5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
8 have been moved out of the public header files. New functions for managing
12 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
13 objects have been moved out of the public header files. New
14 functions for managing these have been added.
17 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
18 have been moved out of the public header files. New functions for managing
19 these have been added.
22 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
23 moved out of the public header files. New functions for managing these
27 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
30 *) Removed the mk1mf build scripts.
33 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
34 it is always safe to #include a header now.
37 *) Removed the aged BC-32 config and all its supporting scripts
40 *) Removed support for Ultrix, Netware, and OS/2.
43 *) Add support for HKDF.
46 *) Add support for blake2b and blake2s
49 *) Added support for "pipelining". Ciphers that have the
50 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
51 encryptions/decryptions simultaneously. There are currently no built-in
52 ciphers with this property but the expectation is that engines will be able
53 to offer it to significantly improve throughput. Support has been extended
54 into libssl so that multiple records for a single connection can be
55 processed in one go (for >=TLS 1.1).
58 *) Added the AFALG engine. This is an async capable engine which is able to
59 offload work to the Linux kernel. In this initial version it only supports
60 AES128-CBC. The kernel must be version 4.1.0 or greater.
63 *) OpenSSL now uses a new threading API. It is no longer necessary to
64 set locking callbacks to use OpenSSL in a multi-threaded environment. There
65 are two supported threading models: pthreads and windows threads. It is
66 also possible to configure OpenSSL at compile time for "no-threads". The
67 old threading API should no longer be used. The functions have been
68 replaced with "no-op" compatibility macros.
69 [Alessandro Ghedini, Matt Caswell]
71 *) Modify behavior of ALPN to invoke callback after SNI/servername
72 callback, such that updates to the SSL_CTX affect ALPN.
75 *) Add SSL_CIPHER queries for authentication and key-exchange.
78 *) Changes to the DEFAULT cipherlist:
79 - Prefer (EC)DHE handshakes over plain RSA.
80 - Prefer AEAD ciphers over legacy ciphers.
81 - Prefer ECDSA over RSA when both certificates are available.
82 - Prefer TLSv1.2 ciphers/PRF.
83 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
87 *) Change the ECC default curve list to be this, in order: x25519,
88 secp256r1, secp521r1, secp384r1.
91 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
92 disabled by default. They can be re-enabled using the
93 enable-weak-ssl-ciphers option to Configure.
96 *) If the server has ALPN configured, but supports no protocols that the
97 client advertises, send a fatal "no_application_protocol" alert.
98 This behaviour is SHALL in RFC 7301, though it isn't universally
99 implemented by other servers.
102 *) Add X25519 support.
103 Integrate support for X25519 into EC library. This includes support
104 for public and private key encoding using the format documented in
105 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
106 OID from that draft, encodes public keys using little endian
107 format in the ECPoint structure and private keys using
108 little endian form in the privateKey field of the ECPrivateKey
109 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
112 Note: the current version supports key generation, public and
113 private key encoding and ECDH key agreement using the EC API.
114 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
118 *) Deprecate SRP_VBASE_get_by_user.
119 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
120 In order to fix an unavoidable memory leak (CVE-2016-0798),
121 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
122 seed, even if the seed is configured.
124 Users should use SRP_VBASE_get1_by_user instead. Note that in
125 SRP_VBASE_get1_by_user, caller must free the returned value. Note
126 also that even though configuring the SRP seed attempts to hide
127 invalid usernames by continuing the handshake with fake
128 credentials, this behaviour is not constant time and no strong
129 guarantees are made that the handshake is indistinguishable from
130 that of a valid user.
133 *) Configuration change; it's now possible to build dynamic engines
134 without having to build shared libraries and vice versa. This
135 only applies to the engines in engines/, those in crypto/engine/
136 will always be built into libcrypto (i.e. "static").
138 Building dynamic engines is enabled by default; to disable, use
139 the configuration option "disable-dynamic-engine".
141 The only requirements for building dynamic engines are the
142 presence of the DSO module and building with position independent
143 code, so they will also automatically be disabled if configuring
144 with "disable-dso" or "disable-pic".
146 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
147 are also taken away from openssl/opensslconf.h, as they are
151 *) Configuration change; if there is a known flag to compile
152 position independent code, it will always be applied on the
153 libcrypto and libssl object files, and never on the application
154 object files. This means other libraries that use routines from
155 libcrypto / libssl can be made into shared libraries regardless
156 of how OpenSSL was configured.
158 If this isn't desirable, the configuration options "disable-pic"
159 or "no-pic" can be used to disable the use of PIC. This will
160 also disable building shared libraries and dynamic engines.
163 *) Removed JPAKE code. It was experimental and has no wide use.
166 *) The INSTALL_PREFIX Makefile variable has been renamed to
167 DESTDIR. That makes for less confusion on what this variable
168 is for. Also, the configuration option --install_prefix is
172 *) Heartbeat for TLS has been removed and is disabled by default
173 for DTLS; configure with enable-heartbeats. Code that uses the
174 old #define's might need to be updated.
175 [Emilia Käsper, Rich Salz]
177 *) Rename REF_CHECK to REF_DEBUG.
180 *) New "unified" build system
182 The "unified" build system is aimed to be a common system for all
183 platforms we support. With it comes new support for VMS.
185 This system builds supports building in a different directory tree
186 than the source tree. It produces one Makefile (for unix family
187 or lookalikes), or one descrip.mms (for VMS).
189 The source of information to make the Makefile / descrip.mms is
190 small files called 'build.info', holding the necessary
191 information for each directory with source to compile, and a
192 template in Configurations, like unix-Makefile.tmpl or
195 We rely heavily on the perl module Text::Template.
198 *) Added support for auto-initialisation and de-initialisation of the library.
199 OpenSSL no longer requires explicit init or deinit routines to be called,
200 except in certain circumstances. See the OPENSSL_init_crypto() and
201 OPENSSL_init_ssl() man pages for further information.
204 *) The arguments to the DTLSv1_listen function have changed. Specifically the
205 "peer" argument is now expected to be a BIO_ADDR object.
207 *) Rewrite of BIO networking library. The BIO library lacked consistent
208 support of IPv6, and adding it required some more extensive
209 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
210 which hold all types of addresses and chains of address information.
211 It also introduces a new API, with functions like BIO_socket,
212 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
213 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
214 have been adapted accordingly.
217 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
221 *) CRIME protection: disable compression by default, even if OpenSSL is
222 compiled with zlib enabled. Applications can still enable compression
223 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
224 using the SSL_CONF library to configure compression.
227 *) The signature of the session callback configured with
228 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
229 was explicitly marked as 'const unsigned char*' instead of
233 *) Always DPURIFY. Remove the use of uninitialized memory in the
234 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
237 *) Removed many obsolete configuration items, including
238 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
239 MD2_CHAR, MD2_INT, MD2_LONG
241 IDEA_SHORT, IDEA_LONG
242 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
243 [Rich Salz, with advice from Andy Polyakov]
245 *) Many BN internals have been moved to an internal header file.
246 [Rich Salz with help from Andy Polyakov]
248 *) Configuration and writing out the results from it has changed.
249 Files such as Makefile include/openssl/opensslconf.h and are now
250 produced through general templates, such as Makefile.in and
251 crypto/opensslconf.h.in and some help from the perl module
254 Also, the center of configuration information is no longer
255 Makefile. Instead, Configure produces a perl module in
256 configdata.pm which holds most of the config data (in the hash
257 table %config), the target data that comes from the target
258 configuration in one of the Configurations/*.conf files (in
262 *) To clarify their intended purposes, the Configure options
263 --prefix and --openssldir change their semantics, and become more
264 straightforward and less interdependent.
266 --prefix shall be used exclusively to give the location INSTALLTOP
267 where programs, scripts, libraries, include files and manuals are
268 going to be installed. The default is now /usr/local.
270 --openssldir shall be used exclusively to give the default
271 location OPENSSLDIR where certificates, private keys, CRLs are
272 managed. This is also where the default openssl.cnf gets
274 If the directory given with this option is a relative path, the
275 values of both the --prefix value and the --openssldir value will
276 be combined to become OPENSSLDIR.
277 The default for --openssldir is INSTALLTOP/ssl.
279 Anyone who uses --openssldir to specify where OpenSSL is to be
280 installed MUST change to use --prefix instead.
283 *) The GOST engine was out of date and therefore it has been removed. An up
284 to date GOST engine is now being maintained in an external repository.
285 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
286 support for GOST ciphersuites (these are only activated if a GOST engine
290 *) EGD is no longer supported by default; use enable-egd when
292 [Ben Kaduk and Rich Salz]
294 *) The distribution now has Makefile.in files, which are used to
295 create Makefile's when Configure is run. *Configure must be run
296 before trying to build now.*
299 *) The return value for SSL_CIPHER_description() for error conditions
303 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
305 Obtaining and performing DNSSEC validation of TLSA records is
306 the application's responsibility. The application provides
307 the TLSA records of its choice to OpenSSL, and these are then
308 used to authenticate the peer.
310 The TLSA records need not even come from DNS. They can, for
311 example, be used to implement local end-entity certificate or
312 trust-anchor "pinning", where the "pin" data takes the form
313 of TLSA records, which can augment or replace verification
314 based on the usual WebPKI public certification authorities.
317 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
318 continues to support deprecated interfaces in default builds.
319 However, applications are strongly advised to compile their
320 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
321 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
322 or the 1.1.0 releases.
324 In environments in which all applications have been ported to
325 not use any deprecated interfaces OpenSSL's Configure script
326 should be used with the --api=1.1.0 option to entirely remove
327 support for the deprecated features from the library and
328 unconditionally disable them in the installed headers.
329 Essentially the same effect can be achieved with the "no-deprecated"
330 argument to Configure, except that this will always restrict
331 the build to just the latest API, rather than a fixed API
334 As applications are ported to future revisions of the API,
335 they should update their compile-time OPENSSL_API_COMPAT define
336 accordingly, but in most cases should be able to continue to
337 compile with later releases.
339 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
340 0x10000000L and 0x00908000L, respectively. However those
341 versions did not support the OPENSSL_API_COMPAT feature, and
342 so applications are not typically tested for explicit support
343 of just the undeprecated features of either release.
346 *) Add support for setting the minimum and maximum supported protocol.
347 It can bet set via the SSL_set_min_proto_version() and
348 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
349 MaxProtcol. It's recommended to use the new APIs to disable
350 protocols instead of disabling individual protocols using
351 SSL_set_options() or SSL_CONF's Protocol. This change also
352 removes support for disabling TLS 1.2 in the OpenSSL TLS
353 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
356 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
359 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
360 and integrates ECDSA and ECDH functionality into EC. Implementations can
361 now redirect key generation and no longer need to convert to or from
364 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
365 include the ec.h header file instead.
368 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
369 ciphers who are no longer supported and drops support the ephemeral RSA key
370 exchange. The LOW ciphers currently doesn't have any ciphers in it.
373 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
374 opaque. For HMAC_CTX, the following constructors and destructors
377 HMAC_CTX *HMAC_CTX_new(void);
378 void HMAC_CTX_free(HMAC_CTX *ctx);
380 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
381 destroy such methods has been added. See EVP_MD_meth_new(3) and
382 EVP_CIPHER_meth_new(3) for documentation.
385 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
386 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
387 EVP_MD_CTX_reset() should be called instead to reinitialise
388 an already created structure.
389 2) For consistency with the majority of our object creators and
390 destructors, EVP_MD_CTX_(create|destroy) were renamed to
391 EVP_MD_CTX_(new|free). The old names are retained as macros
392 for deprecated builds.
395 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
396 cryptographic operations to be performed asynchronously as long as an
397 asynchronous capable engine is used. See the ASYNC_start_job() man page for
398 further details. Libssl has also had this capability integrated with the
399 introduction of the new mode SSL_MODE_ASYNC and associated error
400 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
401 pages. This work was developed in partnership with Intel Corp.
404 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
405 always enabled now. If you want to disable the support you should
406 exclude it using the list of supported ciphers. This also means that the
407 "-no_ecdhe" option has been removed from s_server.
410 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
411 SSL_{CTX_}set1_curves() which can set a list.
414 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
415 curve you want to support using SSL_{CTX_}set1_curves().
418 *) State machine rewrite. The state machine code has been significantly
419 refactored in order to remove much duplication of code and solve issues
420 with the old code (see ssl/statem/README for further details). This change
421 does have some associated API changes. Notably the SSL_state() function
422 has been removed and replaced by SSL_get_state which now returns an
423 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
424 altogether. The previous handshake states defined in ssl.h and ssl3.h have
428 *) All instances of the string "ssleay" in the public API were replaced
429 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
430 Some error codes related to internal RSA_eay API's were renamed.
433 *) The demo files in crypto/threads were moved to demo/threads.
436 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
438 [Matt Caswell, Rich Salz]
440 *) New ASN.1 embed macro.
442 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
443 structure is not allocated: it is part of the parent. That is instead of
451 This reduces memory fragmentation and make it impossible to accidentally
452 set a mandatory field to NULL.
454 This currently only works for some fields specifically a SEQUENCE, CHOICE,
455 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
456 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
460 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
463 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
464 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
465 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
466 DES and RC4 ciphersuites.
469 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
470 This changes the decoding behaviour for some invalid messages,
471 though the change is mostly in the more lenient direction, and
472 legacy behaviour is preserved as much as possible.
475 *) Fix no-stdio build.
476 [ David Woodhouse <David.Woodhouse@intel.com> and also
477 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
479 *) New testing framework
480 The testing framework has been largely rewritten and is now using
481 perl and the perl modules Test::Harness and an extended variant of
482 Test::More called OpenSSL::Test to do its work. All test scripts in
483 test/ have been rewritten into test recipes, and all direct calls to
484 executables in test/Makefile have become individual recipes using the
485 simplified testing OpenSSL::Test::Simple.
487 For documentation on our testing modules, do:
489 perldoc test/testlib/OpenSSL/Test/Simple.pm
490 perldoc test/testlib/OpenSSL/Test.pm
494 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
495 are used; the latter aborts on memory leaks (usually checked on exit).
496 Some undocumented "set malloc, etc., hooks" functions were removed
497 and others were changed. All are now documented.
500 *) In DSA_generate_parameters_ex, if the provided seed is too short,
502 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
504 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
505 from RFC4279, RFC4785, RFC5487, RFC5489.
507 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
508 original RSA_PSK patch.
511 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
512 era flag was never set throughout the codebase (only read). Also removed
513 SSL3_FLAGS_POP_BUFFER which was only used if
514 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
517 *) Changed the default name options in the "ca", "crl", "req" and "x509"
518 to be "oneline" instead of "compat".
521 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
522 not aware of clients that still exhibit this bug, and the workaround
523 hasn't been working properly for a while.
526 *) The return type of BIO_number_read() and BIO_number_written() as well as
527 the corresponding num_read and num_write members in the BIO structure has
528 changed from unsigned long to uint64_t. On platforms where an unsigned
529 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
533 *) Given the pervasive nature of TLS extensions it is inadvisable to run
534 OpenSSL without support for them. It also means that maintaining
535 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
536 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
539 *) Removed support for the two export grade static DH ciphersuites
540 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
541 were newly added (along with a number of other static DH ciphersuites) to
542 1.0.2. However the two export ones have *never* worked since they were
543 introduced. It seems strange in any case to be adding new export
544 ciphersuites, and given "logjam" it also does not seem correct to fix them.
547 *) Version negotiation has been rewritten. In particular SSLv23_method(),
548 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
549 and turned into macros which simply call the new preferred function names
550 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
551 should use the new names instead. Also as part of this change the ssl23.h
552 header file has been removed.
555 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
556 code and the associated standard is no longer considered fit-for-purpose.
559 *) RT2547 was closed. When generating a private key, try to make the
560 output file readable only by the owner. This behavior change might
561 be noticeable when interacting with other software.
563 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
567 *) Added HTTP GET support to the ocsp command.
570 *) Changed default digest for the dgst and enc commands from MD5 to
574 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
577 *) Added support for TLS extended master secret from
578 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
579 initial patch which was a great help during development.
582 *) All libssl internal structures have been removed from the public header
583 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
584 now redundant). Users should not attempt to access internal structures
585 directly. Instead they should use the provided API functions.
588 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
589 Access to deprecated functions can be re-enabled by running config with
590 "enable-deprecated". In addition applications wishing to use deprecated
591 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
592 will, by default, disable some transitive includes that previously existed
593 in the header files (e.g. ec.h will no longer, by default, include bn.h)
596 *) Added support for OCB mode. OpenSSL has been granted a patent license
597 compatible with the OpenSSL license for use of OCB. Details are available
598 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
599 for OCB can be removed by calling config with no-ocb.
602 *) SSLv2 support has been removed. It still supports receiving a SSLv2
603 compatible client hello.
606 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
607 done while fixing the error code for the key-too-small case.
608 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
610 *) CA.sh has been removmed; use CA.pl instead.
613 *) Removed old DES API.
616 *) Remove various unsupported platforms:
622 Sinix/ReliantUNIX RM400
627 16-bit platforms such as WIN16
630 *) Clean up OPENSSL_NO_xxx #define's
631 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
632 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
633 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
634 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
635 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
636 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
637 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
638 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
639 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
640 Remove MS_STATIC; it's a relic from platforms <32 bits.
643 *) Cleaned up dead code
644 Remove all but one '#ifdef undef' which is to be looked at.
647 *) Clean up calling of xxx_free routines.
648 Just like free(), fix most of the xxx_free routines to accept
649 NULL. Remove the non-null checks from callers. Save much code.
652 *) Add secure heap for storage of private keys (when possible).
653 Add BIO_s_secmem(), CBIGNUM, etc.
654 Contributed by Akamai Technologies under our Corporate CLA.
657 *) Experimental support for a new, fast, unbiased prime candidate generator,
658 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
659 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
661 *) New output format NSS in the sess_id command line tool. This allows
662 exporting the session id and the master key in NSS keylog format.
