ID (e.g. when two threads
16288 concurrently obtain them from an external cache).
16289 The internal cache can handle only one SSL_SESSION with a given ID,
16290 so if there's a conflict, we now throw out the old one to achieve
16295 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16296 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16297 some routines that use cipher OIDs: some ciphers do not have OIDs
16298 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16303 * Simplify the trust setting structure and code. Now we just have
16304 two sequences of OIDs for trusted and rejected settings. These will
16305 typically have values the same as the extended key usage extension
16306 and any application specific purposes.
16308 The trust checking code now has a default behaviour: it will just
16309 check for an object with the same NID as the passed id. Functions can
16310 be provided to override either the default behaviour or the behaviour
16311 for a given id. SSL client, server and email already have functions
16312 in place for compatibility: they check the NID and also return "trusted"
16313 if the certificate is self signed.
16317 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16318 traditional format into an EVP_PKEY structure.
16322 * Add a password callback function PEM_cb() which either prompts for
16323 a password if usr_data is NULL or otherwise assumes it is a null
16324 terminated password. Allow passwords to be passed on command line
16325 environment or config files in a few more utilities.
16329 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16330 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16331 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16332 Update documentation.
16336 * Support for ASN1 "NULL" type. This could be handled before by using
16337 ASN1_TYPE but there wasn't any function that would try to read a NULL
16338 and produce an error if it couldn't. For compatibility we also have
16339 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16340 don't allocate anything because they don't need to.
16344 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16347 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16349 * Rebuild of the memory allocation routines used by OpenSSL code and
16350 possibly others as well. The purpose is to make an interface that
16351 provide hooks so anyone can build a separate set of allocation and
16352 deallocation routines to be used by OpenSSL, for example memory
16353 pool implementations, or something else, which was previously hard
16354 since Malloc(), Realloc() and Free() were defined as macros having
16355 the values malloc, realloc and free, respectively (except for Win32
16356 compilations). The same is provided for memory debugging code.
16357 OpenSSL already comes with functionality to find memory leaks, but
16358 this gives people a chance to debug other memory problems.
16360 With these changes, a new set of functions and macros have appeared:
16362 CRYPTO_set_mem_debug_functions() [F]
16363 CRYPTO_get_mem_debug_functions() [F]
16364 CRYPTO_dbg_set_options() [F]
16365 CRYPTO_dbg_get_options() [F]
16366 CRYPTO_malloc_debug_init() [M]
16368 The memory debug functions are NULL by default, unless the library
16369 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16370 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16371 gives the standard debugging functions that come with OpenSSL) or
16372 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16373 provided by the library user) must be used. When the standard
16374 debugging functions are used, CRYPTO_dbg_set_options can be used to
16375 request additional information:
16376 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16377 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16379 Also, things like CRYPTO_set_mem_functions will always give the
16380 expected result (the new set of functions is used for allocation
16381 and deallocation) at all times, regardless of platform and compiler
16384 To finish it up, some functions that were never use in any other
16385 way than through macros have a new API and new semantic:
16387 CRYPTO_dbg_malloc()
16388 CRYPTO_dbg_realloc()
16391 All macros of value have retained their old syntax.
16393 *Richard Levitte and Bodo Moeller*
16395 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16396 ordering of SMIMECapabilities wasn't in "strength order" and there
16397 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16402 * Some ASN1 types with illegal zero length encoding (INTEGER,
16403 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16405 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16407 * Merge in my S/MIME library for OpenSSL. This provides a simple
16408 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16409 functionality to handle multipart/signed properly) and a utility
16410 called 'smime' to call all this stuff. This is based on code I
16411 originally wrote for Celo who have kindly allowed it to be
16412 included in OpenSSL.
16416 * Add variants des_set_key_checked and des_set_key_unchecked of
16417 des_set_key (aka des_key_sched). Global variable des_check_key
16418 decides which of these is called by des_set_key; this way
16419 des_check_key behaves as it always did, but applications and
16420 the library itself, which was buggy for des_check_key == 1,
16421 have a cleaner way to pick the version they need.
16425 * New function PKCS12_newpass() which changes the password of a
16430 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16431 dynamic mix. In both cases the ids can be used as an index into the
16432 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16433 functions so they accept a list of the field values and the
16434 application doesn't need to directly manipulate the X509_TRUST
16439 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16444 * Modify the way the V3 extension code looks up extensions. This now
16445 works in a similar way to the object code: we have some "standard"
16446 extensions in a static table which is searched with OBJ_bsearch()
16447 and the application can add dynamic ones if needed. The file
16448 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16449 updated whenever a new extension is added to the core code and kept
16450 in ext_nid order. There is a simple program 'tabtest.c' which checks
16451 this. New extensions are not added too often so this file can readily
16452 be maintained manually.
16454 There are two big advantages in doing things this way. The extensions
16455 can be looked up immediately and no longer need to be "added" using
16456 X509V3_add_standard_extensions(): this function now does nothing.
16457 Side note: I get *lots* of email saying the extension code doesn't
16458 work because people forget to call this function.
16459 Also no dynamic allocation is done unless new extensions are added:
16460 so if we don't add custom extensions there is no need to call
16461 X509V3_EXT_cleanup().
16465 * Modify enc utility's salting as follows: make salting the default. Add a
16466 magic header, so unsalted files fail gracefully instead of just decrypting
16467 to garbage. This is because not salting is a big security hole, so people
16468 should be discouraged from doing it.
16472 * Fixes and enhancements to the 'x509' utility. It allowed a message
16473 digest to be passed on the command line but it only used this
16474 parameter when signing a certificate. Modified so all relevant
16475 operations are affected by the digest parameter including the
16476 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16477 DSA key was used because it didn't fix the digest.
16481 * Initial certificate chain verify code. Currently tests the untrusted
16482 certificates for consistency with the verify purpose (which is set
16483 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16485 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16486 this is because it will reject chains with invalid extensions whereas
16487 every previous version of OpenSSL and SSLeay made no checks at all.
16489 Trust code: checks the root CA for the relevant trust settings. Trust
16490 settings have an initial value consistent with the verify purpose: e.g.
16491 if the verify purpose is for SSL client use it expects the CA to be
16492 trusted for SSL client use. However the default value can be changed to
16493 permit custom trust settings: one example of this would be to only trust
16494 certificates from a specific "secure" set of CAs.
16496 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16497 which should be used for version portability: especially since the
16498 verify structure is likely to change more often now.
16500 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16501 to set them. If not set then assume SSL clients will verify SSL servers
16504 Two new options to the verify program: -untrusted allows a set of
16505 untrusted certificates to be passed in and -purpose which sets the
16506 intended purpose of the certificate. If a purpose is set then the
16507 new chain verify code is used to check extension consistency.
16511 * Support for the authority information access extension.
16515 * Modify RSA and DSA PEM read routines to transparently handle
16516 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16517 public keys in a format compatible with certificate
16518 SubjectPublicKeyInfo structures. Unfortunately there were already
16519 functions called *_PublicKey_* which used various odd formats so
16520 these are retained for compatibility: however the DSA variants were
16521 never in a public release so they have been deleted. Changed dsa/rsa
16522 utilities to handle the new format: note no releases ever handled public
16523 keys so we should be OK.
16525 The primary motivation for this change is to avoid the same fiasco
16526 that dogs private keys: there are several incompatible private key
16527 formats some of which are standard and some OpenSSL specific and
16528 require various evil hacks to allow partial transparent handling and
16529 even then it doesn't work with DER formats. Given the option anything
16530 other than PKCS#8 should be dumped: but the other formats have to
16531 stay in the name of compatibility.
16533 With public keys and the benefit of hindsight one standard format
16534 is used which works with EVP_PKEY, RSA or DSA structures: though
16535 it clearly returns an error if you try to read the wrong kind of key.