663 [Martin Kaiser <martin@kaiser.cx>]
665 *) Harmonize version and its documentation. -f flag is used to display
667 [mancha <mancha1@zoho.com>]
669 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
670 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
671 [mancha <mancha1@zoho.com>]
673 *) Fix some double frees. These are not thought to be exploitable.
674 [mancha <mancha1@zoho.com>]
676 *) A missing bounds check in the handling of the TLS heartbeat extension
677 can be used to reveal up to 64k of memory to a connected client or
680 Thanks for Neel Mehta of Google Security for discovering this bug and to
681 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
682 preparing the fix (CVE-2014-0160)
683 [Adam Langley, Bodo Moeller]
685 *) Fix for the attack described in the paper "Recovering OpenSSL
686 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
687 by Yuval Yarom and Naomi Benger. Details can be obtained from:
688 http://eprint.iacr.org/2014/140
690 Thanks to Yuval Yarom and Naomi Benger for discovering this
691 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
692 [Yuval Yarom and Naomi Benger]
694 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
695 this fixes a limitation in previous versions of OpenSSL.
698 *) Experimental encrypt-then-mac support.
700 Experimental support for encrypt then mac from
701 draft-gutmann-tls-encrypt-then-mac-02.txt
703 To enable it set the appropriate extension number (0x42 for the test
704 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
706 For non-compliant peers (i.e. just about everything) this should have no
709 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
713 *) Add EVP support for key wrapping algorithms, to avoid problems with
714 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
715 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
716 algorithms and include tests cases.
719 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
723 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
724 MGF1 digest and OAEP label.
727 *) Make openssl verify return errors.
728 [Chris Palmer <palmer@google.com> and Ben Laurie]
730 *) New function ASN1_TIME_diff to calculate the difference between two
731 ASN1_TIME structures or one structure and the current time.
734 *) Update fips_test_suite to support multiple command line options. New
735 test to induce all self test errors in sequence and check expected
739 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
740 sign or verify all in one operation.
743 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
744 test programs and fips_test_suite. Includes functionality to parse
745 the minimal script output of fipsalgest.pl directly.
748 *) Add authorisation parameter to FIPS_module_mode_set().
751 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
754 *) Use separate DRBG fields for internal and external flags. New function
755 FIPS_drbg_health_check() to perform on demand health checking. Add
756 generation tests to fips_test_suite with reduced health check interval to
757 demonstrate periodic health checking. Add "nodh" option to
758 fips_test_suite to skip very slow DH test.
761 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
765 *) More extensive health check for DRBG checking many more failure modes.
766 New function FIPS_selftest_drbg_all() to handle every possible DRBG
767 combination: call this in fips_test_suite.
770 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
771 and POST to handle Dual EC cases.
774 *) Add support for canonical generation of DSA parameter 'g'. See
777 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
778 POST to handle HMAC cases.
781 *) Add functions FIPS_module_version() and FIPS_module_version_text()
782 to return numerical and string versions of the FIPS module number.
785 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
786 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
787 outside the validated module in the FIPS capable OpenSSL.
790 *) Minor change to DRBG entropy callback semantics. In some cases
791 there is no multiple of the block length between min_len and
792 max_len. Allow the callback to return more than max_len bytes
793 of entropy but discard any extra: it is the callback's responsibility
794 to ensure that the extra data discarded does not impact the
795 requested amount of entropy.
798 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
799 information in FIPS186-3, SP800-57 and SP800-131A.
802 *) CCM support via EVP. Interface is very similar to GCM case except we
803 must supply all data in one chunk (i.e. no update, final) and the
804 message length must be supplied if AAD is used. Add algorithm test
808 *) Initial version of POST overhaul. Add POST callback to allow the status
809 of POST to be monitored and/or failures induced. Modify fips_test_suite
810 to use callback. Always run all selftests even if one fails.
813 *) XTS support including algorithm test driver in the fips_gcmtest program.
814 Note: this does increase the maximum key length from 32 to 64 bytes but
815 there should be no binary compatibility issues as existing applications
816 will never use XTS mode.
819 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
820 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
821 performs algorithm blocking for unapproved PRNG types. Also do not
822 set PRNG type in FIPS_mode_set(): leave this to the application.
823 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
824 the standard OpenSSL PRNG: set additional data to a date time vector.
827 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
828 This shouldn't present any incompatibility problems because applications
829 shouldn't be using these directly and any that are will need to rethink
830 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
833 *) Extensive self tests and health checking required by SP800-90 DRBG.
834 Remove strength parameter from FIPS_drbg_instantiate and always
835 instantiate at maximum supported strength.
838 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
841 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
844 *) New function DH_compute_key_padded() to compute a DH key and pad with
845 leading zeroes if needed: this complies with SP800-56A et al.
848 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
849 anything, incomplete, subject to change and largely untested at present.
852 *) Modify fipscanisteronly build option to only build the necessary object
853 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
856 *) Add experimental option FIPSSYMS to give all symbols in
857 fipscanister.o and FIPS or fips prefix. This will avoid
858 conflicts with future versions of OpenSSL. Add perl script
859 util/fipsas.pl to preprocess assembly language source files
860 and rename any affected symbols.
863 *) Add selftest checks and algorithm block of non-fips algorithms in
864 FIPS mode. Remove DES2 from selftests.
867 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
868 return internal method without any ENGINE dependencies. Add new
869 tiny fips sign and verify functions.
872 *) New build option no-ec2m to disable characteristic 2 code.
875 *) New build option "fipscanisteronly". This only builds fipscanister.o
876 and (currently) associated fips utilities. Uses the file Makefile.fips
877 instead of Makefile.org as the prototype.
880 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
881 Update fips_gcmtest to use IV generator.
884 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
885 setting output buffer to NULL. The *Final function must be
886 called although it will not retrieve any additional data. The tag
887 can be set or retrieved with a ctrl. The IV length is by default 12
888 bytes (96 bits) but can be set to an alternative value. If the IV
889 length exceeds the maximum IV length (currently 16 bytes) it cannot be
893 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
894 underlying do_cipher function handles all cipher semantics itself
895 including padding and finalisation. This is useful if (for example)
896 an ENGINE cipher handles block padding itself. The behaviour of
897 do_cipher is subtly changed if this flag is set: the return value
898 is the number of characters written to the output buffer (zero is
899 no longer an error code) or a negative error code. Also if the
900 input buffer is NULL and length 0 finalisation should be performed.
903 *) If a candidate issuer certificate is already part of the constructed
904 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
907 *) Improve forward-security support: add functions
909 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
910 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
912 for use by SSL/TLS servers; the callback function will be called whenever a
913 new session is created, and gets to decide whether the session may be
914 cached to make it resumable (return 0) or not (return 1). (As by the
915 SSL/TLS protocol specifications, the session_id sent by the server will be
916 empty to indicate that the session is not resumable; also, the server will
917 not generate RFC 4507 (RFC 5077) session tickets.)
919 A simple reasonable callback implementation is to return is_forward_secure.
920 This parameter will be set to 1 or 0 depending on the ciphersuite selected
921 by the SSL/TLS server library, indicating whether it can provide forward
923 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
925 *) New -verify_name option in command line utilities to set verification
929 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
930 Add CMAC pkey methods.
933 *) Experimental renegotiation in s_server -www mode. If the client
934 browses /reneg connection is renegotiated. If /renegcert it is
935 renegotiated requesting a certificate.
938 *) Add an "external" session cache for debugging purposes to s_server. This
939 should help trace issues which normally are only apparent in deployed
940 multi-process servers.
943 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
944 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
945 BIO_set_cipher() and some obscure PEM functions were changed so they
946 can now return an error. The RAND changes required a change to the
947 RAND_METHOD structure.
950 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
951 a gcc attribute to warn if the result of a function is ignored. This
952 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
953 whose return value is often ignored.
956 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
957 These allow SCTs (signed certificate timestamps) to be requested and
958 validated when establishing a connection.
959 [Rob Percival <robpercival@google.com>]
961 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
963 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
964 Builds that are not configured with "enable-weak-ssl-ciphers" will not
965 provide any "EXPORT" or "LOW" strength ciphers.
968 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
969 is by default disabled at build-time. Builds that are not configured with
970 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
971 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
972 will need to explicitly call either of:
974 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
976 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
978 as appropriate. Even if either of those is used, or the application
979 explicitly uses the version-specific SSLv2_method() or its client and
980 server variants, SSLv2 ciphers vulnerable to exhaustive search key
981 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
982 ciphers, and SSLv2 56-bit DES are no longer available.
986 *) Fix a double-free in DSA code
988 A double free bug was discovered when OpenSSL parses malformed DSA private
989 keys and could lead to a DoS attack or memory corruption for applications
990 that receive DSA private keys from untrusted sources. This scenario is
993 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
998 *) Disable SRP fake user seed to address a server memory leak.
1000 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1002 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1003 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1004 was changed to ignore the "fake user" SRP seed, even if the seed
1007 Users should use SRP_VBASE_get1_by_user instead. Note that in
1008 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1009 also that even though configuring the SRP seed attempts to hide
1010 invalid usernames by continuing the handshake with fake
1011 credentials, this behaviour is not constant time and no strong
1012 guarantees are made that the handshake is indistinguishable from
1013 that of a valid user.
1017 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1019 In the BN_hex2bn function the number of hex digits is calculated using an
1020 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1021 large values of |i| this can result in |bn_expand| not allocating any
1022 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1023 field as NULL leading to a subsequent NULL ptr deref. For very large values
1024 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1025 In this case memory is allocated to the internal BIGNUM data field, but it
1026 is insufficiently sized leading to heap corruption. A similar issue exists
1027 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1028 is ever called by user applications with very large untrusted hex/dec data.
1029 This is anticipated to be a rare occurrence.
1031 All OpenSSL internal usage of these functions use data that is not expected
1032 to be untrusted, e.g. config file data or application command line
1033 arguments. If user developed applications generate config file data based
1034 on untrusted data then it is possible that this could also lead to security
1035 consequences. This is also anticipated to be rare.
1037 This issue was reported to OpenSSL by Guido Vranken.
1041 *) Fix memory issues in BIO_*printf functions
1043 The internal |fmtstr| function used in processing a "%s" format string in
1044 the BIO_*printf functions could overflow while calculating the length of a
1045 string and cause an OOB read when printing very long strings.
1047 Additionally the internal |doapr_outch| function can attempt to write to an
1048 OOB memory location (at an offset from the NULL pointer) in the event of a
1049 memory allocation failure. In 1.0.2 and below this could be caused where
1050 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1051 could be in processing a very long "%s" format string. Memory leaks can
1054 The first issue may mask the second issue dependent on compiler behaviour.
1055 These problems could enable attacks where large amounts of untrusted data
1056 is passed to the BIO_*printf functions. If applications use these functions
1057 in this way then they could be vulnerable. OpenSSL itself uses these
1058 functions when printing out human-readable dumps of ASN.1 data. Therefore
1059 applications that print this data could be vulnerable if the data is from
1060 untrusted sources. OpenSSL command line applications could also be
1061 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1062 as command line arguments.
1064 Libssl is not considered directly vulnerable. Additionally certificates etc
1065 received via remote connections via libssl are also unlikely to be able to
1066 trigger these issues because of message size limits enforced within libssl.
1068 This issue was reported to OpenSSL Guido Vranken.
1072 *) Side channel attack on modular exponentiation
1074 A side-channel attack was found which makes use of cache-bank conflicts on
1075 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1076 of RSA keys. The ability to exploit this issue is limited as it relies on
1077 an attacker who has control of code in a thread running on the same
1078 hyper-threaded core as the victim thread which is performing decryptions.
1080 This issue was reported to OpenSSL by Yuval Yarom, The University of
1081 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1082 Nadia Heninger, University of Pennsylvania with more information at
1083 http://cachebleed.info.
1087 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1088 if no keysize is specified with default_bits. This fixes an
1089 omission in an earlier change that changed all RSA/DSA key generation
1090 apps to use 2048 bits by default.
1093 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1094 *) DH small subgroups
1096 Historically OpenSSL only ever generated DH parameters based on "safe"
1097 primes. More recently (in version 1.0.2) support was provided for
1098 generating X9.42 style parameter files such as those required for RFC 5114
1099 support. The primes used in such files may not be "safe". Where an
1100 application is using DH configured with parameters based on primes that are
1101 not "safe" then an attacker could use this fact to find a peer's private
1102 DH exponent. This attack requires that the attacker complete multiple
1103 handshakes in which the peer uses the same private DH exponent. For example
1104 this could be used to discover a TLS server's private DH exponent if it's
1105 reusing the private DH exponent or it's using a static DH ciphersuite.
1107 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1108 TLS. It is not on by default. If the option is not set then the server
1109 reuses the same private DH exponent for the life of the server process and
1110 would be vulnerable to this attack. It is believed that many popular
1111 applications do set this option and would therefore not be at risk.
1113 The fix for this issue adds an additional check where a "q" parameter is
1114 available (as is the case in X9.42 based parameters). This detects the
1115 only known attack, and is the only possible defense for static DH
1116 ciphersuites. This could have some performance impact.
1118 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1119 default and cannot be disabled. This could have some performance impact.
1121 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1125 *) SSLv2 doesn't block disabled ciphers
1127 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1128 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1129 been disabled, provided that the SSLv2 protocol was not also disabled via
1132 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1133 and Sebastian Schinzel.
1137 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1139 *) BN_mod_exp may produce incorrect results on x86_64
1141 There is a carry propagating bug in the x86_64 Montgomery squaring
1142 procedure. No EC algorithms are affected. Analysis suggests that attacks
1143 against RSA and DSA as a result of this defect would be very difficult to
1144 perform and are not believed likely. Attacks against DH are considered just
1145 feasible (although very difficult) because most of the work necessary to
1146 deduce information about a private key may be performed offline. The amount
1147 of resources required for such an attack would be very significant and
1148 likely only accessible to a limited number of attackers. An attacker would
1149 additionally need online access to an unpatched system using the target
1150 private key in a scenario with persistent DH parameters and a private
1151 key that is shared between multiple clients. For example this can occur by
1152 default in OpenSSL DHE based SSL/TLS ciphersuites.
1154 This issue was reported to OpenSSL by Hanno Böck.
1158 *) Certificate verify crash with missing PSS parameter
1160 The signature verification routines will crash with a NULL pointer
1161 dereference if presented with an ASN.1 signature using the RSA PSS
1162 algorithm and absent mask generation function parameter. Since these
1163 routines are used to verify certificate signature algorithms this can be
1164 used to crash any certificate verification operation and exploited in a
1165 DoS attack. Any application which performs certificate verification is
1166 vulnerable including OpenSSL clients and servers which enable client
1169 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1173 *) X509_ATTRIBUTE memory leak
1175 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1176 memory. This structure is used by the PKCS#7 and CMS routines so any
1177 application which reads PKCS#7 or CMS data from untrusted sources is
1178 affected. SSL/TLS is not affected.
1180 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1185 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1186 This changes the decoding behaviour for some invalid messages,
1187 though the change is mostly in the more lenient direction, and
1188 legacy behaviour is preserved as much as possible.
1191 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1193 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1195 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1197 *) Alternate chains certificate forgery
1199 During certificate verfification, OpenSSL will attempt to find an
1200 alternative certificate chain if the first attempt to build such a chain
1201 fails. An error in the implementation of this logic can mean that an
1202 attacker could cause certain checks on untrusted certificates to be
1203 bypassed, such as the CA flag, enabling them to use a valid leaf
1204 certificate to act as a CA and "issue" an invalid certificate.
1206 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1210 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1212 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1213 incompatibility in the handling of HMAC. The previous ABI has now been
1217 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1219 *) Malformed ECParameters causes infinite loop
1221 When processing an ECParameters structure OpenSSL enters an infinite loop
1222 if the curve specified is over a specially malformed binary polynomial
1225 This can be used to perform denial of service against any
1226 system which processes public keys, certificate requests or
1227 certificates. This includes TLS clients and TLS servers with
1228 client authentication enabled.
1230 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1234 *) Exploitable out-of-bounds read in X509_cmp_time
1236 X509_cmp_time does not properly check the length of the ASN1_TIME
1237 string and can read a few bytes out of bounds. In addition,
1238 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1241 An attacker can use this to craft malformed certificates and CRLs of
1242 various sizes and potentially cause a segmentation fault, resulting in
1243 a DoS on applications that verify certificates or CRLs. TLS clients
1244 that verify CRLs are affected. TLS clients and servers with client
1245 authentication enabled may be affected if they use custom verification
1248 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1249 independently by Hanno Böck.
1253 *) PKCS7 crash with missing EnvelopedContent
1255 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1256 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1257 with missing content and trigger a NULL pointer dereference on parsing.
1259 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1260 structures from untrusted sources are affected. OpenSSL clients and
1261 servers are not affected.
1263 This issue was reported to OpenSSL by Michal Zalewski (Google).
1267 *) CMS verify infinite loop with unknown hash function
1269 When verifying a signedData message the CMS code can enter an infinite loop
1270 if presented with an unknown hash function OID. This can be used to perform
1271 denial of service against any system which verifies signedData messages using
1273 This issue was reported to OpenSSL by Johannes Bauer.
1277 *) Race condition handling NewSessionTicket
1279 If a NewSessionTicket is received by a multi-threaded client when attempting to
1280 reuse a previous ticket then a race condition can occur potentially leading to
1281 a double free of the ticket data.
1285 *) Only support 256-bit or stronger elliptic curves with the
1286 'ecdh_auto' setting (server) or by default (client). Of supported
1287 curves, prefer P-256 (both).
1290 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1292 *) ClientHello sigalgs DoS fix
1294 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1295 invalid signature algorithms extension a NULL pointer dereference will
1296 occur. This can be exploited in a DoS attack against the server.
1298 This issue was was reported to OpenSSL by David Ramos of Stanford
1301 [Stephen Henson and Matt Caswell]
1303 *) Multiblock corrupted pointer fix
1305 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1306 feature only applies on 64 bit x86 architecture platforms that support AES
1307 NI instructions. A defect in the implementation of "multiblock" can cause
1308 OpenSSL's internal write buffer to become incorrectly set to NULL when
1309 using non-blocking IO. Typically, when the user application is using a
1310 socket BIO for writing, this will only result in a failed connection.