16537 Added a -pubkey option to the 'x509' utility to output the public key.
16538 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16539 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16540 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16541 that do the same as the `EVP_PKEY_assign_*()` except they up the
16542 reference count of the added key (they don't "swallow" the
16547 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16548 CRLs would fail if the file contained no certificates or no CRLs:
16549 added a new function to read in both types and return the number
16550 read: this means that if none are read it will be an error. The
16551 DER versions of the certificate and CRL reader would always fail
16552 because it isn't possible to mix certificates and CRLs in DER format
16553 without choking one or the other routine. Changed this to just read
16554 a certificate: this is the best we can do. Also modified the code
16555 in `apps/verify.c` to take notice of return codes: it was previously
16556 attempting to read in certificates from NULL pointers and ignoring
16557 any errors: this is one reason why the cert and CRL reader seemed
16558 to work. It doesn't check return codes from the default certificate
16559 routines: these may well fail if the certificates aren't installed.
16563 * Code to support otherName option in GeneralName.
16567 * First update to verify code. Change the verify utility
16568 so it warns if it is passed a self signed certificate:
16569 for consistency with the normal behaviour. X509_verify
16570 has been modified to it will now verify a self signed
16571 certificate if *exactly* the same certificate appears
16572 in the store: it was previously impossible to trust a
16573 single self signed certificate. This means that:
16574 openssl verify ss.pem
16575 now gives a warning about a self signed certificate but
16576 openssl verify -CAfile ss.pem ss.pem
16581 * For servers, store verify_result in SSL_SESSION data structure
16582 (and add it to external session representation).
16583 This is needed when client certificate verifications fails,
16584 but an application-provided verification callback (set by
16585 SSL_CTX_set_cert_verify_callback) allows accepting the session
16586 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16587 but returns 1): When the session is reused, we have to set
16588 ssl->verify_result to the appropriate error code to avoid
16591 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16593 * Fix a bug in the new PKCS#7 code: it didn't consider the
16594 case in PKCS7_dataInit() where the signed PKCS7 structure
16595 didn't contain any existing data because it was being created.
16597 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16599 * Add a salt to the key derivation routines in enc.c. This
16600 forms the first 8 bytes of the encrypted file. Also add a
16601 -S option to allow a salt to be input on the command line.
16605 * New function X509_cmp(). Oddly enough there wasn't a function
16606 to compare two certificates. We do this by working out the SHA1
16607 hash and comparing that. X509_cmp() will be needed by the trust
16612 * SSL_get1_session() is like SSL_get_session(), but increments
16613 the reference count in the SSL_SESSION returned.
16615 *Geoff Thorpe <geoff@eu.c2.net>*
16617 * Fix for 'req': it was adding a null to request attributes.
16618 Also change the X509_LOOKUP and X509_INFO code to handle
16619 certificate auxiliary information.
16623 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16628 * Add the possibility to add extra information to the memory leak
16629 detecting output, to form tracebacks, showing from where each
16630 allocation was originated: CRYPTO_push_info("constant string") adds
16631 the string plus current file name and line number to a per-thread
16632 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16633 is like calling CYRPTO_pop_info() until the stack is empty.
16634 Also updated memory leak detection code to be multi-thread-safe.
16638 * Add options -text and -noout to pkcs7 utility and delete the
16639 encryption options which never did anything. Update docs.
16643 * Add options to some of the utilities to allow the pass phrase
16644 to be included on either the command line (not recommended on
16645 OSes like Unix) or read from the environment. Update the
16646 manpages and fix a few bugs.
16650 * Add a few manpages for some of the openssl commands.
16654 * Fix the -revoke option in ca. It was freeing up memory twice,
16655 leaking and not finding already revoked certificates.
16659 * Extensive changes to support certificate auxiliary information.
16660 This involves the use of X509_CERT_AUX structure and X509_AUX
16661 functions. An X509_AUX function such as PEM_read_X509_AUX()
16662 can still read in a certificate file in the usual way but it
16663 will also read in any additional "auxiliary information". By
16664 doing things this way a fair degree of compatibility can be
16665 retained: existing certificates can have this information added
16666 using the new 'x509' options.
16668 Current auxiliary information includes an "alias" and some trust
16669 settings. The trust settings will ultimately be used in enhanced
16670 certificate chain verification routines: currently a certificate
16671 can only be trusted if it is self signed and then it is trusted
16676 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16677 The problem was that one of the replacement routines had not been working
16678 since SSLeay releases. For now the offending routine has been replaced
16679 with non-optimised assembler. Even so, this now gives around 95%
16680 performance improvement for 1024 bit RSA signs.
16684 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16685 handling. Most clients have the effective key size in bits equal to
16686 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16687 A few however don't do this and instead use the size of the decrypted key
16688 to determine the RC2 key length and the AlgorithmIdentifier to determine
16689 the effective key length. In this case the effective key length can still
16690 be 40 bits but the key length can be 168 bits for example. This is fixed
16691 by manually forcing an RC2 key into the EVP_PKEY structure because the
16692 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16693 the key length and effective key length are equal.
16697 * Add a bunch of functions that should simplify the creation of
16698 X509_NAME structures. Now you should be able to do:
16699 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16700 and have it automatically work out the correct field type and fill in
16701 the structures. The more adventurous can try:
16702 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16703 and it will (hopefully) work out the correct multibyte encoding.
16707 * Change the 'req' utility to use the new field handling and multibyte
16708 copy routines. Before the DN field creation was handled in an ad hoc
16709 way in req, ca, and x509 which was rather broken and didn't support
16710 BMPStrings or UTF8Strings. Since some software doesn't implement
16711 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16712 using the dirstring_type option. See the new comment in the default
16713 openssl.cnf for more info.
16717 * Make crypto/rand/md_rand.c more robust:
16718 - Assure unique random numbers after fork().
16719 - Make sure that concurrent threads access the global counter and
16720 md serializably so that we never lose entropy in them
16721 or use exactly the same state in multiple threads.
16722 Access to the large state is not always serializable because
16723 the additional locking could be a performance killer, and
16724 md should be large enough anyway.
16728 * New file `apps/app_rand.c` with commonly needed functionality
16729 for handling the random seed file.
16731 Use the random seed file in some applications that previously did not:
16733 dsaparam -genkey (which also ignored its '-rand' option),
16736 x509 (when signing).
16737 Except on systems with /dev/urandom, it is crucial to have a random
16738 seed file at least for key creation, DSA signing, and for DH exchanges;
16739 for RSA signatures we could do without one.
16741 gendh and gendsa (unlike genrsa) used to read only the first byte
16742 of each file listed in the '-rand' option. The function as previously
16743 found in genrsa is now in app_rand.c and is used by all programs
16744 that support '-rand'.
16748 * In RAND_write_file, use mode 0600 for creating files;
16749 don't just chmod when it may be too late.
16753 * Report an error from X509_STORE_load_locations
16754 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16758 * New function ASN1_mbstring_copy() this copies a string in either
16759 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16760 into an ASN1_STRING type. A mask of permissible types is passed
16761 and it chooses the "minimal" type to use or an error if not type
16766 * Add function equivalents to the various macros in asn1.h. The old
16767 macros are retained with an `M_` prefix. Code inside the library can
16768 use the `M_` macros. External code (including the openssl utility)
16769 should *NOT* in order to be "shared library friendly".
16773 * Add various functions that can check a certificate's extensions
16774 to see if it usable for various purposes such as SSL client,
16775 server or S/MIME and CAs of these types. This is currently
16776 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16777 verification. Also added a -purpose flag to x509 utility to
16778 print out all the purposes.
16782 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16787 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16788 for, obtain and decode and extension and obtain its critical flag.
16789 This allows all the necessary extension code to be handled in a
16790 single function call.
16794 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16795 platforms. See crypto/rc4/rc4_enc.c for further details.
16799 * New -noout option to asn1parse. This causes no output to be produced
16800 its main use is when combined with -strparse and -out to extract data
16801 from a file (which may not be in ASN.1 format).
16805 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16806 when producing the local key id.
16808 *Richard Levitte <levitte@stacken.kth.se>*
16810 * New option -dhparam in s_server. This allows a DH parameter file to be
16811 stated explicitly. If it is not stated then it tries the first server
16812 certificate file. The previous behaviour hard coded the filename
16817 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16818 a public key to be input or output. For example:
16819 openssl rsa -in key.pem -pubout -out pubkey.pem
16820 Also added necessary DSA public key functions to handle this.
16824 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16825 in the message. This was handled by allowing
16826 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16828 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16830 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16831 to the end of the strings whereas this didn't. This would cause problems
16832 if strings read with d2i_ASN1_bytes() were later modified.