1311 However if some other BIO is used then it is likely that a segmentation
1312 fault will be triggered, thus enabling a potential DoS attack.
1314 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1318 *) Segmentation fault in DTLSv1_listen fix
1320 The DTLSv1_listen function is intended to be stateless and processes the
1321 initial ClientHello from many peers. It is common for user code to loop
1322 over the call to DTLSv1_listen until a valid ClientHello is received with
1323 an associated cookie. A defect in the implementation of DTLSv1_listen means
1324 that state is preserved in the SSL object from one invocation to the next
1325 that can lead to a segmentation fault. Errors processing the initial
1326 ClientHello can trigger this scenario. An example of such an error could be
1327 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1330 This issue was reported to OpenSSL by Per Allansson.
1334 *) Segmentation fault in ASN1_TYPE_cmp fix
1336 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1337 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1338 certificate signature algorithm consistency this can be used to crash any
1339 certificate verification operation and exploited in a DoS attack. Any
1340 application which performs certificate verification is vulnerable including
1341 OpenSSL clients and servers which enable client authentication.
1345 *) Segmentation fault for invalid PSS parameters fix
1347 The signature verification routines will crash with a NULL pointer
1348 dereference if presented with an ASN.1 signature using the RSA PSS
1349 algorithm and invalid parameters. Since these routines are used to verify
1350 certificate signature algorithms this can be used to crash any
1351 certificate verification operation and exploited in a DoS attack. Any
1352 application which performs certificate verification is vulnerable including
1353 OpenSSL clients and servers which enable client authentication.
1355 This issue was was reported to OpenSSL by Brian Carpenter.
1359 *) ASN.1 structure reuse memory corruption fix
1361 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1362 memory corruption via an invalid write. Such reuse is and has been
1363 strongly discouraged and is believed to be rare.
1365 Applications that parse structures containing CHOICE or ANY DEFINED BY
1366 components may be affected. Certificate parsing (d2i_X509 and related
1367 functions) are however not affected. OpenSSL clients and servers are
1372 *) PKCS7 NULL pointer dereferences fix
1374 The PKCS#7 parsing code does not handle missing outer ContentInfo
1375 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1376 missing content and trigger a NULL pointer dereference on parsing.
1378 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1379 otherwise parse PKCS#7 structures from untrusted sources are
1380 affected. OpenSSL clients and servers are not affected.
1382 This issue was reported to OpenSSL by Michal Zalewski (Google).
1386 *) DoS via reachable assert in SSLv2 servers fix
1388 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1389 servers that both support SSLv2 and enable export cipher suites by sending
1390 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1392 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1393 (OpenSSL development team).
1397 *) Empty CKE with client auth and DHE fix
1399 If client auth is used then a server can seg fault in the event of a DHE
1400 ciphersuite being selected and a zero length ClientKeyExchange message
1401 being sent by the client. This could be exploited in a DoS attack.
1405 *) Handshake with unseeded PRNG fix
1407 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1408 with an unseeded PRNG. The conditions are:
1409 - The client is on a platform where the PRNG has not been seeded
1410 automatically, and the user has not seeded manually
1411 - A protocol specific client method version has been used (i.e. not
1412 SSL_client_methodv23)
1413 - A ciphersuite is used that does not require additional random data from
1414 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1416 If the handshake succeeds then the client random that has been used will
1417 have been generated from a PRNG with insufficient entropy and therefore the
1418 output may be predictable.
1420 For example using the following command with an unseeded openssl will
1421 succeed on an unpatched platform:
1423 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1427 *) Use After Free following d2i_ECPrivatekey error fix
1429 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1430 could cause a use after free condition. This, in turn, could cause a double
1431 free in several private key parsing functions (such as d2i_PrivateKey
1432 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1433 for applications that receive EC private keys from untrusted
1434 sources. This scenario is considered rare.
1436 This issue was discovered by the BoringSSL project and fixed in their
1441 *) X509_to_X509_REQ NULL pointer deref fix
1443 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1444 the certificate key is invalid. This function is rarely used in practice.
1446 This issue was discovered by Brian Carpenter.
1450 *) Removed the export ciphers from the DEFAULT ciphers
1453 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1455 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1456 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1457 So far those who have to target multiple plaforms would compromise
1458 and argue that binary targeting say ARMv5 would still execute on
1459 ARMv8. "Universal" build resolves this compromise by providing
1460 near-optimal performance even on newer platforms.
1463 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1464 (other platforms pending).
1465 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1467 *) Add support for the SignedCertificateTimestampList certificate and
1468 OCSP response extensions from RFC6962.
1471 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1472 for corner cases. (Certain input points at infinity could lead to
1473 bogus results, with non-infinity inputs mapped to infinity too.)
1476 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1477 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1478 common cases are optimized and there still is room for further
1479 improvements. Vector Permutation AES for Altivec is also added.
1482 *) Add support for little-endian ppc64 Linux target.
1483 [Marcelo Cerri (IBM)]
1485 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1486 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1487 are optimized and there still is room for further improvements.
1488 Both 32- and 64-bit modes are supported.
1489 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1491 *) Improved ARMv7 NEON support.
1494 *) Support for SPARC Architecture 2011 crypto extensions, first
1495 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1496 SHA256/512, MD5, GHASH and modular exponentiation.
1497 [Andy Polyakov, David Miller]
1499 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1501 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1503 *) Support for new and upcoming Intel processors, including AVX2,
1504 BMI and SHA ISA extensions. This includes additional "stitched"
1505 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1508 This work was sponsored by Intel Corp.
1511 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1512 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1513 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1516 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1517 this fixes a limiation in previous versions of OpenSSL.
1520 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1521 MGF1 digest and OAEP label.
1524 *) Add EVP support for key wrapping algorithms, to avoid problems with
1525 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1526 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1527 algorithms and include tests cases.
1530 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1532 [Douglas E. Engert, Steve Henson]
1534 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1535 difference in days and seconds between two tm or ASN1_TIME structures.
1538 *) Add -rev test option to s_server to just reverse order of characters
1539 received by client and send back to server. Also prints an abbreviated
1540 summary of the connection parameters.
1543 *) New option -brief for s_client and s_server to print out a brief summary
1544 of connection parameters.
1547 *) Add callbacks for arbitrary TLS extensions.
1548 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1550 *) New option -crl_download in several openssl utilities to download CRLs
1551 from CRLDP extension in certificates.
1554 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1557 *) New function X509_CRL_diff to generate a delta CRL from the difference
1558 of two full CRLs. Add support to "crl" utility.
1561 *) New functions to set lookup_crls function and to retrieve
1562 X509_STORE from X509_STORE_CTX.
1565 *) Print out deprecated issuer and subject unique ID fields in
1569 *) Extend OCSP I/O functions so they can be used for simple general purpose
1570 HTTP as well as OCSP. New wrapper function which can be used to download
1571 CRLs using the OCSP API.
1574 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1577 *) SSL_CONF* functions. These provide a common framework for application
1578 configuration using configuration files or command lines.
1581 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1582 message callback and prints the results. Needs compile time option
1583 "enable-ssl-trace". New options to s_client and s_server to enable
1587 *) New ctrl and macro to retrieve supported points extensions.
1588 Print out extension in s_server and s_client.
1591 *) New functions to retrieve certificate signature and signature
1595 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1599 *) New Suite B modes for TLS code. These use and enforce the requirements
1600 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1601 only use Suite B curves. The Suite B modes can be set by using the
1602 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1605 *) New chain verification flags for Suite B levels of security. Check
1606 algorithms are acceptable when flags are set in X509_verify_cert.
1609 *) Make tls1_check_chain return a set of flags indicating checks passed
1610 by a certificate chain. Add additional tests to handle client
1611 certificates: checks for matching certificate type and issuer name
1615 *) If an attempt is made to use a signature algorithm not in the peer
1616 preference list abort the handshake. If client has no suitable
1617 signature algorithms in response to a certificate request do not
1618 use the certificate.
1621 *) If server EC tmp key is not in client preference list abort handshake.
1624 *) Add support for certificate stores in CERT structure. This makes it
1625 possible to have different stores per SSL structure or one store in
1626 the parent SSL_CTX. Include distint stores for certificate chain
1627 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1628 to build and store a certificate chain in CERT structure: returing
1629 an error if the chain cannot be built: this will allow applications
1630 to test if a chain is correctly configured.
1632 Note: if the CERT based stores are not set then the parent SSL_CTX
1633 store is used to retain compatibility with existing behaviour.
1637 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1638 mask based on the current session, check mask when sending client
1639 hello and checking the requested ciphersuite.
1642 *) New ctrls to retrieve and set certificate types in a certificate
1643 request message. Print out received values in s_client. If certificate
1644 types is not set with custom values set sensible values based on
1645 supported signature algorithms.
1648 *) Support for distinct client and server supported signature algorithms.
1651 *) Add certificate callback. If set this is called whenever a certificate
1652 is required by client or server. An application can decide which
1653 certificate chain to present based on arbitrary criteria: for example
1654 supported signature algorithms. Add very simple example to s_server.
1655 This fixes many of the problems and restrictions of the existing client
1656 certificate callback: for example you can now clear an existing
1657 certificate and specify the whole chain.
1660 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1661 the certificate can be used for (if anything). Set valid_flags field
1662 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1663 to have similar checks in it.
1665 Add new "cert_flags" field to CERT structure and include a "strict mode".
1666 This enforces some TLS certificate requirements (such as only permitting
1667 certificate signature algorithms contained in the supported algorithms
1668 extension) which some implementations ignore: this option should be used
1669 with caution as it could cause interoperability issues.
1672 *) Update and tidy signature algorithm extension processing. Work out
1673 shared signature algorithms based on preferences and peer algorithms
1674 and print them out in s_client and s_server. Abort handshake if no
1675 shared signature algorithms.
1678 *) Add new functions to allow customised supported signature algorithms
1679 for SSL and SSL_CTX structures. Add options to s_client and s_server
1683 *) New function SSL_certs_clear() to delete all references to certificates
1684 from an SSL structure. Before this once a certificate had been added
1685 it couldn't be removed.
1688 *) Integrate hostname, email address and IP address checking with certificate
1689 verification. New verify options supporting checking in opensl utility.
1692 *) Fixes and wildcard matching support to hostname and email checking
1693 functions. Add manual page.
1694 [Florian Weimer (Red Hat Product Security Team)]
1696 *) New functions to check a hostname email or IP address against a
1697 certificate. Add options x509 utility to print results of checks against
1701 *) Fix OCSP checking.
1702 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1704 *) Initial experimental support for explicitly trusted non-root CAs.
1705 OpenSSL still tries to build a complete chain to a root but if an
1706 intermediate CA has a trust setting included that is used. The first
1707 setting is used: whether to trust (e.g., -addtrust option to the x509
1711 *) Add -trusted_first option which attempts to find certificates in the
1712 trusted store even if an untrusted chain is also supplied.
1715 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1716 platform support for Linux and Android.
1719 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1722 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1723 When in FIPS mode the approved implementations are used as normal,
1724 when not in FIPS mode the internal unapproved versions are used instead.
1725 This means that the FIPS capable OpenSSL isn't forced to use the
1726 (often lower performance) FIPS implementations outside FIPS mode.
1729 *) Transparently support X9.42 DH parameters when calling
1730 PEM_read_bio_DHparameters. This means existing applications can handle
1731 the new parameter format automatically.
1734 *) Initial experimental support for X9.42 DH parameter format: mainly
1735 to support use of 'q' parameter for RFC5114 parameters.
1738 *) Add DH parameters from RFC5114 including test data to dhtest.
1741 *) Support for automatic EC temporary key parameter selection. If enabled
1742 the most preferred EC parameters are automatically used instead of
1743 hardcoded fixed parameters. Now a server just has to call:
1744 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1745 support ECDH and use the most appropriate parameters.
1748 *) Enhance and tidy EC curve and point format TLS extension code. Use
1749 static structures instead of allocation if default values are used.
1750 New ctrls to set curves we wish to support and to retrieve shared curves.
1751 Print out shared curves in s_server. New options to s_server and s_client
1752 to set list of supported curves.
1755 *) New ctrls to retrieve supported signature algorithms and
1756 supported curve values as an array of NIDs. Extend openssl utility
1757 to print out received values.
1760 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1761 between NIDs and the more common NIST names such as "P-256". Enhance
1762 ecparam utility and ECC method to recognise the NIST names for curves.
1765 *) Enhance SSL/TLS certificate chain handling to support different
1766 chains for each certificate instead of one chain in the parent SSL_CTX.
1769 *) Support for fixed DH ciphersuite client authentication: where both
1770 server and client use DH certificates with common parameters.
1773 *) Support for fixed DH ciphersuites: those requiring DH server
1777 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1779 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1780 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1781 X509_CINF_get_signature were reverted post internal team review.
1783 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1785 *) Build fixes for the Windows and OpenVMS platforms
1786 [Matt Caswell and Richard Levitte]
1788 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1790 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1791 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1792 dereference. This could lead to a Denial Of Service attack. Thanks to
1793 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1797 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1798 dtls1_buffer_record function under certain conditions. In particular this
1799 could occur if an attacker sent repeated DTLS records with the same
1800 sequence number but for the next epoch. The memory leak could be exploited
1801 by an attacker in a Denial of Service attack through memory exhaustion.
1802 Thanks to Chris Mueller for reporting this issue.
1806 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1807 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1808 method would be set to NULL which could later result in a NULL pointer
1809 dereference. Thanks to Frank Schmirler for reporting this issue.
1813 *) Abort handshake if server key exchange message is omitted for ephemeral
1816 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1817 reporting this issue.
1821 *) Remove non-export ephemeral RSA code on client and server. This code
1822 violated the TLS standard by allowing the use of temporary RSA keys in
1823 non-export ciphersuites and could be used by a server to effectively
1824 downgrade the RSA key length used to a value smaller than the server
1825 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1826 INRIA or reporting this issue.
1830 *) Fixed issue where DH client certificates are accepted without verification.
1831 An OpenSSL server will accept a DH certificate for client authentication
1832 without the certificate verify message. This effectively allows a client to
1833 authenticate without the use of a private key. This only affects servers
1834 which trust a client certificate authority which issues certificates
1835 containing DH keys: these are extremely rare and hardly ever encountered.
1836 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1841 *) Ensure that the session ID context of an SSL is updated when its
1842 SSL_CTX is updated via SSL_set_SSL_CTX.
1844 The session ID context is typically set from the parent SSL_CTX,
1845 and can vary with the CTX.
1848 *) Fix various certificate fingerprint issues.
1850 By using non-DER or invalid encodings outside the signed portion of a
1851 certificate the fingerprint can be changed without breaking the signature.
1852 Although no details of the signed portion of the certificate can be changed
1853 this can cause problems with some applications: e.g. those using the
1854 certificate fingerprint for blacklists.
1856 1. Reject signatures with non zero unused bits.
1858 If the BIT STRING containing the signature has non zero unused bits reject
1859 the signature. All current signature algorithms require zero unused bits.
1861 2. Check certificate algorithm consistency.
1863 Check the AlgorithmIdentifier inside TBS matches the one in the
1864 certificate signature. NB: this will result in signature failure
1865 errors for some broken certificates.
1867 Thanks to Konrad Kraszewski from Google for reporting this issue.
1869 3. Check DSA/ECDSA signatures use DER.
1871 Reencode DSA/ECDSA signatures and compare with the original received
1872 signature. Return an error if there is a mismatch.
1874 This will reject various cases including garbage after signature
1875 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1876 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1877 (negative or with leading zeroes).
1879 Further analysis was conducted and fixes were developed by Stephen Henson
1880 of the OpenSSL core team.
1885 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1886 results on some platforms, including x86_64. This bug occurs at random
1887 with a very low probability, and is not known to be exploitable in any
1888 way, though its exact impact is difficult to determine. Thanks to Pieter
1889 Wuille (Blockstream) who reported this issue and also suggested an initial
1890 fix. Further analysis was conducted by the OpenSSL development team and
1891 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1892 the OpenSSL core team.
1896 *) Do not resume sessions on the server if the negotiated protocol
1897 version does not match the session's version. Resuming with a different
1898 version, while not strictly forbidden by the RFC, is of questionable
1899 sanity and breaks all known clients.
1900 [David Benjamin, Emilia Käsper]
1902 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1903 early CCS messages during renegotiation. (Note that because
1904 renegotiation is encrypted, this early CCS was not exploitable.)
1907 *) Tighten client-side session ticket handling during renegotiation:
1908 ensure that the client only accepts a session ticket if the server sends
1909 the extension anew in the ServerHello. Previously, a TLS client would
1910 reuse the old extension state and thus accept a session ticket if one was
1911 announced in the initial ServerHello.
1913 Similarly, ensure that the client requires a session ticket if one
1914 was advertised in the ServerHello. Previously, a TLS client would
1915 ignore a missing NewSessionTicket message.
1918 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1920 *) SRTP Memory Leak.
1922 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1923 sends a carefully crafted handshake message, to cause OpenSSL to fail
1924 to free up to 64k of memory causing a memory leak. This could be
1925 exploited in a Denial Of Service attack. This issue affects OpenSSL
1926 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1927 whether SRTP is used or configured. Implementations of OpenSSL that
1928 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1930 The fix was developed by the OpenSSL team.
1934 *) Session Ticket Memory Leak.
1936 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1937 integrity of that ticket is first verified. In the event of a session
1938 ticket integrity check failing, OpenSSL will fail to free memory
1939 causing a memory leak. By sending a large number of invalid session
1940 tickets an attacker could exploit this issue in a Denial Of Service
1945 *) Build option no-ssl3 is incomplete.
1947 When OpenSSL is configured with "no-ssl3" as a build option, servers
1948 could accept and complete a SSL 3.0 handshake, and clients could be
1949 configured to send them.