16834 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16836 * Fix for base64 decode bug. When a base64 bio reads only one line of
16837 data and it contains EOF it will end up returning an error. This is
16838 caused by input 46 bytes long. The cause is due to the way base64
16839 BIOs find the start of base64 encoded data. They do this by trying a
16840 trial decode on each line until they find one that works. When they
16841 do a flag is set and it starts again knowing it can pass all the
16842 data directly through the decoder. Unfortunately it doesn't reset
16843 the context it uses. This means that if EOF is reached an attempt
16844 is made to pass two EOFs through the context and this causes the
16845 resulting error. This can also cause other problems as well. As is
16846 usual with these problems it takes *ages* to find and the fix is
16847 trivial: move one line.
16849 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16851 * Ugly workaround to get s_client and s_server working under Windows. The
16852 old code wouldn't work because it needed to select() on sockets and the
16853 tty (for keypresses and to see if data could be written). Win32 only
16854 supports select() on sockets so we select() with a 1s timeout on the
16855 sockets and then see if any characters are waiting to be read, if none
16856 are present then we retry, we also assume we can always write data to
16857 the tty. This isn't nice because the code then blocks until we've
16858 received a complete line of data and it is effectively polling the
16859 keyboard at 1s intervals: however it's quite a bit better than not
16860 working at all :-) A dedicated Windows application might handle this
16861 with an event loop for example.
16865 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16866 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16867 will be called when RSA_sign() and RSA_verify() are used. This is useful
16868 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16869 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16870 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16871 This necessitated the support of an extra signature type NID_md5_sha1
16872 for SSL signatures and modifications to the SSL library to use it instead
16873 of calling RSA_public_decrypt() and RSA_private_encrypt().
16877 * Add new -verify -CAfile and -CApath options to the crl program, these
16878 will lookup a CRL issuers certificate and verify the signature in a
16879 similar way to the verify program. Tidy up the crl program so it
16880 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16881 less strict. It will now permit CRL extensions even if it is not
16882 a V2 CRL: this will allow it to tolerate some broken CRLs.
16886 * Initialize all non-automatic variables each time one of the openssl
16887 sub-programs is started (this is necessary as they may be started
16888 multiple times from the "OpenSSL>" prompt).
16890 *Lennart Bang, Bodo Moeller*
16892 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16893 removing all other RSA functionality (this is what NO_RSA does). This
16894 is so (for example) those in the US can disable those operations covered
16895 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16900 * Non-copying interface to BIO pairs.
16901 (still largely untested)
16905 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16906 ASCII string. This was handled independently in various places before.
16910 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16911 UTF8 strings a character at a time.
16915 * Use client_version from client hello to select the protocol
16916 (s23_srvr.c) and for RSA client key exchange verification
16917 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16921 * Add various utility functions to handle SPKACs, these were previously
16922 handled by poking round in the structure internals. Added new function
16923 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16924 print, verify and generate SPKACs. Based on an original idea from
16925 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16929 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16933 * Allow the config file extension section to be overwritten on the
16934 command line. Based on an original idea from Massimiliano Pala
16935 <madwolf@comune.modena.it>. The new option is called -extensions
16936 and can be applied to ca, req and x509. Also -reqexts to override
16937 the request extensions in req and -crlexts to override the crl extensions
16942 * Add new feature to the SPKAC handling in ca. Now you can include
16943 the same field multiple times by preceding it by "XXXX." for example:
16946 this is the same syntax as used in the req config file.
16950 * Allow certificate extensions to be added to certificate requests. These
16951 are specified in a 'req_extensions' option of the req section of the
16952 config file. They can be printed out with the -text option to req but
16953 are otherwise ignored at present.
16957 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16958 data read consists of only the final block it would not decrypted because
16959 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16960 A misplaced 'break' also meant the decrypted final block might not be
16961 copied until the next read.
16965 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16966 a few extra parameters to the DH structure: these will be useful if
16967 for example we want the value of 'q' or implement X9.42 DH.
16971 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16972 provides hooks that allow the default DSA functions or functions on a
16973 "per key" basis to be replaced. This allows hardware acceleration and
16974 hardware key storage to be handled without major modification to the
16975 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16976 associated functions.
16980 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16981 as "read only": it can't be written to and the buffer it points to will
16982 not be freed. Reading from a read only BIO is much more efficient than
16983 a normal memory BIO. This was added because there are several times when
16984 an area of memory needs to be read from a BIO. The previous method was
16985 to create a memory BIO and write the data to it, this results in two
16986 copies of the data and an O(n^2) reading algorithm. There is a new
16987 function BIO_new_mem_buf() which creates a read only memory BIO from
16988 an area of memory. Also modified the PKCS#7 routines to use read only
16993 * Bugfix: ssl23_get_client_hello did not work properly when called in
16994 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16995 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16996 but a retry condition occurred while trying to read the rest.
17000 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17001 NID_pkcs7_encrypted by default: this was wrong since this should almost
17002 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17003 the encrypted data type: this is a more sensible place to put it and it
17004 allows the PKCS#12 code to be tidied up that duplicated this
17009 * Changed obj_dat.pl script so it takes its input and output files on
17010 the command line. This should avoid shell escape redirection problems
17015 * Initial support for certificate extension requests, these are included
17016 in things like Xenroll certificate requests. Included functions to allow
17017 extensions to be obtained and added.
17021 * -crlf option to s_client and s_server for sending newlines as
17022 CRLF (as required by many protocols).
17026 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17028 * Install libRSAglue.a when OpenSSL is built with RSAref.
17030 *Ralf S. Engelschall*
17032 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17034 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17036 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17041 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17042 DH parameters/keys (q is lost during that conversion, but the resulting
17043 DH parameters contain its length).
17045 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17046 much faster than DH_generate_parameters (which creates parameters
17047 where `p = 2*q + 1`), and also the smaller q makes DH computations
17048 much more efficient (160-bit exponentiation instead of 1024-bit
17049 exponentiation); so this provides a convenient way to support DHE
17050 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17051 utter importance to use
17052 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17054 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17055 when such DH parameters are used, because otherwise small subgroup
17056 attacks may become possible!
17060 * Avoid memory leak in i2d_DHparams.
17064 * Allow the -k option to be used more than once in the enc program:
17065 this allows the same encrypted message to be read by multiple recipients.
17069 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17070 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17071 it will always use the numerical form of the OID, even if it has a short
17076 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17077 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17078 otherwise bn_mod_exp was called. In the case of hardware keys for example
17079 no private key components need be present and it might store extra data
17080 in the RSA structure, which cannot be accessed from bn_mod_exp.
17081 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17082 private key operations.
17086 * Added support for SPARC Linux.
17090 * pem_password_cb function type incompatibly changed from
17091 typedef int pem_password_cb(char *buf, int size, int rwflag);
17093 ....(char *buf, int size, int rwflag, void *userdata);
17094 so that applications can pass data to their callbacks:
17095 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17096 additional void * argument, which is just handed through whenever
17097 the password callback is called.
17099 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17101 New function SSL_CTX_set_default_passwd_cb_userdata.
17103 Compatibility note: As many C implementations push function arguments
17104 onto the stack in reverse order, the new library version is likely to
17105 interoperate with programs that have been compiled with the old
17106 pem_password_cb definition (PEM_whatever takes some data that
17107 happens to be on the stack as its last argument, and the callback
17108 just ignores this garbage); but there is no guarantee whatsoever that
17111 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17112 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17113 problems not only on Windows, but also on some Unix platforms.
17114 To avoid problematic command lines, these definitions are now in an
17115 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17116 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17120 * MIPS III/IV assembler module is reimplemented.
17124 * More DES library cleanups: remove references to srand/rand and
17125 delete an unused file.
17129 * Add support for the free Netwide assembler (NASM) under Win32,
17130 since not many people have MASM (ml) and it can be hard to obtain.
17131 This is currently experimental but it seems to work OK and pass all
17132 the tests. Check out INSTALL.W32 for info.
17136 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17137 without temporary keys kept an extra copy of the server key,
17138 and connections with temporary keys did not free everything in case
17143 * New function RSA_check_key and new openssl rsa option -check
17144 for verifying the consistency of RSA keys.