1951 [Akamai and the OpenSSL team]
1953 *) Add support for TLS_FALLBACK_SCSV.
1954 Client applications doing fallback retries should call
1955 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1957 [Adam Langley, Bodo Moeller]
1959 *) Add additional DigestInfo checks.
1961 Reencode DigestInto in DER and check against the original when
1962 verifying RSA signature: this will reject any improperly encoded
1963 DigestInfo structures.
1965 Note: this is a precautionary measure and no attacks are currently known.
1969 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1971 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1972 SRP code can be overrun an internal buffer. Add sanity check that
1973 g, A, B < N to SRP code.
1975 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1976 Group for discovering this issue.
1980 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1981 TLS 1.0 instead of higher protocol versions when the ClientHello message
1982 is badly fragmented. This allows a man-in-the-middle attacker to force a
1983 downgrade to TLS 1.0 even if both the server and the client support a
1984 higher protocol version, by modifying the client's TLS records.
1986 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1987 researching this issue.
1991 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1992 to a denial of service attack. A malicious server can crash the client
1993 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1994 ciphersuite and sending carefully crafted handshake messages.
1996 Thanks to Felix Gröbert (Google) for discovering and researching this
2001 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2002 to leak memory. This can be exploited through a Denial of Service attack.
2003 Thanks to Adam Langley for discovering and researching this issue.
2007 *) An attacker can force openssl to consume large amounts of memory whilst
2008 processing DTLS handshake messages. This can be exploited through a
2009 Denial of Service attack.
2010 Thanks to Adam Langley for discovering and researching this issue.
2014 *) An attacker can force an error condition which causes openssl to crash
2015 whilst processing DTLS packets due to memory being freed twice. This
2016 can be exploited through a Denial of Service attack.
2017 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2022 *) If a multithreaded client connects to a malicious server using a resumed
2023 session and the server sends an ec point format extension it could write
2024 up to 255 bytes to freed memory.
2026 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2031 *) A malicious server can crash an OpenSSL client with a null pointer
2032 dereference (read) by specifying an SRP ciphersuite even though it was not
2033 properly negotiated with the client. This can be exploited through a
2034 Denial of Service attack.
2036 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2037 discovering and researching this issue.
2041 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2042 X509_name_oneline, X509_name_print_ex et al. to leak some information
2043 from the stack. Applications may be affected if they echo pretty printing
2044 output to the attacker.
2046 Thanks to Ivan Fratric (Google) for discovering this issue.
2048 [Emilia Käsper, and Steve Henson]
2050 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2051 for corner cases. (Certain input points at infinity could lead to
2052 bogus results, with non-infinity inputs mapped to infinity too.)
2055 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2057 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2058 handshake can force the use of weak keying material in OpenSSL
2059 SSL/TLS clients and servers.
2061 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2062 researching this issue. (CVE-2014-0224)
2063 [KIKUCHI Masashi, Steve Henson]
2065 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2066 OpenSSL DTLS client the code can be made to recurse eventually crashing
2069 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2071 [Imre Rad, Steve Henson]
2073 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2074 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2075 client or server. This is potentially exploitable to run arbitrary
2076 code on a vulnerable client or server.
2078 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2079 [Jüri Aedla, Steve Henson]
2081 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2082 are subject to a denial of service attack.
2084 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2085 this issue. (CVE-2014-3470)
2086 [Felix Gröbert, Ivan Fratric, Steve Henson]
2088 *) Harmonize version and its documentation. -f flag is used to display
2090 [mancha <mancha1@zoho.com>]
2092 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2093 in i2d_ECPrivateKey.
2094 [mancha <mancha1@zoho.com>]
2096 *) Fix some double frees. These are not thought to be exploitable.
2097 [mancha <mancha1@zoho.com>]
2099 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2101 *) A missing bounds check in the handling of the TLS heartbeat extension
2102 can be used to reveal up to 64k of memory to a connected client or
2105 Thanks for Neel Mehta of Google Security for discovering this bug and to
2106 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2107 preparing the fix (CVE-2014-0160)
2108 [Adam Langley, Bodo Moeller]
2110 *) Fix for the attack described in the paper "Recovering OpenSSL
2111 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2112 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2113 http://eprint.iacr.org/2014/140
2115 Thanks to Yuval Yarom and Naomi Benger for discovering this
2116 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2117 [Yuval Yarom and Naomi Benger]
2119 *) TLS pad extension: draft-agl-tls-padding-03
2121 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2122 TLS client Hello record length value would otherwise be > 255 and
2123 less that 512 pad with a dummy extension containing zeroes so it
2124 is at least 512 bytes long.
2126 [Adam Langley, Steve Henson]
2128 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2130 *) Fix for TLS record tampering bug. A carefully crafted invalid
2131 handshake could crash OpenSSL with a NULL pointer exception.
2132 Thanks to Anton Johansson for reporting this issues.
2135 *) Keep original DTLS digest and encryption contexts in retransmission
2136 structures so we can use the previous session parameters if they need
2137 to be resent. (CVE-2013-6450)
2140 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2141 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2142 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2143 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2144 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2145 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2146 [Rob Stradling, Adam Langley]
2148 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2150 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2151 supporting platforms or when small records were transferred.
2152 [Andy Polyakov, Steve Henson]
2154 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2156 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2158 This addresses the flaw in CBC record processing discovered by
2159 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2160 at: http://www.isg.rhul.ac.uk/tls/
2162 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2163 Security Group at Royal Holloway, University of London
2164 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2165 Emilia Käsper for the initial patch.
2167 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2169 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2170 ciphersuites which can be exploited in a denial of service attack.
2171 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2172 and detecting this bug and to Wolfgang Ettlinger
2173 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2177 *) Return an error when checking OCSP signatures when key is NULL.
2178 This fixes a DoS attack. (CVE-2013-0166)
2181 *) Make openssl verify return errors.
2182 [Chris Palmer <palmer@google.com> and Ben Laurie]
2184 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2185 the right response is stapled. Also change SSL_get_certificate()
2186 so it returns the certificate actually sent.
2187 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2188 [Rob Stradling <rob.stradling@comodo.com>]
2190 *) Fix possible deadlock when decoding public keys.
2193 *) Don't use TLS 1.0 record version number in initial client hello
2197 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2199 *) Sanity check record length before skipping explicit IV in TLS
2200 1.2, 1.1 and DTLS to fix DoS attack.
2202 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2203 fuzzing as a service testing platform.
2207 *) Initialise tkeylen properly when encrypting CMS messages.
2208 Thanks to Solar Designer of Openwall for reporting this issue.
2211 *) In FIPS mode don't try to use composite ciphers as they are not
2215 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2217 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2218 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2219 mean any application compiled against OpenSSL 1.0.0 headers setting
2220 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2221 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2222 0x10000000L Any application which was previously compiled against
2223 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2224 will need to be recompiled as a result. Letting be results in
2225 inability to disable specifically TLS 1.1 and in client context,
2226 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2229 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2230 disable just protocol X, but all protocols above X *if* there are
2231 protocols *below* X still enabled. In more practical terms it means
2232 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2233 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2234 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2238 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2240 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2241 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2242 in CRYPTO_realloc_clean.
2244 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2245 issue and to Adam Langley <agl@chromium.org> for fixing it.
2247 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2249 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2252 *) Workarounds for some broken servers that "hang" if a client hello
2253 record length exceeds 255 bytes.
2255 1. Do not use record version number > TLS 1.0 in initial client
2256 hello: some (but not all) hanging servers will now work.
2257 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2258 the number of ciphers sent in the client hello. This should be
2259 set to an even number, such as 50, for example by passing:
2260 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2261 Most broken servers should now work.
2262 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2263 TLS 1.2 client support entirely.
2266 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2269 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2271 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2272 STRING form instead of a DigestInfo.
2275 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2276 and the RSA_sign/RSA_verify functions. This was made more apparent when
2277 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2278 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2279 the correct format in RSA_verify so both forms transparently work.
2282 *) Some servers which support TLS 1.0 can choke if we initially indicate
2283 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2284 encrypted premaster secret. As a workaround use the maximum permitted
2285 client version in client hello, this should keep such servers happy
2286 and still work with previous versions of OpenSSL.
2289 *) Add support for TLS/DTLS heartbeats.
2290 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2292 *) Add support for SCTP.
2293 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2295 *) Improved PRNG seeding for VOS.
2296 [Paul Green <Paul.Green@stratus.com>]
2298 *) Extensive assembler packs updates, most notably:
2300 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2301 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2302 - x86_64: bit-sliced AES implementation;
2303 - ARM: NEON support, contemporary platforms optimizations;
2304 - s390x: z196 support;
2305 - *: GHASH and GF(2^m) multiplication implementations;
2309 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2310 (removal of unnecessary code)
2311 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2313 *) Add TLS key material exporter from RFC 5705.
2316 *) Add DTLS-SRTP negotiation from RFC 5764.
2319 *) Add Next Protocol Negotiation,
2320 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2321 disabled with a no-npn flag to config or Configure. Code donated
2323 [Adam Langley <agl@google.com> and Ben Laurie]
2325 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2326 NIST-P256, NIST-P521, with constant-time single point multiplication on
2327 typical inputs. Compiler support for the nonstandard type __uint128_t is
2328 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2329 Code made available under Apache License version 2.0.
2331 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2332 line to include this in your build of OpenSSL, and run "make depend" (or
2333 "make update"). This enables the following EC_METHODs:
2335 EC_GFp_nistp224_method()
2336 EC_GFp_nistp256_method()
2337 EC_GFp_nistp521_method()
2339 EC_GROUP_new_by_curve_name() will automatically use these (while
2340 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2342 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2344 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2345 all platforms. Move ssize_t definition from e_os.h to the public
2346 header file e_os2.h as it now appears in public header file cms.h
2349 *) New -sigopt option to the ca, req and x509 utilities. Additional
2350 signature parameters can be passed using this option and in
2354 *) Add RSA PSS signing function. This will generate and set the
2355 appropriate AlgorithmIdentifiers for PSS based on those in the
2356 corresponding EVP_MD_CTX structure. No application support yet.
2359 *) Support for companion algorithm specific ASN1 signing routines.
2360 New function ASN1_item_sign_ctx() signs a pre-initialised
2361 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2362 the appropriate parameters.
2365 *) Add new algorithm specific ASN1 verification initialisation function
2366 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2367 handling will be the same no matter what EVP_PKEY_METHOD is used.
2368 Add a PSS handler to support verification of PSS signatures: checked
2369 against a number of sample certificates.
2372 *) Add signature printing for PSS. Add PSS OIDs.
2373 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2375 *) Add algorithm specific signature printing. An individual ASN1 method
2376 can now print out signatures instead of the standard hex dump.
2378 More complex signatures (e.g. PSS) can print out more meaningful
2379 information. Include DSA version that prints out the signature
2383 *) Password based recipient info support for CMS library: implementing
2387 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2388 neatly separates the code into cipher and PBE sections and is required
2389 for some algorithms that split PBES2 into separate pieces (such as
2390 password based CMS).
2393 *) Session-handling fixes:
2394 - Fix handling of connections that are resuming with a session ID,
2395 but also support Session Tickets.
2396 - Fix a bug that suppressed issuing of a new ticket if the client
2397 presented a ticket with an expired session.
2398 - Try to set the ticket lifetime hint to something reasonable.
2399 - Make tickets shorter by excluding irrelevant information.
2400 - On the client side, don't ignore renewed tickets.
2401 [Adam Langley, Bodo Moeller (Google)]
2403 *) Fix PSK session representation.
2406 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2408 This work was sponsored by Intel.
2411 *) Add GCM support to TLS library. Some custom code is needed to split
2412 the IV between the fixed (from PRF) and explicit (from TLS record)
2413 portions. This adds all GCM ciphersuites supported by RFC5288 and
2414 RFC5289. Generalise some AES* cipherstrings to include GCM and
2415 add a special AESGCM string for GCM only.
2418 *) Expand range of ctrls for AES GCM. Permit setting invocation
2419 field on decrypt and retrieval of invocation field only on encrypt.
2422 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2423 As required by RFC5289 these ciphersuites cannot be used if for
2424 versions of TLS earlier than 1.2.
2427 *) For FIPS capable OpenSSL interpret a NULL default public key method
2428 as unset and return the appropriate default but do *not* set the default.
2429 This means we can return the appropriate method in applications that
2430 switch between FIPS and non-FIPS modes.
2433 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2434 ENGINE is used then we cannot handle that in the FIPS module so we
2435 keep original code iff non-FIPS operations are allowed.
2438 *) Add -attime option to openssl utilities.
2439 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2441 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2444 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2445 FIPS EC methods unconditionally for now.
2448 *) New build option no-ec2m to disable characteristic 2 code.
2451 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2452 all cases can be covered as some introduce binary incompatibilities.
2455 *) Redirect RSA operations to FIPS module including keygen,
2456 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2459 *) Add similar low level API blocking to ciphers.
2462 *) Low level digest APIs are not approved in FIPS mode: any attempt
2463 to use these will cause a fatal error. Applications that *really* want
2464 to use them can use the private_* version instead.
2467 *) Redirect cipher operations to FIPS module for FIPS builds.
2470 *) Redirect digest operations to FIPS module for FIPS builds.
2473 *) Update build system to add "fips" flag which will link in fipscanister.o
2474 for static and shared library builds embedding a signature if needed.
2477 *) Output TLS supported curves in preference order instead of numerical
2478 order. This is currently hardcoded for the highest order curves first.
2479 This should be configurable so applications can judge speed vs strength.
2482 *) Add TLS v1.2 server support for client authentication.
2485 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2489 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2490 FIPS modules versions.
2493 *) Add TLS v1.2 client side support for client authentication. Keep cache
2494 of handshake records longer as we don't know the hash algorithm to use
2495 until after the certificate request message is received.
2498 *) Initial TLS v1.2 client support. Add a default signature algorithms
2499 extension including all the algorithms we support. Parse new signature
2500 format in client key exchange. Relax some ECC signing restrictions for
2501 TLS v1.2 as indicated in RFC5246.
2504 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2505 to new signature format when needed using client digest preference.
2506 All server ciphersuites should now work correctly in TLS v1.2. No client
2507 support yet and no support for client certificates.
2510 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2511 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2512 ciphersuites. At present only RSA key exchange ciphersuites work with
2513 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2514 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2515 and version checking.
2518 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2519 with this defined it will not be affected by any changes to ssl internal
2520 structures. Add several utility functions to allow openssl application
2521 to work with OPENSSL_NO_SSL_INTERN defined.
2525 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2527 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2530 *) Permit abbreviated handshakes when renegotiating using the function
2531 SSL_renegotiate_abbreviated().
2532 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2534 *) Add call to ENGINE_register_all_complete() to
2535 ENGINE_load_builtin_engines(), so some implementations get used
2536 automatically instead of needing explicit application support.
2539 *) Add support for TLS key exporter as described in RFC5705.
2540 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2542 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2543 a few changes are required:
2545 Add SSL_OP_NO_TLSv1_1 flag.
2546 Add TLSv1_1 methods.
2547 Update version checking logic to handle version 1.1.
2548 Add explicit IV handling (ported from DTLS code).
2549 Add command line options to s_client/s_server.
2552 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2554 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2555 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2556 content decryption and always return the same error. Note: this attack
2557 needs on average 2^20 messages so it only affects automated senders. The
2558 old behaviour can be reenabled in the CMS code by setting the
2559 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2560 an MMA defence is not necessary.
2561 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2562 this issue. (CVE-2012-0884)
2565 *) Fix CVE-2011-4619: make sure we really are receiving a
2566 client hello before rejecting multiple SGC restarts. Thanks to
2567 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2570 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2572 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2573 Thanks to Antonio Martin, Enterprise Secure Access Research and
2574 Development, Cisco Systems, Inc. for discovering this bug and
2575 preparing a fix. (CVE-2012-0050)
2578 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2580 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2581 of the Vaudenay padding oracle attack on CBC mode encryption
2582 which enables an efficient plaintext recovery attack against
2583 the OpenSSL implementation of DTLS. Their attack exploits timing
2584 differences arising during decryption processing. A research
2585 paper describing this attack can be found at:
2586 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2587 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2588 Security Group at Royal Holloway, University of London
2589 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2590 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2591 for preparing the fix. (CVE-2011-4108)
2592 [Robin Seggelmann, Michael Tuexen]
2594 *) Clear bytes used for block padding of SSL 3.0 records.
2596 [Adam Langley (Google)]
2598 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2599 Kadianakis <desnacked@gmail.com> for discovering this issue and
2600 Adam Langley for preparing the fix. (CVE-2011-4619)
2601 [Adam Langley (Google)]
2603 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2604 [Andrey Kulikov <amdeich@gmail.com>]
2606 *) Prevent malformed RFC3779 data triggering an assertion failure.
2607 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2608 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2609 [Rob Austein <sra@hactrn.net>]
2611 *) Improved PRNG seeding for VOS.
2612 [Paul Green <Paul.Green@stratus.com>]
2614 *) Fix ssl_ciph.c set-up race.
2615 [Adam Langley (Google)]
2617 *) Fix spurious failures in ecdsatest.c.
2618 [Emilia Käsper (Google)]
2620 *) Fix the BIO_f_buffer() implementation (which was mixing different
2621 interpretations of the '..._len' fields).
2622 [Adam Langley (Google)]
2624 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2625 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2626 threads won't reuse the same blinding coefficients.
2628 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2629 lock to call BN_BLINDING_invert_ex, and avoids one use of
2630 BN_BLINDING_update for each BN_BLINDING structure (previously,
2631 the last update always remained unused).
2632 [Emilia Käsper (Google)]
2634 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2635 [Bob Buckholz (Google)]
2637 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2639 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2640 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2641 [Kaspar Brand <ossl@velox.ch>]
2643 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2644 for multi-threaded use of ECDH. (CVE-2011-3210)
2645 [Adam Langley (Google)]
2647 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2650 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2651 signature public key algorithm by using OID xref utilities instead.