17146 *Ulf Moeller, Bodo Moeller*
17148 * Various changes to make Win32 compile work:
17149 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17150 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17151 comparison" warnings.
17152 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17156 * Add a debugging option to PKCS#5 v2 key generation function: when
17157 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17158 derived keys are printed to stderr.
17162 * Copy the flags in ASN1_STRING_dup().
17164 *Roman E. Pavlov <pre@mo.msk.ru>*
17166 * The x509 application mishandled signing requests containing DSA
17167 keys when the signing key was also DSA and the parameters didn't match.
17169 It was supposed to omit the parameters when they matched the signing key:
17170 the verifying software was then supposed to automatically use the CA's
17171 parameters if they were absent from the end user certificate.
17173 Omitting parameters is no longer recommended. The test was also
17174 the wrong way round! This was probably due to unusual behaviour in
17175 EVP_cmp_parameters() which returns 1 if the parameters match.
17176 This meant that parameters were omitted when they *didn't* match and
17177 the certificate was useless. Certificates signed with 'ca' didn't have
17180 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17182 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17183 The interface is as follows:
17184 Applications can use
17185 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17186 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17187 "off" is now the default.
17188 The library internally uses
17189 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17190 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17191 to disable memory-checking temporarily.
17193 Some inconsistent states that previously were possible (and were
17194 even the default) are now avoided.
17196 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17197 with each memory chunk allocated; this is occasionally more helpful
17198 than just having a counter.
17200 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17202 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17207 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17208 which largely parallels "options", but is for changing API behaviour,
17209 whereas "options" are about protocol behaviour.
17210 Initial "mode" flags are:
17212 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17213 a single record has been written.
17214 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17215 retries use the same buffer location.
17216 (But all of the contents must be
17221 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17224 * Fix problems with no-hmac etc.
17226 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17228 * New functions RSA_get_default_method(), RSA_set_method() and
17229 RSA_get_method(). These allows replacement of RSA_METHODs without having
17230 to mess around with the internals of an RSA structure.
17234 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17235 Also really enable memory leak checks in openssl.c and in some
17238 *Chad C. Mulligan, Bodo Moeller*
17240 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17241 up the length of negative integers. This has now been simplified to just
17242 store the length when it is first determined and use it later, rather
17243 than trying to keep track of where data is copied and updating it to
17245 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17247 * Add a new function PKCS7_signatureVerify. This allows the verification
17248 of a PKCS#7 signature but with the signing certificate passed to the
17249 function itself. This contrasts with PKCS7_dataVerify which assumes the
17250 certificate is present in the PKCS#7 structure. This isn't always the
17251 case: certificates can be omitted from a PKCS#7 structure and be
17252 distributed by "out of band" means (such as a certificate database).
17256 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17257 function prototypes in pem.h, also change util/mkdef.pl to add the
17258 necessary function names.
17262 * mk1mf.pl (used by Windows builds) did not properly read the
17263 options set by Configure in the top level Makefile, and Configure
17264 was not even able to write more than one option correctly.
17265 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17269 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17270 file to be loaded from a BIO or FILE pointer. The BIO version will
17271 for example allow memory BIOs to contain config info.
17275 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17276 Whoever hopes to achieve shared-library compatibility across versions
17277 must use this, not the compile-time macro.
17278 (Exercise 0.9.4: Which is the minimum library version required by
17280 Note: All this applies only to multi-threaded programs, others don't
17285 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17286 through a BIO pair triggered the default case, i.e.
17287 SSLerr(...,SSL_R_UNKNOWN_STATE).
17291 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17292 can use the SSL library even if none of the specific BIOs is
17297 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17298 for the encoded length.
17300 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17302 * Add initial documentation of the X509V3 functions.
17306 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17307 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17308 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17309 secure PKCS#8 private key format with a high iteration count.
17313 * Fix determination of Perl interpreter: A perl or perl5
17314 *directory* in $PATH was also accepted as the interpreter.
17316 *Ralf S. Engelschall*
17318 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17319 wrong with it but it was very old and did things like calling
17320 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17321 unusual formatting.
17325 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17326 to use the new extension code.
17330 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17331 with macros. This should make it easier to change their form, add extra
17332 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17337 * Add to configuration table a new entry that can specify an alternative
17338 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17339 according to Mark Crispin <MRC@Panda.COM>.
17343 * DES CBC did not update the IV. Weird.
17347 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17348 Changing the behaviour of the former might break existing programs --
17349 where IV updating is needed, des_ncbc_encrypt can be used.
17352 * When bntest is run from "make test" it drives bc to check its
17353 calculations, as well as internally checking them. If an internal check
17354 fails, it needs to cause bc to give a non-zero result or make test carries
17355 on without noticing the failure. Fixed.
17359 * DES library cleanups.
17363 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17364 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17365 ciphers. NOTE: although the key derivation function has been verified
17366 against some published test vectors it has not been extensively tested
17367 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17372 * Instead of "mkdir -p", which is not fully portable, use new
17373 Perl script "util/mkdir-p.pl".
17377 * Rewrite the way password based encryption (PBE) is handled. It used to
17378 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17379 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17380 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17381 the 'parameter' field of the AlgorithmIdentifier is passed to the
17382 underlying key generation function so it must do its own ASN1 parsing.
17383 This has also changed the EVP_PBE_CipherInit() function which now has a
17384 'parameter' argument instead of literal salt and iteration count values
17385 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17389 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17390 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17391 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17392 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17393 value was just used as a "magic string" and not used directly its
17394 value doesn't matter.
17398 * Introduce some semblance of const correctness to BN. Shame C doesn't
17403 * "linux-sparc64" configuration (ultrapenguin).
17405 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17406 "linux-sparc" configuration.
17408 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17410 * config now generates no-xxx options for missing ciphers.
17414 * Support the EBCDIC character set (work in progress).
17415 File ebcdic.c not yet included because it has a different license.
17417 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17419 * Support BS2000/OSD-POSIX.
17421 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17423 * Make callbacks for key generation use `void *` instead of `char *`.
17427 * Make S/MIME samples compile (not yet tested).
17431 * Additional typesafe stacks.
17435 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17439 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17441 * New configuration variant "sco5-gcc".
17443 * Updated some demos.
17445 *Sean O Riordain, Wade Scholine*
17447 * Add missing BIO_free at exit of pkcs12 application.
17451 * Fix memory leak in conf.c.
17455 * Updates for Win32 to assembler version of MD5.
17459 * Set #! path to perl in `apps/der_chop` to where we found it
17460 instead of using a fixed path.
17464 * SHA library changes for irix64-mips4-cc.
17468 * Improvements for VMS support.
17472 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17474 * Bignum library bug fix. IRIX 6 passes "make test" now!
17475 This also avoids the problems with SC4.2 and unpatched SC5.
17477 *Andy Polyakov <appro@fy.chalmers.se>*
17479 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17480 These are required because of the typesafe stack would otherwise break
17481 existing code. If old code used a structure member which used to be STACK
17482 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17483 sk_num or sk_value it would produce an error because the num, data members
17484 are not present in STACK_OF. Now it just produces a warning. sk_set
17485 replaces the old method of assigning a value to sk_value
17486 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17487 that does this will no longer work (and should use sk_set instead) but
17488 this could be regarded as a "questionable" behaviour anyway.
17492 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17493 correctly handle encrypted S/MIME data.
17497 * Change type of various DES function arguments from des_cblock
17498 (which means, in function argument declarations, pointer to char)
17499 to des_cblock * (meaning pointer to array with 8 char elements),
17500 which allows the compiler to do more typechecking; it was like
17501 that back in SSLeay, but with lots of ugly casts.
17503 Introduce new type const_des_cblock.
17507 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17508 problems: find RecipientInfo structure that matches recipient certificate
17509 and initialise the ASN1 structures properly based on passed cipher.
17513 * Belatedly make the BN tests actually check the results.
17517 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17518 to and from BNs: it was completely broken. New compilation option
17519 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17520 key elements as negative integers.
17524 * Reorganize and speed up MD5.
17526 *Andy Polyakov <appro@fy.chalmers.se>*
17530 *Richard Levitte <richard@levitte.org>*
17532 * New option -out to asn1parse to allow the parsed structure to be
17533 output to a file. This is most useful when combined with the -strparse
17534 option to examine the output of things like OCTET STRINGS.
17538 * Make SSL library a little more fool-proof by not requiring any longer
17539 that `SSL_set_{accept,connect}_state` be called before
17540 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17541 in many applications because usually everything *appeared* to work as
17542 intended anyway -- now it really works as intended).
17546 * Move openssl.cnf out of lib/.
17550 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17551 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17552 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17554 *Ralf S. Engelschall*
17556 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17557 handle PKCS#7 enveloped data properly.
17559 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17561 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17562 copying pointers. The cert_st handling is changed by this in
17563 various ways (and thus what used to be known as ctx->default_cert
17564 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17565 any longer when s->cert does not give us what we need).
17566 ssl_cert_instantiate becomes obsolete by this change.
17567 As soon as we've got the new code right (possibly it already is?),
17568 we have solved a couple of bugs of the earlier code where s->cert
17569 was used as if it could not have been shared with other SSL structures.
17571 Note that using the SSL API in certain dirty ways now will result
17572 in different behaviour than observed with earlier library versions:
17573 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17574 does not influence s as it used to.
17576 In order to clean up things more thoroughly, inside SSL_SESSION
17577 we don't use CERT any longer, but a new structure SESS_CERT
17578 that holds per-session data (if available); currently, this is
17579 the peer's certificate chain and, for clients, the server's certificate
17580 and temporary key. CERT holds only those values that can have
17581 meaningful defaults in an SSL_CTX.
17585 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17586 from the internal representation. Various PKCS#7 fixes: remove some
17587 evil casts and set the enc_dig_alg field properly based on the signing
17592 * Allow PKCS#12 password to be set from the command line or the
17593 environment. Let 'ca' get its config file name from the environment
17594 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17599 * Allow certificate policies extension to use an IA5STRING for the
17600 organization field. This is contrary to the PKIX definition but
17601 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17606 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17607 without disallowing inline assembler and the like for non-pedantic builds.
17611 * Support Borland C++ builder.
17613 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17619 * SHA-1 cleanups and performance enhancements.
17621 *Andy Polyakov <appro@fy.chalmers.se>*
17623 * Sparc v8plus assembler for the bignum library.
17625 *Andy Polyakov <appro@fy.chalmers.se>*
17627 * Accept any -xxx and +xxx compiler options in Configure.
17631 * Update HPUX configuration.
17635 * Add missing `sk_<type>_unshift()` function to safestack.h
17637 *Ralf S. Engelschall*
17639 * New function SSL_CTX_use_certificate_chain_file that sets the
17640 "extra_cert"s in addition to the certificate. (This makes sense
17641 only for "PEM" format files, as chains as a whole are not
17646 * Support verify_depth from the SSL API.
17647 x509_vfy.c had what can be considered an off-by-one-error:
17648 Its depth (which was not part of the external interface)
17649 was actually counting the number of certificates in a chain;
17650 now it really counts the depth.
17654 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17655 instead of X509err, which often resulted in confusing error
17656 messages since the error codes are not globally unique
17657 (e.g. an alleged error in ssl3_accept when a certificate
17658 didn't match the private key).
17660 * New function SSL_CTX_set_session_id_context that allows to set a default
17661 value (so that you don't need SSL_set_session_id_context for each
17662 connection using the SSL_CTX).
17666 * OAEP decoding bug fix.
17670 * Support INSTALL_PREFIX for package builders, as proposed by
17675 * New Configure options "threads" and "no-threads". For systems
17676 where the proper compiler options are known (currently Solaris
17677 and Linux), "threads" is the default.
17681 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17685 * Install various scripts to $(OPENSSLDIR)/misc, not to
17686 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17687 such as /usr/local/bin.
17691 * "make linux-shared" to build shared libraries.
17693 *Niels Poppe <niels@netbox.org>*
17695 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17699 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17700 extension adding in x509 utility.
17704 * Remove NOPROTO sections and error code comments.
17708 * Partial rewrite of the DEF file generator to now parse the ANSI
17713 * New Configure options --prefix=DIR and --openssldir=DIR.
17717 * Complete rewrite of the error code script(s). It is all now handled
17718 by one script at the top level which handles error code gathering,
17719 header rewriting and C source file generation. It should be much better
17720 than the old method: it now uses a modified version of Ulf's parser to
17721 read the ANSI prototypes in all header files (thus the old K&R definitions
17722 aren't needed for error creation any more) and do a better job of
17723 translating function codes into names. The old 'ASN1 error code embedded
17724 in a comment' is no longer necessary and it doesn't use .err files which
17725 have now been deleted. Also the error code call doesn't have to appear all
17726 on one line (which resulted in some large lines...).
17730 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17734 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17735 0 (which usually indicates a closed connection), but continue reading.
17739 * Fix some race conditions.
17743 * Add support for CRL distribution points extension. Add Certificate
17744 Policies and CRL distribution points documentation.
17748 * Move the autogenerated header file parts to crypto/opensslconf.h.
17752 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17753 8 of keying material. Merlin has also confirmed interop with this fix
17754 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17756 *Merlin Hughes <merlin@baltimore.ie>*
17758 * Fix lots of warnings.
17760 *Richard Levitte <levitte@stacken.kth.se>*
17762 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17763 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17765 *Richard Levitte <levitte@stacken.kth.se>*
17767 * Fix problems with sizeof(long) == 8.
17769 *Andy Polyakov <appro@fy.chalmers.se>*
17771 * Change functions to ANSI C.
17775 * Fix typos in error codes.
17777 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17779 * Remove defunct assembler files from Configure.
17783 * SPARC v8 assembler BIGNUM implementation.
17785 *Andy Polyakov <appro@fy.chalmers.se>*
17787 * Support for Certificate Policies extension: both print and set.
17788 Various additions to support the r2i method this uses.
17792 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17793 return a const string when you are expecting an allocated buffer.
17797 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17798 types DirectoryString and DisplayText.
17802 * Add code to allow r2i extensions to access the configuration database,
17803 add an LHASH database driver and add several ctx helper functions.
17807 * Fix an evil bug in bn_expand2() which caused various BN functions to
17808 fail when they extended the size of a BIGNUM.
17812 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17813 support typesafe stack.
17817 * Fix typo in SSL_[gs]et_options().
17819 *Nils Frostberg <nils@medcom.se>*
17821 * Delete various functions and files that belonged to the (now obsolete)
17822 old X509V3 handling code.
17826 * New Configure option "rsaref".
17830 * Don't auto-generate pem.h.
17834 * Introduce type-safe ASN.1 SETs.
17838 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17840 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17842 * Introduce type-safe STACKs. This will almost certainly break lots of code
17843 that links with OpenSSL (well at least cause lots of warnings), but fear
17844 not: the conversion is trivial, and it eliminates loads of evil casts. A
17845 few STACKed things have been converted already. Feel free to convert more.
17846 In the fullness of time, I'll do away with the STACK type altogether.
17850 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17851 specified in `<certfile>` by updating the entry in the index.txt file.
17852 This way one no longer has to edit the index.txt file manually for
17853 revoking a certificate. The -revoke option does the gory details now.
17855 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17857 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17858 `-text` option at all and this way the `-noout -text` combination was
17859 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17861 *Ralf S. Engelschall*
17863 * Make sure a corresponding plain text error message exists for the
17864 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17865 verify callback function determined that a certificate was revoked.
17867 *Ralf S. Engelschall*
17869 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17870 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17871 all available ciphers including rc5, which was forgotten until now.
17872 In order to let the testing shell script know which algorithms
17873 are available, a new (up to now undocumented) command
17874 `openssl list-cipher-commands` is used.
17878 * Bugfix: s_client occasionally would sleep in select() when
17879 it should have checked SSL_pending() first.
17883 * New functions DSA_do_sign and DSA_do_verify to provide access to
17884 the raw DSA values prior to ASN.1 encoding.
17888 * Tweaks to Configure
17890 *Niels Poppe <niels@netbox.org>*
17892 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17897 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17901 * New config option to avoid instructions that are illegal on the 80386.
17902 The default code is faster, but requires at least a 486.
17906 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17907 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17908 same as SSL2_VERSION anyway.