2652 Before this you could only use some ECC ciphersuites with SHA1 only.
2655 *) Add protection against ECDSA timing attacks as mentioned in the paper
2656 by Billy Bob Brumley and Nicola Tuveri, see:
2658 http://eprint.iacr.org/2011/232.pdf
2660 [Billy Bob Brumley and Nicola Tuveri]
2662 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2664 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2665 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2667 *) Fix bug in string printing code: if *any* escaping is enabled we must
2668 escape the escape character (backslash) or the resulting string is
2672 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2674 *) Disable code workaround for ancient and obsolete Netscape browsers
2675 and servers: an attacker can use it in a ciphersuite downgrade attack.
2676 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2679 *) Fixed J-PAKE implementation error, originally discovered by
2680 Sebastien Martini, further info and confirmation from Stefan
2681 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2684 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2686 *) Fix extension code to avoid race conditions which can result in a buffer
2687 overrun vulnerability: resumed sessions must not be modified as they can
2688 be shared by multiple threads. CVE-2010-3864
2691 *) Fix WIN32 build system to correctly link an ENGINE directory into
2695 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2697 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2699 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2701 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2703 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2704 context. The operation can be customised via the ctrl mechanism in
2705 case ENGINEs want to include additional functionality.
2708 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2711 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2712 output hashes compatible with older versions of OpenSSL.
2713 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2715 *) Fix compression algorithm handling: if resuming a session use the
2716 compression algorithm of the resumed session instead of determining
2717 it from client hello again. Don't allow server to change algorithm.
2720 *) Add load_crls() function to apps tidying load_certs() too. Add option
2721 to verify utility to allow additional CRLs to be included.
2724 *) Update OCSP request code to permit adding custom headers to the request:
2725 some responders need this.
2728 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2730 [Julia Lawall <julia@diku.dk>]
2732 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2733 needlessly dereferenced structures, used obsolete functions and
2734 didn't handle all updated verify codes correctly.
2737 *) Disable MD2 in the default configuration.
2740 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2741 indicate the initial BIO being pushed or popped. This makes it possible
2742 to determine whether the BIO is the one explicitly called or as a result
2743 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2744 it handles reference counts correctly and doesn't zero out the I/O bio
2745 when it is not being explicitly popped. WARNING: applications which
2746 included workarounds for the old buggy behaviour will need to be modified
2747 or they could free up already freed BIOs.
2750 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2751 renaming to all platforms (within the 0.9.8 branch, this was
2752 done conditionally on Netware platforms to avoid a name clash).
2753 [Guenter <lists@gknw.net>]
2755 *) Add ECDHE and PSK support to DTLS.
2756 [Michael Tuexen <tuexen@fh-muenster.de>]
2758 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2762 *) Add "missing" function EVP_MD_flags() (without this the only way to
2763 retrieve a digest flags is by accessing the structure directly. Update
2764 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2765 or cipher is registered as in the "from" argument. Print out all
2766 registered digests in the dgst usage message instead of manually
2767 attempting to work them out.
2770 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2771 this allows the use of compression and extensions. Change default cipher
2772 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2773 by default unless an application cipher string requests it.
2776 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2777 key ids to find matching certificates and keys but some PKCS#12 files
2778 don't follow the (somewhat unwritten) rules and this strategy fails.
2779 Now just gather all certificates together and the first private key
2780 then look for the first certificate that matches the key.
2783 *) Support use of registered digest and cipher names for dgst and cipher
2784 commands instead of having to add each one as a special case. So now
2791 openssl dgst -sha256 foo
2793 and this works for ENGINE based algorithms too.
2797 *) Update Gost ENGINE to support parameter files.
2798 [Victor B. Wagner <vitus@cryptocom.ru>]
2800 *) Support GeneralizedTime in ca utility.
2801 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2803 *) Enhance the hash format used for certificate directory links. The new
2804 form uses the canonical encoding (meaning equivalent names will work
2805 even if they aren't identical) and uses SHA1 instead of MD5. This form
2806 is incompatible with the older format and as a result c_rehash should
2807 be used to rebuild symbolic links.
2810 *) Make PKCS#8 the default write format for private keys, replacing the
2811 traditional format. This form is standardised, more secure and doesn't
2812 include an implicit MD5 dependency.
2815 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2816 committed to OpenSSL should pass this lot as a minimum.
2819 *) Add session ticket override functionality for use by EAP-FAST.
2820 [Jouni Malinen <j@w1.fi>]
2822 *) Modify HMAC functions to return a value. Since these can be implemented
2823 in an ENGINE errors can occur.
2826 *) Type-checked OBJ_bsearch_ex.
2829 *) Type-checked OBJ_bsearch. Also some constification necessitated
2830 by type-checking. Still to come: TXT_DB, bsearch(?),
2831 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2835 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2836 seconds to a tm structure directly, instead of going through OS
2837 specific date routines. This avoids any issues with OS routines such
2838 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2839 and X509_time_adj_ex() to cover the extended range. The existing
2840 X509_time_adj() is still usable and will no longer have any date issues.
2843 *) Delta CRL support. New use deltas option which will attempt to locate
2844 and search any appropriate delta CRLs available.
2846 This work was sponsored by Google.
2849 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2850 code and add additional score elements. Validate alternate CRL paths
2851 as part of the CRL checking and indicate a new error "CRL path validation
2852 error" in this case. Applications wanting additional details can use
2853 the verify callback and check the new "parent" field. If this is not
2854 NULL CRL path validation is taking place. Existing applications wont
2855 see this because it requires extended CRL support which is off by
2858 This work was sponsored by Google.
2861 *) Support for freshest CRL extension.
2863 This work was sponsored by Google.
2866 *) Initial indirect CRL support. Currently only supported in the CRLs
2867 passed directly and not via lookup. Process certificate issuer
2868 CRL entry extension and lookup CRL entries by bother issuer name
2869 and serial number. Check and process CRL issuer entry in IDP extension.
2871 This work was sponsored by Google.
2874 *) Add support for distinct certificate and CRL paths. The CRL issuer
2875 certificate is validated separately in this case. Only enabled if
2876 an extended CRL support flag is set: this flag will enable additional
2877 CRL functionality in future.
2879 This work was sponsored by Google.
2882 *) Add support for policy mappings extension.
2884 This work was sponsored by Google.
2887 *) Fixes to pathlength constraint, self issued certificate handling,
2888 policy processing to align with RFC3280 and PKITS tests.
2890 This work was sponsored by Google.
2893 *) Support for name constraints certificate extension. DN, email, DNS
2894 and URI types are currently supported.
2896 This work was sponsored by Google.
2899 *) To cater for systems that provide a pointer-based thread ID rather
2900 than numeric, deprecate the current numeric thread ID mechanism and
2901 replace it with a structure and associated callback type. This
2902 mechanism allows a numeric "hash" to be extracted from a thread ID in
2903 either case, and on platforms where pointers are larger than 'long',
2904 mixing is done to help ensure the numeric 'hash' is usable even if it
2905 can't be guaranteed unique. The default mechanism is to use "&errno"
2906 as a pointer-based thread ID to distinguish between threads.
2908 Applications that want to provide their own thread IDs should now use
2909 CRYPTO_THREADID_set_callback() to register a callback that will call
2910 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2912 Note that ERR_remove_state() is now deprecated, because it is tied
2913 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2914 to free the current thread's error state should be replaced by
2915 ERR_remove_thread_state(NULL).
2917 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2918 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2919 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2920 application was previously providing a numeric thread callback that
2921 was inappropriate for distinguishing threads, then uniqueness might
2922 have been obtained with &errno that happened immediately in the
2923 intermediate development versions of OpenSSL; this is no longer the
2924 case, the numeric thread callback will now override the automatic use
2926 [Geoff Thorpe, with help from Bodo Moeller]
2928 *) Initial support for different CRL issuing certificates. This covers a
2929 simple case where the self issued certificates in the chain exist and
2930 the real CRL issuer is higher in the existing chain.
2932 This work was sponsored by Google.
2935 *) Removed effectively defunct crypto/store from the build.
2938 *) Revamp of STACK to provide stronger type-checking. Still to come:
2939 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2940 ASN1_STRING, CONF_VALUE.
2943 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2944 RAM on SSL connections. This option can save about 34k per idle SSL.
2947 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2948 STACK, TXT_DB, bsearch, qsort.
2951 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2952 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2953 support for data, signedData, compressedData, digestedData and
2954 encryptedData, envelopedData types included. Scripts to check against
2955 RFC4134 examples draft and interop and consistency checks of many
2956 content types and variants.
2959 *) Add options to enc utility to support use of zlib compression BIO.
2962 *) Extend mk1mf to support importing of options and assembly language
2963 files from Configure script, currently only included in VC-WIN32.
2964 The assembly language rules can now optionally generate the source
2965 files from the associated perl scripts.
2968 *) Implement remaining functionality needed to support GOST ciphersuites.
2969 Interop testing has been performed using CryptoPro implementations.
2970 [Victor B. Wagner <vitus@cryptocom.ru>]
2972 *) s390x assembler pack.
2975 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2979 *) Implement Opaque PRF Input TLS extension as specified in
2980 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2981 official specification yet and no extension type assignment by
2982 IANA exists, this extension (for now) will have to be explicitly
2983 enabled when building OpenSSL by providing the extension number
2984 to use. For example, specify an option
2986 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2988 to the "config" or "Configure" script to enable the extension,
2989 assuming extension number 0x9527 (which is a completely arbitrary
2990 and unofficial assignment based on the MD5 hash of the Internet
2991 Draft). Note that by doing so, you potentially lose
2992 interoperability with other TLS implementations since these might
2993 be using the same extension number for other purposes.
2995 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2996 opaque PRF input value to use in the handshake. This will create
2997 an interal copy of the length-'len' string at 'src', and will
2998 return non-zero for success.
3000 To get more control and flexibility, provide a callback function
3003 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3004 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3008 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3011 Callback function 'cb' will be called in handshakes, and is
3012 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3013 Argument 'arg' is for application purposes (the value as given to
3014 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3015 be provided to the callback function). The callback function
3016 has to return non-zero to report success: usually 1 to use opaque
3017 PRF input just if possible, or 2 to enforce use of the opaque PRF
3018 input. In the latter case, the library will abort the handshake
3019 if opaque PRF input is not successfully negotiated.
3021 Arguments 'peerinput' and 'len' given to the callback function
3022 will always be NULL and 0 in the case of a client. A server will
3023 see the client's opaque PRF input through these variables if
3024 available (NULL and 0 otherwise). Note that if the server
3025 provides an opaque PRF input, the length must be the same as the
3026 length of the client's opaque PRF input.
3028 Note that the callback function will only be called when creating
3029 a new session (session resumption can resume whatever was
3030 previously negotiated), and will not be called in SSL 2.0
3031 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3032 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3033 for applications that need to enforce opaque PRF input.
3037 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3040 [Victor B. Wagner <vitus@cryptocom.ru>]
3042 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3043 RFC4507bis. The encrypted ticket format is an encrypted encoded
3044 SSL_SESSION structure, that way new session features are automatically
3047 If a client application caches session in an SSL_SESSION structure
3048 support is transparent because tickets are now stored in the encoded
3051 The SSL_CTX structure automatically generates keys for ticket
3052 protection in servers so again support should be possible
3053 with no application modification.
3055 If a client or server wishes to disable RFC4507 support then the option
3056 SSL_OP_NO_TICKET can be set.
3058 Add a TLS extension debugging callback to allow the contents of any client
3059 or server extensions to be examined.
3061 This work was sponsored by Google.
3064 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3065 OpenSSL should now compile cleanly on gcc 4.2
3066 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3068 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3069 support including streaming MAC support: this is required for GOST
3070 ciphersuite support.
3071 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3073 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3074 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3075 to output in BER and PEM format.
3078 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3079 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3080 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3081 ENGINE support for HMAC keys which are unextractable. New -mac and
3082 -macopt options to dgst utility.
3085 *) New option -sigopt to dgst utility. Update dgst to use
3086 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3087 alternative signing parameters such as X9.31 or PSS in the dgst
3091 *) Change ssl_cipher_apply_rule(), the internal function that does
3092 the work each time a ciphersuite string requests enabling
3093 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3094 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3095 the order of disabled ciphersuites such that those ciphersuites
3096 that most recently went from enabled to disabled not only stay
3097 in order with respect to each other, but also have higher priority
3098 than other disabled ciphersuites the next time ciphersuites are
3101 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3102 the same ciphersuites as with "HIGH" alone, but in a specific
3103 order where the PSK ciphersuites come first (since they are the
3104 most recently disabled ciphersuites when "HIGH" is parsed).
3106 Also, change ssl_create_cipher_list() (using this new
3107 funcionality) such that between otherwise identical
3108 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3112 *) Change ssl_create_cipher_list() so that it automatically
3113 arranges the ciphersuites in reasonable order before starting
3114 to process the rule string. Thus, the definition for "DEFAULT"
3115 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3116 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3117 This makes it much easier to arrive at a reasonable default order
3118 in applications for which anonymous ciphers are OK (meaning
3119 that you can't actually use DEFAULT).
3120 [Bodo Moeller; suggested by Victor Duchovni]
3122 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3123 processing) into multiple integers instead of setting
3124 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3125 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3126 (These masks as well as the individual bit definitions are hidden
3127 away into the non-exported interface ssl/ssl_locl.h, so this
3128 change to the definition of the SSL_CIPHER structure shouldn't
3129 affect applications.) This give us more bits for each of these
3130 categories, so there is no longer a need to coagulate AES128 and
3131 AES256 into a single algorithm bit, and to coagulate Camellia128
3132 and Camellia256 into a single algorithm bit, which has led to all
3135 Thus, among other things, the kludge introduced in 0.9.7m and
3136 0.9.8e for masking out AES256 independently of AES128 or masking
3137 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3139 With the change, we also introduce new ciphersuite aliases that
3140 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3144 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3145 Use the leftmost N bytes of the signature input if the input is
3146 larger than the prime q (with N being the size in bytes of q).
3149 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3150 it yet and it is largely untested.
3153 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3156 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3157 some compilers (gcc 4.2 and later) reject their use. Safestack is
3158 reimplemented. Update ASN1 to avoid use of legacy functions.
3161 *) Win32/64 targets are linked with Winsock2.
3164 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3165 to external functions. This can be used to increase CRL handling
3166 efficiency especially when CRLs are very large by (for example) storing
3167 the CRL revoked certificates in a database.
3170 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3171 new CRLs added to a directory can be used. New command line option
3172 -verify_return_error to s_client and s_server. This causes real errors
3173 to be returned by the verify callback instead of carrying on no matter
3174 what. This reflects the way a "real world" verify callback would behave.
3177 *) GOST engine, supporting several GOST algorithms and public key formats.
3178 Kindly donated by Cryptocom.
3181 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3182 partitioned by DP are handled but no indirect CRL or reason partitioning
3183 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3184 selected via a scoring technique which handles IDP and AKID in CRLs.
3187 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3188 will ultimately be used for all verify operations: this will remove the
3189 X509_STORE dependency on certificate verification and allow alternative
3190 lookup methods. X509_STORE based implementations of these two callbacks.
3193 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3194 Modify get_crl() to find a valid (unexpired) CRL if possible.
3197 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3198 this would be called X509_CRL_cmp() but that name is already used by
3199 a function that just compares CRL issuer names. Cache several CRL
3200 extensions in X509_CRL structure and cache CRLDP in X509.
3203 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3204 this maps equivalent X509_NAME structures into a consistent structure.
3205 Name comparison can then be performed rapidly using memcmp().
3208 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3212 *) Allow digests to supply their own micalg string for S/MIME type using
3213 the ctrl EVP_MD_CTRL_MICALG.
3216 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3217 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3218 ctrl. It can then customise the structure before and/or after signing
3222 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3223 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3224 to free up any added signature OIDs.
3227 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3228 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3229 digest and cipher tables. New options added to openssl utility:
3230 list-message-digest-algorithms and list-cipher-algorithms.
3233 *) Change the array representation of binary polynomials: the list
3234 of degrees of non-zero coefficients is now terminated with -1.
3235 Previously it was terminated with 0, which was also part of the
3236 value; thus, the array representation was not applicable to
3237 polynomials where t^0 has coefficient zero. This change makes
3238 the array representation useful in a more general context.
3241 *) Various modifications and fixes to SSL/TLS cipher string
3242 handling. For ECC, the code now distinguishes between fixed ECDH
3243 with RSA certificates on the one hand and with ECDSA certificates
3244 on the other hand, since these are separate ciphersuites. The
3245 unused code for Fortezza ciphersuites has been removed.
3247 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3248 (not "ECDHE"). For consistency with the code for DH
3249 certificates, use of ECDH certificates is now considered ECDH
3250 authentication, not RSA or ECDSA authentication (the latter is
3251 merely the CA's signing algorithm and not actively used in the
3254 The temporary ciphersuite alias "ECCdraft" is no longer
3255 available, and ECC ciphersuites are no longer excluded from "ALL"
3256 and "DEFAULT". The following aliases now exist for RFC 4492
3257 ciphersuites, most of these by analogy with the DH case:
3259 kECDHr - ECDH cert, signed with RSA
3260 kECDHe - ECDH cert, signed with ECDSA
3261 kECDH - ECDH cert (signed with either RSA or ECDSA)
3262 kEECDH - ephemeral ECDH
3263 ECDH - ECDH cert or ephemeral ECDH
3269 AECDH - anonymous ECDH
3270 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3274 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3275 Use correct micalg parameters depending on digest(s) in signed message.
3278 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3279 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3282 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3283 an engine to register a method. Add ENGINE lookups for methods and
3284 functional reference processing.
3287 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3288 EVP_{Sign,Verify}* which allow an application to customise the signature
3292 *) New -resign option to smime utility. This adds one or more signers
3293 to an existing PKCS#7 signedData structure. Also -md option to use an
3294 alternative message digest algorithm for signing.