17912 * New "-showcerts" option for s_client.
17916 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17917 application. Various cleanups and fixes.
17921 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17922 modify error routines to work internally. Add error codes and PBE init
17923 to library startup routines.
17927 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17928 packing functions to asn1 and evp. Changed function names and error
17929 codes along the way.
17933 * PKCS12 integration: and so it begins... First of several patches to
17934 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17935 objects to objects.h
17939 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17940 and display support for Thawte strong extranet extension.
17944 * Add LinuxPPC support.
17946 *Jeff Dubrule <igor@pobox.org>*
17948 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17949 bn_div_words in alpha.s.
17951 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17953 * Make sure the RSA OAEP test is skipped under -DRSAref because
17954 OAEP isn't supported when OpenSSL is built with RSAref.
17956 *Ulf Moeller <ulf@fitug.de>*
17958 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17959 so they no longer are missing under -DNOPROTO.
17961 *Soren S. Jorvang <soren@t.dk>*
17963 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17965 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17966 doesn't work when the session is reused. Coming soon!
17970 * Fix a security hole, that allows sessions to be reused in the wrong
17971 context thus bypassing client cert protection! All software that uses
17972 client certs and session caches in multiple contexts NEEDS PATCHING to
17973 allow session reuse! A fuller solution is in the works.
17975 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17977 * Some more source tree cleanups (removed obsolete files
17978 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17979 permission on "config" script to be executable) and a fix for the INSTALL
17982 *Ulf Moeller <ulf@fitug.de>*
17984 * Remove some legacy and erroneous uses of malloc, free instead of
17987 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17989 * Make rsa_oaep_test return non-zero on error.
17991 *Ulf Moeller <ulf@fitug.de>*
17993 * Add support for native Solaris shared libraries. Configure
17994 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17995 if someone would make that last step automatic.
17997 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17999 * ctx_size was not built with the right compiler during "make links". Fixed.
18003 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18004 except NULL ciphers". This means the default cipher list will no longer
18005 enable NULL ciphers. They need to be specifically enabled e.g. with
18006 the string "DEFAULT:eNULL".
18010 * Fix to RSA private encryption routines: if p < q then it would
18011 occasionally produce an invalid result. This will only happen with
18012 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18016 * Be less restrictive and allow also `perl util/perlpath.pl
18017 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18018 because this way one can also use an interpreter named `perl5` (which is
18019 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18020 installed as `perl`).
18022 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18024 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18026 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18028 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18029 advapi32.lib to Win32 build and change the pem test comparison
18030 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18031 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18032 and crypto/des/ede_cbcm_enc.c.
18036 * DES quad checksum was broken on big-endian architectures. Fixed.
18040 * Comment out two functions in bio.h that aren't implemented. Fix up the
18041 Win32 test batch file so it (might) work again. The Win32 test batch file
18042 is horrible: I feel ill....
18046 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18047 in e_os.h. Audit of header files to check ANSI and non ANSI
18048 sections: 10 functions were absent from non ANSI section and not exported
18049 from Windows DLLs. Fixed up libeay.num for new functions.
18053 * Make `openssl version` output lines consistent.
18055 *Ralf S. Engelschall*
18057 * Fix Win32 symbol export lists for BIO functions: Added
18058 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18059 to ms/libeay{16,32}.def.
18061 *Ralf S. Engelschall*
18063 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18064 fine under Unix and passes some trivial tests I've now added. But the
18065 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18066 added to make sure no one expects that this stuff really works in the
18067 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18068 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18071 *Ralf S. Engelschall*
18073 * Fix the generation of two part addresses in perl.
18075 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18077 * Add config entry for Linux on MIPS.
18079 *John Tobey <jtobey@channel1.com>*
18081 * Make links whenever Configure is run, unless we are on Windoze.
18085 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18086 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18091 * Add a useful kludge to allow package maintainers to specify compiler and
18092 other platforms details on the command line without having to patch the
18093 Configure script every time: One now can use
18094 `perl Configure <id>:<details>`,
18095 i.e. platform ids are allowed to have details appended
18096 to them (separated by colons). This is treated as there would be a static
18097 pre-configured entry in Configure's %table under key `<id>` with value
18098 `<details>` and `perl Configure <id>` is called. So, when you want to
18099 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18100 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18101 now, which overrides the FreeBSD-elf entry on-the-fly.
18103 *Ralf S. Engelschall*
18105 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18109 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18110 on the `perl Configure ...` command line. This way one can compile
18111 OpenSSL libraries with Position Independent Code (PIC) which is needed
18112 for linking it into DSOs.
18114 *Ralf S. Engelschall*
18116 * Remarkably, export ciphers were totally broken and no-one had noticed!
18121 * Cleaned up the LICENSE document: The official contact for any license
18122 questions now is the OpenSSL core team under openssl-core@openssl.org.
18123 And add a paragraph about the dual-license situation to make sure people
18124 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18125 to the OpenSSL toolkit.
18127 *Ralf S. Engelschall*
18129 * General source tree makefile cleanups: Made `making xxx in yyy...`
18130 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18131 Additionally cleaned up the `make links` target: Remove unnecessary
18132 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18133 to speed processing and no longer clutter the display with confusing
18134 stuff. Instead only the actually done links are displayed.
18136 *Ralf S. Engelschall*
18138 * Permit null encryption ciphersuites, used for authentication only. It used
18139 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18140 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18145 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18146 signed attributes when verifying signatures (this would break them),
18147 the detached data encoding was wrong and public keys obtained using
18148 X509_get_pubkey() weren't freed.
18152 * Add text documentation for the BUFFER functions. Also added a work around
18153 to a Win95 console bug. This was triggered by the password read stuff: the
18154 last character typed gets carried over to the next fread(). If you were
18155 generating a new cert request using 'req' for example then the last
18156 character of the passphrase would be CR which would then enter the first
18161 * Added the new 'Includes OpenSSL Cryptography Software' button as
18162 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18163 button and can be used by applications based on OpenSSL to show the
18164 relationship to the OpenSSL project.
18166 *Ralf S. Engelschall*
18168 * Remove confusing variables in function signatures in files
18169 ssl/ssl_lib.c and ssl/ssl.h.
18171 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18173 * Don't install bss_file.c under PREFIX/include/
18175 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18177 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18178 functions that return function pointers and has support for NT specific
18179 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18180 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18181 unsigned to signed types: this was killing the Win32 compile.
18185 * Add new certificate file to stack functions,
18186 SSL_add_dir_cert_subjects_to_stack() and
18187 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18188 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18189 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18190 This means that Apache-SSL and similar packages don't have to mess around
18191 to add as many CAs as they want to the preferred list.
18195 * Experiment with doxygen documentation. Currently only partially applied to
18197 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18198 openssl.doxy as the configuration file.
18202 * Get rid of remaining C++-style comments which strict C compilers hate.
18204 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18206 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18207 compiled in by default: it has problems with large keys.
18211 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18212 DH private keys and/or callback functions which directly correspond to
18213 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18214 is needed for applications which have to configure certificates on a
18215 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18217 For the RSA certificate situation is makes no difference, but
18218 for the DSA certificate situation this fixes the "no shared cipher"
18219 problem where the OpenSSL cipher selection procedure failed because the
18220 temporary keys were not overtaken from the context and the API provided
18221 no way to reconfigure them.
18222 The new functions now let applications reconfigure the stuff and they
18223 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18224 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18225 non-public-API function ssl_cert_instantiate() is used as a helper
18226 function and also to reduce code redundancy inside ssl_rsa.c.
18228 *Ralf S. Engelschall*
18230 * Move s_server -dcert and -dkey options out of the undocumented feature
18231 area because they are useful for the DSA situation and should be
18232 recognized by the users.
18234 *Ralf S. Engelschall*
18236 * Fix the cipher decision scheme for export ciphers: the export bits are
18237 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18238 SSL_EXP_MASK. So, the original variable has to be used instead of the
18239 already masked variable.
18241 *Richard Levitte <levitte@stacken.kth.se>*
18243 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18245 *Richard Levitte <levitte@stacken.kth.se>*
18247 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18248 from `int` to `unsigned int` because it is a length and initialized by
18249 EVP_DigestFinal() which expects an `unsigned int *`.
18251 *Richard Levitte <levitte@stacken.kth.se>*
18253 * Don't hard-code path to Perl interpreter on shebang line of Configure
18254 script. Instead use the usual Shell->Perl transition trick.
18256 *Ralf S. Engelschall*
18258 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18259 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18260 -noout -modulus` as it's already the case for `openssl rsa -noout
18261 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18262 currently the public key is printed (a decision which was already done by
18263 `openssl dsa -modulus` in the past) which serves a similar purpose.
18264 Additionally the NO_RSA no longer completely removes the whole -modulus
18265 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18268 *Ralf S. Engelschall*
18270 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18271 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18273 *Arne Ansper <arne@ats.cyber.ee>*
18275 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18276 to be added. Now both 'req' and 'ca' can use new objects defined in the
18281 * Add cool BIO that does syslog (or event log on NT).
18283 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18285 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18286 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18287 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18288 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18292 * Add preliminary config info for new extension code.
18296 * Make RSA_NO_PADDING really use no padding.
18298 *Ulf Moeller <ulf@fitug.de>*
18300 * Generate errors when private/public key check is done.
18304 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18305 for some CRL extensions and new objects added.
18309 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18310 key usage extension and fuller support for authority key id.
18314 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18315 padding method for RSA, which is recommended for new applications in PKCS
18316 #1 v2.0 (RFC 2437, October 1998).
18317 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18318 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18319 against Bleichbacher's attack on RSA.
18320 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18323 * Updates to the new SSL compression code
18325 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18327 * Fix so that the version number in the master secret, when passed
18328 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18329 (because the server will not accept higher), that the version number
18330 is 0x03,0x01, not 0x03,0x00
18332 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18334 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18335 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18336 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18340 * Support for RAW extensions where an arbitrary extension can be
18341 created by including its DER encoding. See `apps/openssl.cnf` for
18346 * Make sure latest Perl versions don't interpret some generated C array
18347 code as Perl array code in the crypto/err/err_genc.pl script.
18349 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18351 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18352 not many people have the assembler. Various Win32 compilation fixes and
18353 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18354 build instructions.
18358 * Modify configure script 'Configure' to automatically create crypto/date.h
18359 file under Win32 and also build pem.h from pem.org. New script
18360 util/mkfiles.pl to create the MINFO file on environments that can't do a
18361 'make files': perl util/mkfiles.pl >MINFO should work.
18365 * Major rework of DES function declarations, in the pursuit of correctness
18366 and purity. As a result, many evil casts evaporated, and some weirdness,
18367 too. You may find this causes warnings in your code. Zapping your evil
18368 casts will probably fix them. Mostly.
18372 * Fix for a typo in asn1.h. Bug fix to object creation script
18373 obj_dat.pl. It considered a zero in an object definition to mean
18374 "end of object": none of the objects in objects.h have any zeros
18375 so it wasn't spotted.
18377 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18379 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18380 Masking (CBCM). In the absence of test vectors, the best I have been able
18381 to do is check that the decrypt undoes the encrypt, so far. Send me test
18382 vectors if you have them.
18386 * Correct calculation of key length for export ciphers (too much space was
18387 allocated for null ciphers). This has not been tested!
18391 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18392 message is now correct (it understands "crypto" and "ssl" on its
18393 command line). There is also now an "update" option. This will update
18394 the util/ssleay.num and util/libeay.num files with any new functions.
18396 perl util/mkdef.pl crypto ssl update
18397 it will update them.
18401 * Overhauled the Perl interface:
18402 - ported BN stuff to OpenSSL's different BN library
18403 - made the perl/ source tree CVS-aware
18404 - renamed the package from SSLeay to OpenSSL (the files still contain
18405 their history because I've copied them in the repository)
18406 - removed obsolete files (the test scripts will be replaced
18407 by better Test::Harness variants in the future)
18409 *Ralf S. Engelschall*
18411 * First cut for a very conservative source tree cleanup:
18412 1. merge various obsolete readme texts into doc/ssleay.txt
18413 where we collect the old documents and readme texts.
18414 2. remove the first part of files where I'm already sure that we no
18415 longer need them because of three reasons: either they are just temporary
18416 files which were left by Eric or they are preserved original files where
18417 I've verified that the diff is also available in the CVS via "cvs diff
18418 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18419 the crypto/md/ stuff).
18421 *Ralf S. Engelschall*
18423 * More extension code. Incomplete support for subject and issuer alt
18424 name, issuer and authority key id. Change the i2v function parameters
18425 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18426 what that's for :-) Fix to ASN1 macro which messed up
18427 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18431 * Preliminary support for ENUMERATED type. This is largely copied from the
18436 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18438 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18440 * Make sure `make rehash` target really finds the `openssl` program.
18442 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18444 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18445 like to hear about it if this slows down other processors.
18449 * Add CygWin32 platform information to Configure script.
18451 *Alan Batie <batie@aahz.jf.intel.com>*
18453 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18455 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18457 * New program nseq to manipulate netscape certificate sequences
18461 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18466 * Fixes to BN code. Previously the default was to define BN_RECURSION
18467 but the BN code had some problems that would cause failures when
18468 doing certificate verification and some other functions.
18470 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18472 * Add ASN1 and PEM code to support netscape certificate sequences.
18476 * Add ASN1 and PEM code to support netscape certificate sequences.
18480 * Add several PKIX and private extended key usage OIDs.
18484 * Modify the 'ca' program to handle the new extension code. Modify
18485 openssl.cnf for new extension format, add comments.
18489 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18490 and add a sample to openssl.cnf so req -x509 now adds appropriate
18495 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18496 error code, add initial support to X509_print() and x509 application.
18500 * Takes a deep breath and start adding X509 V3 extension support code. Add
18501 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18502 stuff is currently isolated and isn't even compiled yet.
18506 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18507 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18508 Removed the versions check from X509 routines when loading extensions:
18509 this allows certain broken certificates that don't set the version
18510 properly to be processed.
18514 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18515 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18516 can still be regenerated with "make depend".
18520 * Spelling mistake in C version of CAST-128.
18522 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18524 * Changes to the error generation code. The perl script err-code.pl
18525 now reads in the old error codes and retains the old numbers, only
18526 adding new ones if necessary. It also only changes the .err files if new
18527 codes are added. The makefiles have been modified to only insert errors
18528 when needed (to avoid needlessly modifying header files). This is done
18529 by only inserting errors if the .err file is newer than the auto generated
18530 C file. To rebuild all the error codes from scratch (the old behaviour)
18531 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18532 or delete all the .err files.
18536 * CAST-128 was incorrectly implemented for short keys. The C version has
18537 been fixed, but is untested. The assembler versions are also fixed, but
18538 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18539 to regenerate it if needed.
18540 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18541 Hagino <itojun@kame.net>*
18543 * File was opened incorrectly in randfile.c.
18545 *Ulf Möller <ulf@fitug.de>*
18547 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18548 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18549 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18550 al: it's just almost always a UTCTime. Note this patch adds new error
18551 codes so do a "make errors" if there are problems.
18555 * Correct Linux 1 recognition in config.
18557 *Ulf Möller <ulf@fitug.de>*
18559 * Remove pointless MD5 hash when using DSA keys in ca.
18561 *Anonymous <nobody@replay.com>*
18563 * Generate an error if given an empty string as a cert directory. Also
18564 generate an error if handed NULL (previously returned 0 to indicate an
18565 error, but didn't set one).
18567 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18569 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18573 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18574 parameters. This was causing a warning which killed off the Win32 compile.
18578 * Remove C++ style comments from crypto/bn/bn_local.h.
18580 *Neil Costigan <neil.costigan@celocom.com>*
18582 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18583 based on a text string, looking up short and long names and finally
18584 "dot" format. The "dot" format stuff didn't work. Added new function
18585 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18586 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18587 OID is not part of the table.
18591 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18592 X509_LOOKUP_by_alias().
18596 * Sort openssl functions by name.
18600 * Get the `gendsa` command working and add it to the `list` command. Remove
18601 encryption from sample DSA keys (in case anyone is interested the password
18606 * Make *all* `*_free` functions accept a NULL pointer.