3297 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3298 create PKCS7 structures containing multiple signers. Update smime
3299 application to support multiple signers.
3302 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3306 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3307 Reorganize PBE internals to lookup from a static table using NIDs,
3308 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3309 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3310 PRF which will be automatically used with PBES2.
3313 *) Replace the algorithm specific calls to generate keys in "req" with the
3317 *) Update PKCS#7 enveloped data routines to use new API. This is now
3318 supported by any public key method supporting the encrypt operation. A
3319 ctrl is added to allow the public key algorithm to examine or modify
3320 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3324 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3325 a default digest type to use. In most cases this will be SHA1 but some
3326 algorithms (such as GOST) need to specify an alternative digest. The
3327 return value indicates how strong the preference is 1 means optional and
3328 2 is mandatory (that is it is the only supported type). Modify
3329 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3330 use the default md. Update openssl utilities to use the default digest
3331 type for signing if it is not explicitly indicated.
3334 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3335 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3336 signing method from the key type. This effectively removes the link
3337 between digests and public key types.
3340 *) Add an OID cross reference table and utility functions. Its purpose is to
3341 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3342 rsaEncryption. This will allow some of the algorithm specific hackery
3343 needed to use the correct OID to be removed.
3346 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3347 structures for PKCS7_sign(). They are now set up by the relevant public
3351 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3354 *) Add support for key derivation (agreement) in the API, DH method and
3358 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3359 public and private key formats. As a side effect these add additional
3360 command line functionality not previously available: DSA signatures can be
3361 generated and verified using pkeyutl and DH key support and generation in
3366 [Oliver Tappe <zooey@hirschkaefer.de>]
3368 *) New make target "install_html_docs" installs HTML renditions of the
3370 [Oliver Tappe <zooey@hirschkaefer.de>]
3372 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3373 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3374 support key and parameter generation and add initial key generation
3375 functionality for RSA.
3378 *) Add functions for main EVP_PKEY_method operations. The undocumented
3379 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3380 EVP_PKEY_{encrypt,decrypt}_old.
3383 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3384 key API, doesn't do much yet.
3387 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3388 public key algorithms. New option to openssl utility:
3389 "list-public-key-algorithms" to print out info.
3392 *) Implement the Supported Elliptic Curves Extension for
3393 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3396 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3397 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3400 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3401 utilities such as rsa, dsa, dsaparam etc except they process any key
3405 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3406 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3407 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3411 *) Initial support for pluggable public key ASN1.
3412 De-spaghettify the public key ASN1 handling. Move public and private
3413 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3414 algorithm specific handling to a single module within the relevant
3415 algorithm directory. Add functions to allow (near) opaque processing
3416 of public and private key structures.
3419 *) Implement the Supported Point Formats Extension for
3420 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3423 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3424 for the psk identity [hint] and the psk callback functions to the
3425 SSL_SESSION, SSL and SSL_CTX structure.
3428 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3432 SSL_CTX_use_psk_identity_hint
3433 SSL_get_psk_identity_hint
3434 SSL_get_psk_identity
3435 SSL_use_psk_identity_hint
3437 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3439 *) Add RFC 3161 compliant time stamp request creation, response generation
3440 and response verification functionality.
3441 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3443 *) Add initial support for TLS extensions, specifically for the server_name
3444 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3445 have new members for a host name. The SSL data structure has an
3446 additional member SSL_CTX *initial_ctx so that new sessions can be
3447 stored in that context to allow for session resumption, even after the
3448 SSL has been switched to a new SSL_CTX in reaction to a client's
3449 server_name extension.
3451 New functions (subject to change):
3453 SSL_get_servername()
3454 SSL_get_servername_type()
3457 New CTRL codes and macros (subject to change):
3459 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3460 - SSL_CTX_set_tlsext_servername_callback()
3461 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3462 - SSL_CTX_set_tlsext_servername_arg()
3463 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3465 openssl s_client has a new '-servername ...' option.
3467 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3468 '-key2 ...', '-servername_fatal' (subject to change). This allows
3469 testing the HostName extension for a specific single host name ('-cert'
3470 and '-key' remain fallbacks for handshakes without HostName
3471 negotiation). If the unrecognized_name alert has to be sent, this by
3472 default is a warning; it becomes fatal with the '-servername_fatal'
3475 [Peter Sylvester, Remy Allais, Christophe Renou]
3477 *) Whirlpool hash implementation is added.
3480 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3481 bn(64,32). Because of instruction set limitations it doesn't have
3482 any negative impact on performance. This was done mostly in order
3483 to make it possible to share assembler modules, such as bn_mul_mont
3484 implementations, between 32- and 64-bit builds without hassle.
3487 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3488 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3492 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3493 dedicated Montgomery multiplication procedure, is introduced.
3494 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3495 "64-bit" performance on certain 32-bit targets.
3498 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3499 in SSL structures. New SSL ctrl to set maximum send fragment size.
3500 Save memory by seeting the I/O buffer sizes dynamically instead of
3501 using the maximum available value.
3504 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3505 in addition to the text details.
3508 *) Very, very preliminary EXPERIMENTAL support for printing of general
3509 ASN1 structures. This currently produces rather ugly output and doesn't
3510 handle several customised structures at all.
3513 *) Integrated support for PVK file format and some related formats such
3514 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3515 these in the 'rsa' and 'dsa' utilities.
3518 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3521 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3522 place for the (very old) "NETSCAPE" format certificates which are now
3523 handled using new ASN1 code equivalents.
3526 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3527 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3528 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3531 *) Modify CRL distribution points extension code to print out previously
3532 unsupported fields. Enhance extension setting code to allow setting of
3536 *) Add print and set support for Issuing Distribution Point CRL extension.
3539 *) Change 'Configure' script to enable Camellia by default.
3542 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3544 *) When rejecting SSL/TLS records due to an incorrect version number, never
3545 update s->server with a new major version number. As of
3546 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3547 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3548 the previous behavior could result in a read attempt at NULL when
3549 receiving specific incorrect SSL/TLS records once record payload
3550 protection is active. (CVE-2010-0740)
3551 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3553 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3554 could be crashed if the relevant tables were not present (e.g. chrooted).
3555 [Tomas Hoger <thoger@redhat.com>]
3557 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3559 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3560 [Martin Olsson, Neel Mehta]
3562 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3563 accommodate for stack sorting, always a write lock!).
3566 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3567 excessive delays in the RAND_poll(): over a minute. As a workaround
3568 include a time check in the inner Heap32Next loop too.
3571 *) The code that handled flushing of data in SSL/TLS originally used the
3572 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3573 the problem outlined in PR#1949. The fix suggested there however can
3574 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3575 of Apache). So instead simplify the code to flush unconditionally.
3576 This should be fine since flushing with no data to flush is a no op.
3579 *) Handle TLS versions 2.0 and later properly and correctly use the
3580 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3581 off ancient servers have a habit of sticking around for a while...
3584 *) Modify compression code so it frees up structures without using the
3585 ex_data callbacks. This works around a problem where some applications
3586 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3587 restarting) then use compression (e.g. SSL with compression) later.
3588 This results in significant per-connection memory leaks and
3589 has caused some security issues including CVE-2008-1678 and
3593 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3594 change when encrypting or decrypting.
3597 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3598 connect and renegotiate with servers which do not support RI.
3599 Until RI is more widely deployed this option is enabled by default.
3602 *) Add "missing" ssl ctrls to clear options and mode.
3605 *) If client attempts to renegotiate and doesn't support RI respond with
3606 a no_renegotiation alert as required by RFC5746. Some renegotiating
3607 TLS clients will continue a connection gracefully when they receive
3608 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3609 waiting for a server hello which it will never receive. Now we treat a
3610 received no_renegotiation alert as a fatal error. This is because
3611 applications requesting a renegotiation might well expect it to succeed
3612 and would have no code in place to handle the server denying it so the
3613 only safe thing to do is to terminate the connection.
3616 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3617 peer supports secure renegotiation and 0 otherwise. Print out peer
3618 renegotiation support in s_client/s_server.
3621 *) Replace the highly broken and deprecated SPKAC certification method with
3622 the updated NID creation version. This should correctly handle UTF8.
3625 *) Implement RFC5746. Re-enable renegotiation but require the extension
3626 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3627 turns out to be a bad idea. It has been replaced by
3628 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3629 SSL_CTX_set_options(). This is really not recommended unless you
3630 know what you are doing.
3631 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3633 *) Fixes to stateless session resumption handling. Use initial_ctx when
3634 issuing and attempting to decrypt tickets in case it has changed during
3635 servername handling. Use a non-zero length session ID when attempting
3636 stateless session resumption: this makes it possible to determine if
3637 a resumption has occurred immediately after receiving server hello
3638 (several places in OpenSSL subtly assume this) instead of later in
3642 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3643 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3644 fixes for a few places where the return code is not checked
3646 [Julia Lawall <julia@diku.dk>]
3648 *) Add --strict-warnings option to Configure script to include devteam
3649 warnings in other configurations.
3652 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3653 makes it possible to install openssl libraries in locations which
3654 have names other than "lib", for example "/usr/lib64" which some
3656 [Steve Henson, based on patch from Jeremy Utley]
3658 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3659 X690 8.9.12 and can produce some misleading textual output of OIDs.
3660 [Steve Henson, reported by Dan Kaminsky]
3662 *) Delete MD2 from algorithm tables. This follows the recommendation in
3663 several standards that it is not used in new applications due to
3664 several cryptographic weaknesses. For binary compatibility reasons
3665 the MD2 API is still compiled in by default.
3668 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3672 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3673 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3675 [Guenter <lists@gknw.net>]
3677 *) Fix the server certificate chain building code to use X509_verify_cert(),
3678 it used to have an ad-hoc builder which was unable to cope with anything
3679 other than a simple chain.
3680 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3682 *) Don't check self signed certificate signatures in X509_verify_cert()
3683 by default (a flag can override this): it just wastes time without
3684 adding any security. As a useful side effect self signed root CAs
3685 with non-FIPS digests are now usable in FIPS mode.
3688 *) In dtls1_process_out_of_seq_message() the check if the current message
3689 is already buffered was missing. For every new message was memory
3690 allocated, allowing an attacker to perform an denial of service attack
3691 with sending out of seq handshake messages until there is no memory
3692 left. Additionally every future messege was buffered, even if the
3693 sequence number made no sense and would be part of another handshake.
3694 So only messages with sequence numbers less than 10 in advance will be
3695 buffered. (CVE-2009-1378)
3696 [Robin Seggelmann, discovered by Daniel Mentz]
3698 *) Records are buffered if they arrive with a future epoch to be
3699 processed after finishing the corresponding handshake. There is
3700 currently no limitation to this buffer allowing an attacker to perform
3701 a DOS attack with sending records with future epochs until there is no
3702 memory left. This patch adds the pqueue_size() function to determine
3703 the size of a buffer and limits the record buffer to 100 entries.
3705 [Robin Seggelmann, discovered by Daniel Mentz]
3707 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3708 parent structure is freed. (CVE-2009-1379)
3711 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3712 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3715 [Ilya O. <vrghost@gmail.com>]
3717 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3719 *) Disable renegotiation completely - this fixes a severe security
3720 problem (CVE-2009-3555) at the cost of breaking all
3721 renegotiation. Renegotiation can be re-enabled by setting
3722 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3723 run-time. This is really not recommended unless you know what
3727 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3729 *) Don't set val to NULL when freeing up structures, it is freed up by
3730 underlying code. If sizeof(void *) > sizeof(long) this can result in
3731 zeroing past the valid field. (CVE-2009-0789)
3732 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3734 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3735 checked correctly. This would allow some invalid signed attributes to
3736 appear to verify correctly. (CVE-2009-0591)
3737 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3739 *) Reject UniversalString and BMPString types with invalid lengths. This
3740 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3741 a legal length. (CVE-2009-0590)
3744 *) Set S/MIME signing as the default purpose rather than setting it
3745 unconditionally. This allows applications to override it at the store
3749 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3750 to handle some structures.
3753 *) Improve efficiency of mem_gets: don't search whole buffer each time
3755 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3757 *) New -hex option for openssl rand.
3760 *) Print out UTF8String and NumericString when parsing ASN1.
3763 *) Support NumericString type for name components.
3766 *) Allow CC in the environment to override the automatically chosen
3767 compiler. Note that nothing is done to ensure flags work with the
3771 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3773 *) Properly check EVP_VerifyFinal() and similar return values
3775 [Ben Laurie, Bodo Moeller, Google Security Team]
3777 *) Enable TLS extensions by default.
3780 *) Allow the CHIL engine to be loaded, whether the application is
3781 multithreaded or not. (This does not release the developer from the
3782 obligation to set up the dynamic locking callbacks.)
3783 [Sander Temme <sander@temme.net>]
3785 *) Use correct exit code if there is an error in dgst command.
3786 [Steve Henson; problem pointed out by Roland Dirlewanger]
3788 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3789 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3792 *) Add experimental JPAKE support, including demo authentication in
3793 s_client and s_server.
3796 *) Set the comparison function in v3_addr_canonize().
3797 [Rob Austein <sra@hactrn.net>]
3799 *) Add support for XMPP STARTTLS in s_client.
3800 [Philip Paeps <philip@freebsd.org>]
3802 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3803 to ensure that even with this option, only ciphersuites in the
3804 server's preference list will be accepted. (Note that the option
3805 applies only when resuming a session, so the earlier behavior was
3806 just about the algorithm choice for symmetric cryptography.)
3809 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3811 *) Fix NULL pointer dereference if a DTLS server received
3812 ChangeCipherSpec as first record (CVE-2009-1386).
3815 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3816 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3819 *) The fix in 0.9.8c that supposedly got rid of unsafe
3820 double-checked locking was incomplete for RSA blinding,
3821 addressing just one layer of what turns out to have been
3822 doubly unsafe triple-checked locking.
3824 So now fix this for real by retiring the MONT_HELPER macro
3825 in crypto/rsa/rsa_eay.c.
3827 [Bodo Moeller; problem pointed out by Marius Schilder]
3829 *) Various precautionary measures:
3831 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3833 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3834 (NB: This would require knowledge of the secret session ticket key
3835 to exploit, in which case you'd be SOL either way.)
3837 - Change bn_nist.c so that it will properly handle input BIGNUMs
3838 outside the expected range.
3840 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3843 [Neel Mehta, Bodo Moeller]
3845 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3846 the load fails. Useful for distros.
3847 [Ben Laurie and the FreeBSD team]
3849 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3852 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3855 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3857 This work was sponsored by Logica.
3860 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3861 keystores. Support for SSL/TLS client authentication too.
3862 Not compiled unless enable-capieng specified to Configure.
3864 This work was sponsored by Logica.
3867 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3868 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3869 attribute creation routines such as certifcate requests and PKCS#12
3873 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3875 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3876 handshake which could lead to a cilent crash as found using the
3877 Codenomicon TLS test suite (CVE-2008-1672)
3878 [Steve Henson, Mark Cox]
3880 *) Fix double free in TLS server name extensions which could lead to
3881 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3884 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3886 Clear the error queue to ensure that error entries left from
3887 older function calls do not interfere with the correct operation.
3888 [Lutz Jaenicke, Erik de Castro Lopo]
3890 *) Remove root CA certificates of commercial CAs:
3892 The OpenSSL project does not recommend any specific CA and does not
3893 have any policy with respect to including or excluding any CA.
3894 Therefore it does not make any sense to ship an arbitrary selection
3895 of root CA certificates with the OpenSSL software.
3898 *) RSA OAEP patches to fix two separate invalid memory reads.
3899 The first one involves inputs when 'lzero' is greater than
3900 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3901 before the beginning of from). The second one involves inputs where
3902 the 'db' section contains nothing but zeroes (there is a one-byte
3903 invalid read after the end of 'db').
3904 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3906 *) Partial backport from 0.9.9-dev:
3908 Introduce bn_mul_mont (dedicated Montgomery multiplication
3909 procedure) as a candidate for BIGNUM assembler implementation.
3910 While 0.9.9-dev uses assembler for various architectures, only
3911 x86_64 is available by default here in the 0.9.8 branch, and
3912 32-bit x86 is available through a compile-time setting.
3914 To try the 32-bit x86 assembler implementation, use Configure
3915 option "enable-montasm" (which exists only for this backport).
3917 As "enable-montasm" for 32-bit x86 disclaims code stability
3918 anyway, in this constellation we activate additional code
3919 backported from 0.9.9-dev for further performance improvements,
3920 namely BN_from_montgomery_word. (To enable this otherwise,
3921 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3923 [Andy Polyakov (backport partially by Bodo Moeller)]
3925 *) Add TLS session ticket callback. This allows an application to set
3926 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3927 values. This is useful for key rollover for example where several key
3928 sets may exist with different names.
3931 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3932 This was broken until now in 0.9.8 releases, such that the only way
3933 a registered ENGINE could be used (assuming it initialises
3934 successfully on the host) was to explicitly set it as the default
3935 for the relevant algorithms. This is in contradiction with 0.9.7
3936 behaviour and the documentation. With this fix, when an ENGINE is
3937 registered into a given algorithm's table of implementations, the
3938 'uptodate' flag is reset so that auto-discovery will be used next
3939 time a new context for that algorithm attempts to select an
3941 [Ian Lister (tweaked by Geoff Thorpe)]
3943 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3944 implemention in the following ways:
3946 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3949 Lack of BER streaming support means one pass streaming processing is
3950 only supported if data is detached: setting the streaming flag is
3951 ignored for embedded content.
3953 CMS support is disabled by default and must be explicitly enabled
3954 with the enable-cms configuration option.
3957 *) Update the GMP engine glue to do direct copies between BIGNUM and
3958 mpz_t when openssl and GMP use the same limb size. Otherwise the
3959 existing "conversion via a text string export" trick is still used.
3960 [Paul Sheer <paulsheer@gmail.com>]
3962 *) Zlib compression BIO. This is a filter BIO which compressed and
3963 uncompresses any data passed through it.
3966 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3967 RFC3394 compatible AES key wrapping.
3970 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3971 sets string data without copying. X509_ALGOR_set0() and
3972 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3973 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3974 from an X509_ATTRIBUTE structure optionally checking it occurs only
3975 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3979 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3980 to get the expected BN_FLG_CONSTTIME behavior.
3981 [Bodo Moeller (Google)]
3985 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3986 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3987 - added some more tests to do_tests.pl
3988 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3989 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3990 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3991 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3992 - various changes to netware.pl to enable gcc-cross builds on Win32
3994 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3995 - various changes to fix missing prototype warnings
3996 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3997 - added AES, WHIRLPOOL and CPUID assembler code to build files
3998 - added missing AES assembler make rules to mk1mf.pl
3999 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4000 [Guenter Knauf <eflash@gmx.net>]
4002 *) Implement certificate status request TLS extension defined in RFC3546.
4003 A client can set the appropriate parameters and receive the encoded
4004 OCSP response via a callback. A server can query the supplied parameters
4005 and set the encoded OCSP response in the callback. Add simplified examples
4006 to s_client and s_server.
4009 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4011 *) Fix various bugs:
4012 + Binary incompatibility of ssl_ctx_st structure
4013 + DTLS interoperation with non-compliant servers
4014 + Don't call get_session_cb() without proposed session
4015 + Fix ia64 assembler code
4016 [Andy Polyakov, Steve Henson]
4018 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4020 *) DTLS Handshake overhaul. There were longstanding issues with
4021 OpenSSL DTLS implementation, which were making it impossible for
4022 RFC 4347 compliant client to communicate with OpenSSL server.
4023 Unfortunately just fixing these incompatibilities would "cut off"
4024 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4025 server keeps tolerating non RFC compliant syntax. The opposite is
4026 not true, 0.9.8f client can not communicate with earlier server.
4027 This update even addresses CVE-2007-4995.
4030 *) Changes to avoid need for function casts in OpenSSL: some compilers
4031 (gcc 4.2 and later) reject their use.
4032 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4035 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4036 RFC4507bis. The encrypted ticket format is an encrypted encoded
4037 SSL_SESSION structure, that way new session features are automatically
4040 If a client application caches session in an SSL_SESSION structure
4041 support is transparent because tickets are now stored in the encoded
4044 The SSL_CTX structure automatically generates keys for ticket
4045 protection in servers so again support should be possible
4046 with no application modification.
4048 If a client or server wishes to disable RFC4507 support then the option
4049 SSL_OP_NO_TICKET can be set.
4051 Add a TLS extension debugging callback to allow the contents of any client
4052 or server extensions to be examined.
4054 This work was sponsored by Google.
4057 *) Add initial support for TLS extensions, specifically for the server_name
4058 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4059 have new members for a host name. The SSL data structure has an
4060 additional member SSL_CTX *initial_ctx so that new sessions can be
4061 stored in that context to allow for session resumption, even after the
4062 SSL has been switched to a new SSL_CTX in reaction to a client's
4063 server_name extension.
4065 New functions (subject to change):
4067 SSL_get_servername()
4068 SSL_get_servername_type()
4071 New CTRL codes and macros (subject to change):
4073 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4074 - SSL_CTX_set_tlsext_servername_callback()
4075 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4076 - SSL_CTX_set_tlsext_servername_arg()
4077 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4079 openssl s_client has a new '-servername ...' option.
4081 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4082 '-key2 ...', '-servername_fatal' (subject to change). This allows
4083 testing the HostName extension for a specific single host name ('-cert'
4084 and '-key' remain fallbacks for handshakes without HostName
4085 negotiation). If the unrecognized_name alert has to be sent, this by
4086 default is a warning; it becomes fatal with the '-servername_fatal'
4089 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4091 *) Add AES and SSE2 assembly language support to VC++ build.
4094 *) Mitigate attack on final subtraction in Montgomery reduction.
4097 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4098 (which previously caused an internal error).
4101 *) Squeeze another 10% out of IGE mode when in != out.
4104 *) AES IGE mode speedup.
4105 [Dean Gaudet (Google)]
4107 *) Add the Korean symmetric 128-bit cipher SEED (see
4108 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4109 add SEED ciphersuites from RFC 4162:
4111 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4112 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4113 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4114 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4116 To minimize changes between patchlevels in the OpenSSL 0.9.8
4117 series, SEED remains excluded from compilation unless OpenSSL
4118 is configured with 'enable-seed'.
4119 [KISA, Bodo Moeller]
4121 *) Mitigate branch prediction attacks, which can be practical if a
4122 single processor is shared, allowing a spy process to extract
4123 information. For detailed background information, see
4124 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4125 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4126 and Necessary Software Countermeasures"). The core of the change
4127 are new versions BN_div_no_branch() and
4128 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4129 respectively, which are slower, but avoid the security-relevant
4130 conditional branches. These are automatically called by BN_div()
4131 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4132 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4133 remove a conditional branch.
4135 BN_FLG_CONSTTIME is the new name for the previous
4136 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4137 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4138 in the exponent causes BN_mod_exp_mont() to use the alternative
4139 implementation in BN_mod_exp_mont_consttime().) The old name
4140 remains as a deprecated alias.
4142 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4143 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4144 constant-time implementations for more than just exponentiation.
4145 Here too the old name is kept as a deprecated alias.
4147 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4148 the BN_BLINDING structure gets an independent copy of the
4149 modulus. This means that the previous "BIGNUM *m" argument to
4150 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4151 essentially becomes "const BIGNUM *m", although we can't actually
4152 change this in the header file before 0.9.9. It allows
4153 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4154 enable BN_FLG_CONSTTIME.
4156 [Matthew D Wood (Intel Corp)]
4158 *) In the SSL/TLS server implementation, be strict about session ID
4159 context matching (which matters if an application uses a single
4160 external cache for different purposes). Previously,
4161 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4162 set. This did ensure strict client verification, but meant that,
4163 with applications using a single external cache for quite
4164 different requirements, clients could circumvent ciphersuite
4165 restrictions for a given session ID context by starting a session
4166 in a different context.
4169 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4170 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4171 authentication-only ciphersuites.
4174 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4175 not complete and could lead to a possible single byte overflow
4176 (CVE-2007-5135) [Ben Laurie]
4178 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4180 *) Since AES128 and AES256 (and similarly Camellia128 and
4181 Camellia256) share a single mask bit in the logic of
4182 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4183 kludge to work properly if AES128 is available and AES256 isn't
4184 (or if Camellia128 is available and Camellia256 isn't).
4187 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4188 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4189 When a point or a seed is encoded in a BIT STRING, we need to
4190 prevent the removal of trailing zero bits to get the proper DER
4191 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4192 of a NamedBitList, for which trailing 0 bits need to be removed.)
4195 *) Have SSL/TLS server implementation tolerate "mismatched" record
4196 protocol version while receiving ClientHello even if the
4197 ClientHello is fragmented. (The server can't insist on the
4198 particular protocol version it has chosen before the ServerHello
4199 message has informed the client about his choice.)
4202 *) Add RFC 3779 support.
4203 [Rob Austein for ARIN, Ben Laurie]
4205 *) Load error codes if they are not already present instead of using a
4206 static variable. This allows them to be cleanly unloaded and reloaded.
4207 Improve header file function name parsing.
4210 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4211 or CAPABILITY handshake as required by RFCs.
4214 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4216 *) Introduce limits to prevent malicious keys being able to
4217 cause a denial of service. (CVE-2006-2940)
4218 [Steve Henson, Bodo Moeller]
4220 *) Fix ASN.1 parsing of certain invalid structures that can result
4221 in a denial of service. (CVE-2006-2937) [Steve Henson]
4223 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4224 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4226 *) Fix SSL client code which could crash if connecting to a
4227 malicious SSLv2 server. (CVE-2006-4343)
4228 [Tavis Ormandy and Will Drewry, Google Security Team]
4230 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4231 match only those. Before that, "AES256-SHA" would be interpreted
4232 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4233 the same strength classification in 0.9.7h) as we currently only
4234 have a single AES bit in the ciphersuite description bitmap.
4235 That change, however, also applied to ciphersuite strings such as
4236 "RC4-MD5" that intentionally matched multiple ciphersuites --
4237 namely, SSL 2.0 ciphersuites in addition to the more common ones
4238 from SSL 3.0/TLS 1.0.
4240 So we change the selection algorithm again: Naming an explicit
4241 ciphersuite selects this one ciphersuite, and any other similar
4242 ciphersuite (same bitmap) from *other* protocol versions.
4243 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4244 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4246 Since SSL 2.0 does not have any ciphersuites for which the
4247 128/256 bit distinction would be relevant, this works for now.
4248 The proper fix will be to use different bits for AES128 and
4249 AES256, which would have avoided the problems from the beginning;
4250 however, bits are scarce, so we can only do this in a new release
4251 (not just a patchlevel) when we can change the SSL_CIPHER
4252 definition to split the single 'unsigned long mask' bitmap into
4253 multiple values to extend the available space.
4257 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4259 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4260 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4262 *) Add AES IGE and biIGE modes.
4265 *) Change the Unix randomness entropy gathering to use poll() when
4266 possible instead of select(), since the latter has some
4267 undesirable limitations.
4268 [Darryl Miles via Richard Levitte and Bodo Moeller]
4270 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4271 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4272 cannot be implicitly activated as part of, e.g., the "AES" alias.
4273 However, please upgrade to OpenSSL 0.9.9[-dev] for
4274 non-experimental use of the ECC ciphersuites to get TLS extension
4275 support, which is required for curve and point format negotiation
4276 to avoid potential handshake problems.
4279 *) Disable rogue ciphersuites:
4281 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4282 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4283 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4285 The latter two were purportedly from
4286 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4289 Also deactivate the remaining ciphersuites from
4290 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4291 unofficial, and the ID has long expired.
4294 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4295 dual-core machines) and other potential thread-safety issues.
4298 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4299 versions), which is now available for royalty-free use
4300 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4301 Also, add Camellia TLS ciphersuites from RFC 4132.
4303 To minimize changes between patchlevels in the OpenSSL 0.9.8
4304 series, Camellia remains excluded from compilation unless OpenSSL
4305 is configured with 'enable-camellia'.
4308 *) Disable the padding bug check when compression is in use. The padding
4309 bug check assumes the first packet is of even length, this is not
4310 necessarily true if compresssion is enabled and can result in false
4311 positives causing handshake failure. The actual bug test is ancient
4312 code so it is hoped that implementations will either have fixed it by
4313 now or any which still have the bug do not support compression.
4316 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4318 *) When applying a cipher rule check to see if string match is an explicit
4319 cipher suite and only match that one cipher suite if it is.
4322 *) Link in manifests for VC++ if needed.
4323 [Austin Ziegler <halostatue@gmail.com>]
4325 *) Update support for ECC-based TLS ciphersuites according to
4326 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4327 TLS extensions, which are supported starting with the 0.9.9
4328 branch, not in the OpenSSL 0.9.8 branch).
4331 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4332 opaque EVP_CIPHER_CTX handling.
4335 *) Fixes and enhancements to zlib compression code. We now only use
4336 "zlib1.dll" and use the default __cdecl calling convention on Win32
4337 to conform with the standards mentioned here:
4338 http://www.zlib.net/DLL_FAQ.txt
4339 Static zlib linking now works on Windows and the new --with-zlib-include
4340 --with-zlib-lib options to Configure can be used to supply the location
4341 of the headers and library. Gracefully handle case where zlib library
4345 *) Several fixes and enhancements to the OID generation code. The old code
4346 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4347 handle numbers larger than ULONG_MAX, truncated printing and had a
4348 non standard OBJ_obj2txt() behaviour.
4351 *) Add support for building of engines under engine/ as shared libraries
4352 under VC++ build system.
4355 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4356 Hopefully, we will not see any false combination of paths any more.
4359 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4361 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4362 (part of SSL_OP_ALL). This option used to disable the
4363 countermeasure against man-in-the-middle protocol-version
4364 rollback in the SSL 2.0 server implementation, which is a bad
4365 idea. (CVE-2005-2969)
4367 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4368 for Information Security, National Institute of Advanced Industrial
4369 Science and Technology [AIST], Japan)]
4371 *) Add two function to clear and return the verify parameter flags.
4374 *) Keep cipherlists sorted in the source instead of sorting them at
4375 runtime, thus removing the need for a lock.
4378 *) Avoid some small subgroup attacks in Diffie-Hellman.
4379 [Nick Mathewson and Ben Laurie]
4381 *) Add functions for well-known primes.
4384 *) Extended Windows CE support.
4385 [Satoshi Nakamura and Andy Polyakov]
4387 *) Initialize SSL_METHOD structures at compile time instead of during
4388 runtime, thus removing the need for a lock.
4391 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4392 attempting to decrypt each encrypted key in turn. Add support to
4396 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4398 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4401 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4404 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4405 key into the same file any more.
4408 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4411 *) Add -utf8 command line and config file option to 'ca'.
4412 [Stefan <stf@udoma.org]
4414 *) Removed the macro des_crypt(), as it seems to conflict with some
4415 libraries. Use DES_crypt().
4418 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4419 involves renaming the source and generated shared-libs for
4420 both. The engines will accept the corrected or legacy ids
4421 ('ncipher' and '4758_cca' respectively) when binding. NB,
4422 this only applies when building 'shared'.
4423 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4425 *) Add attribute functions to EVP_PKEY structure. Modify
4426 PKCS12_create() to recognize a CSP name attribute and
4427 use it. Make -CSP option work again in pkcs12 utility.
4430 *) Add new functionality to the bn blinding code:
4431 - automatic re-creation of the BN_BLINDING parameters after
4432 a fixed number of uses (currently 32)
4433 - add new function for parameter creation
4434 - introduce flags to control the update behaviour of the
4435 BN_BLINDING parameters
4436 - hide BN_BLINDING structure
4437 Add a second BN_BLINDING slot to the RSA structure to improve
4438 performance when a single RSA object is shared among several
4442 *) Add support for DTLS.
4443 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4445 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4446 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4449 *) Remove buggy and incomplete DH cert support from
4450 ssl/ssl_rsa.c and ssl/s3_both.c
4453 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4454 the apps/openssl applications.
4457 *) Compile clean with "-Wall -Wmissing-prototypes
4458 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4459 DEBUG_SAFESTACK must also be set.
4462 *) Change ./Configure so that certain algorithms can be disabled by default.
4463 The new counterpiece to "no-xxx" is "enable-xxx".
4465 The patented RC5 and MDC2 algorithms will now be disabled unless
4466 "enable-rc5" and "enable-mdc2", respectively, are specified.
4468 (IDEA remains enabled despite being patented. This is because IDEA
4469 is frequently required for interoperability, and there is no license
4470 fee for non-commercial use. As before, "no-idea" can be used to
4471 avoid this algorithm.)
4475 *) Add processing of proxy certificates (see RFC 3820). This work was
4476 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4477 EGEE (Enabling Grids for E-science in Europe).
4480 *) RC4 performance overhaul on modern architectures/implementations, such
4481 as Intel P4, IA-64 and AMD64.
4484 *) New utility extract-section.pl. This can be used specify an alternative
4485 section number in a pod file instead of having to treat each file as
4486 a separate case in Makefile. This can be done by adding two lines to the
4489 =for comment openssl_section:XXX
4491 The blank line is mandatory.
4495 *) New arguments -certform, -keyform and -pass for s_client and s_server
4496 to allow alternative format key and certificate files and passphrase
4500 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4501 update associated structures and add various utility functions.
4503 Add new policy related verify parameters, include policy checking in
4504 standard verify code. Enhance 'smime' application with extra parameters
4505 to support policy checking and print out.
4508 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4509 Nehemiah processors. These extensions support AES encryption in hardware
4510 as well as RNG (though RNG support is currently disabled).
4511 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4513 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4516 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4517 [Andy Polyakov and a number of other people]
4519 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4520 implementation contributed by IBM.
4521 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4523 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4524 exponent rather than 'unsigned long'. There is a corresponding change to
4525 the new 'rsa_keygen' element of the RSA_METHOD structure.
4526 [Jelte Jansen, Geoff Thorpe]
4528 *) Functionality for creating the initial serial number file is now
4529 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4531 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4532 number file to 1, which is bound to cause problems. To avoid
4533 the problems while respecting compatibility between different 0.9.7
4534 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4535 CA.pl for serial number initialization. With the new release 0.9.8,
4536 we can fix the problem directly in the 'ca' utility.)
4539 *) Reduced header interdepencies by declaring more opaque objects in
4540 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4541 give fewer recursive includes, which could break lazy source code - so
4542 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4543 developers should define this symbol when building and using openssl to
4544 ensure they track the recommended behaviour, interfaces, [etc], but
4545 backwards-compatible behaviour prevails when this isn't defined.
4548 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4551 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4552 This will generate a random key of the appropriate length based on the
4553 cipher context. The EVP_CIPHER can provide its own random key generation
4554 routine to support keys of a specific form. This is used in the des and
4555 3des routines to generate a key of the correct parity. Update S/MIME
4556 code to use new functions and hence generate correct parity DES keys.
4557 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4558 valid (weak or incorrect parity).
4561 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4562 as looking them up. This is useful when the verified structure may contain
4563 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4564 present unless the new PKCS7_NO_CRL flag is asserted.
4567 *) Extend ASN1 oid configuration module. It now additionally accepts the
4570 shortName = some long name, 1.2.3.4
4573 *) Reimplemented the BN_CTX implementation. There is now no more static
4574 limitation on the number of variables it can handle nor the depth of the
4575 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4576 information can now expand as required, and rather than having a single
4577 static array of bignums, BN_CTX now uses a linked-list of such arrays
4578 allowing it to expand on demand whilst maintaining the usefulness of
4579 BN_CTX's "bundling".
4582 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4583 to allow all RSA operations to function using a single BN_CTX.
4586 *) Preliminary support for certificate policy evaluation and checking. This
4587 is initially intended to pass the tests outlined in "Conformance Testing
4588 of Relying Party Client Certificate Path Processing Logic" v1.07.