18608 *Frans Heymans <fheymans@isaserver.be>*
18610 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18613 *Anonymous <nobody@replay.com>*
18615 * s_server should send the CAfile as acceptable CAs, not its own cert.
18617 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18619 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18621 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18623 * Temp key "for export" tests were wrong in s3_srvr.c.
18625 *Anonymous <nobody@replay.com>*
18627 * Add prototype for temp key callback functions
18628 SSL_CTX_set_tmp_{rsa,dh}_callback().
18632 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18633 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18637 * X509_name_add_entry() freed the wrong thing after an error.
18639 *Arne Ansper <arne@ats.cyber.ee>*
18641 * rsa_eay.c would attempt to free a NULL context.
18643 *Arne Ansper <arne@ats.cyber.ee>*
18645 * BIO_s_socket() had a broken should_retry() on Windoze.
18647 *Arne Ansper <arne@ats.cyber.ee>*
18649 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18651 *Arne Ansper <arne@ats.cyber.ee>*
18653 * Make sure the already existing X509_STORE->depth variable is initialized
18654 in X509_STORE_new(), but document the fact that this variable is still
18655 unused in the certificate verification process.
18657 *Ralf S. Engelschall*
18659 * Fix the various library and `apps/` files to free up pkeys obtained from
18660 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18664 * Fix reference counting in X509_PUBKEY_get(). This makes
18665 demos/maurice/example2.c work, amongst others, probably.
18667 *Steve Henson and Ben Laurie*
18669 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18670 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18671 are no longer created. This way we have a single and consistent command
18672 line interface `openssl <command>`, similar to `cvs <command>`.
18674 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18676 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18677 BIT STRING wrapper always have zero unused bits.
18681 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18685 * Make the top-level INSTALL documentation easier to understand.
18689 * Makefiles updated to exit if an error occurs in a sub-directory
18690 make (including if user presses ^C) [Paul Sutton]
18692 * Make Montgomery context stuff explicit in RSA data structure.
18696 * Fix build order of pem and err to allow for generated pem.h.
18700 * Fix renumbering bug in X509_NAME_delete_entry().
18704 * Enhanced the err-ins.pl script so it makes the error library number
18705 global and can add a library name. This is needed for external ASN1 and
18706 other error libraries.
18710 * Fixed sk_insert which never worked properly.
18714 * Fix ASN1 macros so they can handle indefinite length constructed
18715 EXPLICIT tags. Some non standard certificates use these: they can now
18720 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18721 into a single doc/ssleay.txt bundle. This way the information is still
18722 preserved but no longer messes up this directory. Now it's new room for
18723 the new set of documentation files.
18725 *Ralf S. Engelschall*
18727 * SETs were incorrectly DER encoded. This was a major pain, because they
18728 shared code with SEQUENCEs, which aren't coded the same. This means that
18729 almost everything to do with SETs or SEQUENCEs has either changed name or
18730 number of arguments.
18732 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18734 * Fix test data to work with the above.
18738 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18739 was already fixed by Eric for 0.9.1 it seems.
18741 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18743 * Autodetect FreeBSD3.
18747 * Fix various bugs in Configure. This affects the following platforms:
18751 unixware-2.0-pentium
18756 * Eliminate generated files from CVS. Reorder tests to regenerate files
18757 before they are needed.
18761 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18765 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18767 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18768 changed SSLeay to OpenSSL in version strings.
18770 *Ralf S. Engelschall*
18772 * Some fixups to the top-level documents.
18776 * Fixed the nasty bug where rsaref.h was not found under compile-time
18777 because the symlink to include/ was missing.
18779 *Ralf S. Engelschall*
18781 * Incorporated the popular no-RSA/DSA-only patches
18782 which allow to compile a RSA-free SSLeay.
18784 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18786 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18787 when "ssleay" is still not found.
18789 *Ralf S. Engelschall*
18791 * Added more platforms to Configure: Cray T3E, HPUX 11,
18793 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18795 * Updated the README file.
18797 *Ralf S. Engelschall*
18799 * Added various .cvsignore files in the CVS repository subdirs
18800 to make a "cvs update" really silent.
18802 *Ralf S. Engelschall*
18804 * Recompiled the error-definition header files and added
18805 missing symbols to the Win32 linker tables.
18807 *Ralf S. Engelschall*
18809 * Cleaned up the top-level documents;
18810 o new files: CHANGES and LICENSE
18811 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18812 o merged COPYRIGHT into LICENSE
18813 o removed obsolete TODO file
18814 o renamed MICROSOFT to INSTALL.W32
18816 *Ralf S. Engelschall*
18818 * Removed dummy files from the 0.9.1b source tree:
18819 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18820 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18821 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18822 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18823 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18825 *Ralf S. Engelschall*
18827 * Added various platform portability fixes.
18831 * The Genesis of the OpenSSL rpject:
18832 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18833 Young and Tim J. Hudson created while they were working for C2Net until
18836 *The OpenSSL Project*
18838 ### Changes between 0.9.0b and 0.9.1b [not released]
18840 * Updated a few CA certificates under certs/
18844 * Changed some BIGNUM api stuff.
18848 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18849 DGUX x86, Linux Alpha, etc.
18853 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18854 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18859 * Add -strparse option to asn1pars program which parses nested
18862 *Dr Stephen Henson <shenson@bigfoot.com>*
18864 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18868 * DSA fix for "ca" program.
18872 * Added "-genkey" option to "dsaparam" program.
18876 * Added RIPE MD160 (rmd160) message digest.
18880 * Added -a (all) option to "ssleay version" command.
18884 * Added PLATFORM define which is the id given to Configure.
18888 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18892 * Extended the ASN.1 parser routines.
18896 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18900 * Added a BN_CTX to the BN library.
18904 * Fixed the weak key values in DES library
18908 * Changed API in EVP library for cipher aliases.
18912 * Added support for RC2/64bit cipher.
18916 * Converted the lhash library to the crypto/mem.c functions.
18920 * Added more recognized ASN.1 object ids.
18924 * Added more RSA padding checks for SSL/TLS.
18928 * Added BIO proxy/filter functionality.
18932 * Added extra_certs to SSL_CTX which can be used
18933 send extra CA certificates to the client in the CA cert chain sending
18934 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18938 * Now Fortezza is denied in the authentication phase because
18939 this is key exchange mechanism is not supported by SSLeay at all.
18943 * Additional PKCS1 checks.
18947 * Support the string "TLSv1" for all TLS v1 ciphers.
18951 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18952 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18956 * Fixed a few memory leaks.
18960 * Fixed various code and comment typos.
18964 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18965 bytes sent in the client random.
18967 *Edward Bishop <ebishop@spyglass.com>*
18971 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18972 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18973 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18974 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18975 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18976 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18977 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18978 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18979 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18980 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18981 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18982 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18983 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18984 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18985 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18986 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18987 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18988 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18989 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18990 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18991 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18992 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18993 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18994 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18995 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18996 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18997 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18998 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18999 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19000 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19001 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19002 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19003 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19004 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19005 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19006 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19007 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19008 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19009 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19010 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19011 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19012 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19013 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19014 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19015 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19016 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19017 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19018 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19019 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19020 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19021 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19022 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19023 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19024 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19025 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19026 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19027 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19028 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19029 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19030 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19031 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19032 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19033 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19034 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19035 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19036 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19037 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19038 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19039 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19040 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19041 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19042 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19043 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19044 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19045 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19046 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19047 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19048 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19049 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19050 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19051 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19052 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19053 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19054 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19055 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19056 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19057 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19058 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19059 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19060 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19061 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19062 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19063 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19064 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19065 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19066 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19067 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19068 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19069 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19070 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19071 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19072 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19073 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19074 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19075 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19076 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19077 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19078 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19079 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19080 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19081 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19082 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19083 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19084 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19085 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19086 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19087 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19088 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19089 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19090 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19091 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19092 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19093 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19094 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19095 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19096 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19097 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19098 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19099 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19100 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19101 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19102 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19103 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19104 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19105 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19106 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19107 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19108 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19109 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19110 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19111 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19112 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19113 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19114 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19115 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19116 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19117 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19118 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19119 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19120 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19121 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19122 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19123 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19124 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19125 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19126 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19127 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19128 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19129 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19130 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19131 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19132 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655