4591 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4592 remained unused and not that useful. A variety of other little bignum
4593 tweaks and fixes have also been made continuing on from the audit (see
4597 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4598 associated ASN1, EVP and SSL functions and old ASN1 macros.
4601 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4602 and this should never fail. So the return value from the use of
4603 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4604 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4607 *) BN_CTX_get() should return zero-valued bignums, providing the same
4608 initialised value as BN_new().
4609 [Geoff Thorpe, suggested by Ulf Möller]
4611 *) Support for inhibitAnyPolicy certificate extension.
4614 *) An audit of the BIGNUM code is underway, for which debugging code is
4615 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4616 is considered valid when processing BIGNUMs, and causes execution to
4617 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4618 further steps are taken to deliberately pollute unused data in BIGNUM
4619 structures to try and expose faulty code further on. For now, openssl will
4620 (in its default mode of operation) continue to tolerate the inconsistent
4621 forms that it has tolerated in the past, but authors and packagers should
4622 consider trying openssl and their own applications when compiled with
4623 these debugging symbols defined. It will help highlight potential bugs in
4624 their own code, and will improve the test coverage for OpenSSL itself. At
4625 some point, these tighter rules will become openssl's default to improve
4626 maintainability, though the assert()s and other overheads will remain only
4627 in debugging configurations. See bn.h for more details.
4628 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4630 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4631 that can only be obtained through BN_CTX_new() (which implicitly
4632 initialises it). The presence of this function only made it possible
4633 to overwrite an existing structure (and cause memory leaks).
4636 *) Because of the callback-based approach for implementing LHASH as a
4637 template type, lh_insert() adds opaque objects to hash-tables and
4638 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4639 to clean up those corresponding objects before destroying the hash table
4640 (and losing the object pointers). So some over-zealous constifications in
4641 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4642 objects as "const" and the lh_doall[_arg] callback wrappers are not
4643 prototyped to have "const" restrictions on the object pointers they are
4644 given (and so aren't required to cast them away any more).
4647 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4648 (speed) prefers to use its own implementation. The two implementations
4649 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4650 its object type properly exposed (MS_TM) instead of casting to/from "char
4651 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4652 aren't necessarily the greatest nomenclatures - but this is what was used
4653 internally to the implementation so I've used that for now.
4656 *) Ensure that deprecated functions do not get compiled when
4657 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4658 the self-tests were still using deprecated key-generation functions so
4659 these have been updated also.
4662 *) Reorganise PKCS#7 code to separate the digest location functionality
4663 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4664 New function PKCS7_set_digest() to set the digest type for PKCS#7
4665 digestedData type. Add additional code to correctly generate the
4666 digestedData type and add support for this type in PKCS7 initialization
4670 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4671 structure of type "other".
4674 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4675 sure the loop does correctly stop and breaking ("division by zero")
4676 modulus operations are not performed. The (pre-generated) prime
4677 table crypto/bn/bn_prime.h was already correct, but it could not be
4678 re-generated on some platforms because of the "division by zero"
4679 situation in the script.
4680 [Ralf S. Engelschall]
4682 *) Update support for ECC-based TLS ciphersuites according to
4683 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4684 SHA-1 now is only used for "small" curves (where the
4685 representation of a field element takes up to 24 bytes); for
4686 larger curves, the field element resulting from ECDH is directly
4687 used as premaster secret.
4688 [Douglas Stebila (Sun Microsystems Laboratories)]
4690 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4691 curve secp160r1 to the tests.
4692 [Douglas Stebila (Sun Microsystems Laboratories)]
4694 *) Add the possibility to load symbols globally with DSO.
4695 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4697 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4698 control of the error stack.
4701 *) Add support for STORE in ENGINE.
4704 *) Add the STORE type. The intention is to provide a common interface
4705 to certificate and key stores, be they simple file-based stores, or
4706 HSM-type store, or LDAP stores, or...
4707 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4710 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4711 pass a list of arguments to any function as well as provide a way
4712 for a function to pass data back to the caller.
4715 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4716 works like BUF_strdup() but can be used to duplicate a portion of
4717 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4721 *) Add the function sk_find_ex() which works like sk_find(), but will
4722 return an index to an element even if an exact match couldn't be
4723 found. The index is guaranteed to point at the element where the
4724 searched-for key would be inserted to preserve sorting order.
4727 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4728 takes an extra flags argument for optional functionality. Currently,
4729 the following flags are defined:
4731 OBJ_BSEARCH_VALUE_ON_NOMATCH
4732 This one gets OBJ_bsearch_ex() to return a pointer to the first
4733 element where the comparing function returns a negative or zero
4736 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4737 This one gets OBJ_bsearch_ex() to return a pointer to the first
4738 element where the comparing function returns zero. This is useful
4739 if there are more than one element where the comparing function
4743 *) Make it possible to create self-signed certificates with 'openssl ca'
4744 in such a way that the self-signed certificate becomes part of the
4745 CA database and uses the same mechanisms for serial number generation
4746 as all other certificate signing. The new flag '-selfsign' enables
4747 this functionality. Adapt CA.sh and CA.pl.in.
4750 *) Add functionality to check the public key of a certificate request
4751 against a given private. This is useful to check that a certificate
4752 request can be signed by that key (self-signing).
4755 *) Make it possible to have multiple active certificates with the same
4756 subject in the CA index file. This is done only if the keyword
4757 'unique_subject' is set to 'no' in the main CA section (default
4758 if 'CA_default') of the configuration file. The value is saved
4759 with the database itself in a separate index attribute file,
4760 named like the index file with '.attr' appended to the name.
4763 *) Generate muti valued AVAs using '+' notation in config files for
4767 *) Support for nameConstraints certificate extension.
4770 *) Support for policyConstraints certificate extension.
4773 *) Support for policyMappings certificate extension.
4776 *) Make sure the default DSA_METHOD implementation only uses its
4777 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4778 and change its own handlers to be NULL so as to remove unnecessary
4779 indirection. This lets alternative implementations fallback to the
4780 default implementation more easily.
4783 *) Support for directoryName in GeneralName related extensions
4787 *) Make it possible to link applications using Makefile.shared.
4788 Make that possible even when linking against static libraries!
4791 *) Support for single pass processing for S/MIME signing. This now
4792 means that S/MIME signing can be done from a pipe, in addition
4793 cleartext signing (multipart/signed type) is effectively streaming
4794 and the signed data does not need to be all held in memory.
4796 This is done with a new flag PKCS7_STREAM. When this flag is set
4797 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4798 is done after the data is output (and digests calculated) in
4799 SMIME_write_PKCS7().
4802 *) Add full support for -rpath/-R, both in shared libraries and
4803 applications, at least on the platforms where it's known how
4807 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4808 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4809 will now compute a table of multiples of the generator that
4810 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4811 faster (notably in the case of a single point multiplication,
4812 scalar * generator).
4813 [Nils Larsch, Bodo Moeller]
4815 *) IPv6 support for certificate extensions. The various extensions
4816 which use the IP:a.b.c.d can now take IPv6 addresses using the
4817 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4821 *) Added an ENGINE that implements RSA by performing private key
4822 exponentiations with the GMP library. The conversions to and from
4823 GMP's mpz_t format aren't optimised nor are any montgomery forms
4824 cached, and on x86 it appears OpenSSL's own performance has caught up.
4825 However there are likely to be other architectures where GMP could
4826 provide a boost. This ENGINE is not built in by default, but it can be
4827 specified at Configure time and should be accompanied by the necessary
4828 linker additions, eg;
4829 ./config -DOPENSSL_USE_GMP -lgmp
4832 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4833 testing availability of engines with "-t" - the old behaviour is
4834 produced by increasing the feature's verbosity with "-tt".
4837 *) ECDSA routines: under certain error conditions uninitialized BN objects
4838 could be freed. Solution: make sure initialization is performed early
4839 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4843 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4844 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4845 software implementations. For DSA and DH, parameter generation can
4846 also be overriden by providing the appropriate method callbacks.
4849 *) Change the "progress" mechanism used in key-generation and
4850 primality testing to functions that take a new BN_GENCB pointer in
4851 place of callback/argument pairs. The new API functions have "_ex"
4852 postfixes and the older functions are reimplemented as wrappers for
4853 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4854 declarations of the old functions to help (graceful) attempts to
4855 migrate to the new functions. Also, the new key-generation API
4856 functions operate on a caller-supplied key-structure and return
4857 success/failure rather than returning a key or NULL - this is to
4858 help make "keygen" another member function of RSA_METHOD etc.
4860 Example for using the new callback interface:
4862 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4866 BN_GENCB_set(&my_cb, my_callback, my_arg);
4868 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4869 /* For the meaning of a, b in calls to my_callback(), see the
4870 * documentation of the function that calls the callback.
4871 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4872 * my_callback should return 1 if it wants BN_is_prime_ex()
4873 * to continue, or 0 to stop.
4878 *) Change the ZLIB compression method to be stateful, and make it
4879 available to TLS with the number defined in
4880 draft-ietf-tls-compression-04.txt.
4883 *) Add the ASN.1 structures and functions for CertificatePair, which
4884 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4886 CertificatePair ::= SEQUENCE {
4887 forward [0] Certificate OPTIONAL,
4888 reverse [1] Certificate OPTIONAL,
4889 -- at least one of the pair shall be present -- }
4891 Also implement the PEM functions to read and write certificate
4892 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4894 This needed to be defined, mostly for the sake of the LDAP
4895 attribute crossCertificatePair, but may prove useful elsewhere as
4899 *) Make it possible to inhibit symlinking of shared libraries in
4900 Makefile.shared, for Cygwin's sake.
4903 *) Extend the BIGNUM API by creating a function
4904 void BN_set_negative(BIGNUM *a, int neg);
4905 and a macro that behave like
4906 int BN_is_negative(const BIGNUM *a);
4908 to avoid the need to access 'a->neg' directly in applications.
4911 *) Implement fast modular reduction for pseudo-Mersenne primes
4912 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4913 EC_GROUP_new_curve_GFp() will now automatically use this
4915 [Nils Larsch <nla@trustcenter.de>]
4917 *) Add new lock type (CRYPTO_LOCK_BN).
4920 *) Change the ENGINE framework to automatically load engines
4921 dynamically from specific directories unless they could be
4922 found to already be built in or loaded. Move all the
4923 current engines except for the cryptodev one to a new
4925 The engines in engines/ are built as shared libraries if
4926 the "shared" options was given to ./Configure or ./config.
4927 Otherwise, they are inserted in libcrypto.a.
4928 /usr/local/ssl/engines is the default directory for dynamic
4929 engines, but that can be overriden at configure time through
4930 the usual use of --prefix and/or --openssldir, and at run
4931 time with the environment variable OPENSSL_ENGINES.
4932 [Geoff Thorpe and Richard Levitte]
4934 *) Add Makefile.shared, a helper makefile to build shared
4935 libraries. Adapt Makefile.org.
4938 *) Add version info to Win32 DLLs.
4939 [Peter 'Luna' Runestig" <peter@runestig.com>]
4941 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4942 can be added using this API to created arbitrary PKCS#12
4943 files while avoiding the low level API.
4945 New options to PKCS12_create(), key or cert can be NULL and
4946 will then be omitted from the output file. The encryption
4947 algorithm NIDs can be set to -1 for no encryption, the mac
4948 iteration count can be set to 0 to omit the mac.
4950 Enhance pkcs12 utility by making the -nokeys and -nocerts
4951 options work when creating a PKCS#12 file. New option -nomac
4952 to omit the mac, NONE can be set for an encryption algorithm.
4953 New code is modified to use the enhanced PKCS12_create()
4954 instead of the low level API.
4957 *) Extend ASN1 encoder to support indefinite length constructed
4958 encoding. This can output sequences tags and octet strings in
4959 this form. Modify pk7_asn1.c to support indefinite length
4960 encoding. This is experimental and needs additional code to
4961 be useful, such as an ASN1 bio and some enhanced streaming
4964 Extend template encode functionality so that tagging is passed
4965 down to the template encoder.
4968 *) Let 'openssl req' fail if an argument to '-newkey' is not
4969 recognized instead of using RSA as a default.
4972 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4973 As these are not official, they are not included in "ALL";
4974 the "ECCdraft" ciphersuite group alias can be used to select them.
4975 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4977 *) Add ECDH engine support.
4978 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4980 *) Add ECDH in new directory crypto/ecdh/.
4981 [Douglas Stebila (Sun Microsystems Laboratories)]
4983 *) Let BN_rand_range() abort with an error after 100 iterations
4984 without success (which indicates a broken PRNG).
4987 *) Change BN_mod_sqrt() so that it verifies that the input value
4988 is really the square of the return value. (Previously,
4989 BN_mod_sqrt would show GIGO behaviour.)
4992 *) Add named elliptic curves over binary fields from X9.62, SECG,
4993 and WAP/WTLS; add OIDs that were still missing.
4995 [Sheueling Chang Shantz and Douglas Stebila
4996 (Sun Microsystems Laboratories)]
4998 *) Extend the EC library for elliptic curves over binary fields
4999 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5002 EC_GF2m_simple_method
5006 EC_GROUP_new_curve_GF2m
5007 EC_GROUP_set_curve_GF2m
5008 EC_GROUP_get_curve_GF2m
5009 EC_POINT_set_affine_coordinates_GF2m
5010 EC_POINT_get_affine_coordinates_GF2m
5011 EC_POINT_set_compressed_coordinates_GF2m
5013 Point compression for binary fields is disabled by default for
5014 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5017 As binary polynomials are represented as BIGNUMs, various members
5018 of the EC_GROUP and EC_POINT data structures can be shared
5019 between the implementations for prime fields and binary fields;
5020 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5021 are essentially identical to their ..._GFp counterparts.
5022 (For simplicity, the '..._GFp' prefix has been dropped from
5023 various internal method names.)
5025 An internal 'field_div' method (similar to 'field_mul' and
5026 'field_sqr') has been added; this is used only for binary fields.
5028 [Sheueling Chang Shantz and Douglas Stebila
5029 (Sun Microsystems Laboratories)]
5031 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5032 through methods ('mul', 'precompute_mult').
5034 The generic implementations (now internally called 'ec_wNAF_mul'
5035 and 'ec_wNAF_precomputed_mult') remain the default if these
5036 methods are undefined.
5038 [Sheueling Chang Shantz and Douglas Stebila
5039 (Sun Microsystems Laboratories)]
5041 *) New function EC_GROUP_get_degree, which is defined through
5042 EC_METHOD. For curves over prime fields, this returns the bit
5043 length of the modulus.
5045 [Sheueling Chang Shantz and Douglas Stebila
5046 (Sun Microsystems Laboratories)]
5048 *) New functions EC_GROUP_dup, EC_POINT_dup.
5049 (These simply call ..._new and ..._copy).
5051 [Sheueling Chang Shantz and Douglas Stebila
5052 (Sun Microsystems Laboratories)]
5054 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5055 Polynomials are represented as BIGNUMs (where the sign bit is not
5056 used) in the following functions [macros]:
5059 BN_GF2m_sub [= BN_GF2m_add]
5060 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5061 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5062 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5064 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5065 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5066 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5067 BN_GF2m_cmp [= BN_ucmp]
5069 (Note that only the 'mod' functions are actually for fields GF(2^m).
5070 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5072 For some functions, an the irreducible polynomial defining a
5073 field can be given as an 'unsigned int[]' with strictly
5074 decreasing elements giving the indices of those bits that are set;
5075 i.e., p[] represents the polynomial
5076 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5078 p[0] > p[1] > ... > p[k] = 0.
5079 This applies to the following functions:
5084 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5085 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5087 BN_GF2m_mod_sqrt_arr
5088 BN_GF2m_mod_solve_quad_arr
5092 Conversion can be performed by the following functions:
5097 bntest.c has additional tests for binary polynomial arithmetic.
5099 Two implementations for BN_GF2m_mod_div() are available.
5100 The default algorithm simply uses BN_GF2m_mod_inv() and
5101 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5102 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5103 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5105 [Sheueling Chang Shantz and Douglas Stebila
5106 (Sun Microsystems Laboratories)]
5108 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5109 functionality is disabled at compile-time.
5110 [Douglas Stebila <douglas.stebila@sun.com>]
5112 *) Change default behaviour of 'openssl asn1parse' so that more
5113 information is visible when viewing, e.g., a certificate:
5115 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5116 mode the content of non-printable OCTET STRINGs is output in a
5117 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5118 avoid the appearance of a printable string.
5119 [Nils Larsch <nla@trustcenter.de>]
5121 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5123 EC_GROUP_set_asn1_flag()
5124 EC_GROUP_get_asn1_flag()
5125 EC_GROUP_set_point_conversion_form()
5126 EC_GROUP_get_point_conversion_form()
5127 These control ASN1 encoding details:
5128 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5129 has been set to OPENSSL_EC_NAMED_CURVE.
5130 - Points are encoded in uncompressed form by default; options for
5131 asn1_for are as for point2oct, namely
5132 POINT_CONVERSION_COMPRESSED
5133 POINT_CONVERSION_UNCOMPRESSED
5134 POINT_CONVERSION_HYBRID
5136 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5139 EC_GROUP_get0_seed()
5140 EC_GROUP_get_seed_len()
5141 This is used only for ASN1 purposes (so far).
5142 [Nils Larsch <nla@trustcenter.de>]
5144 *) Add 'field_type' member to EC_METHOD, which holds the NID
5145 of the appropriate field type OID. The new function
5146 EC_METHOD_get_field_type() returns this value.
5147 [Nils Larsch <nla@trustcenter.de>]
5152 EC_POINT_point2hex()
5153 EC_POINT_hex2point()
5154 providing useful interfaces to EC_POINT_point2oct() and
5155 EC_POINT_oct2point().
5156 [Nils Larsch <nla@trustcenter.de>]
5158 *) Change internals of the EC library so that the functions
5159 EC_GROUP_set_generator()
5160 EC_GROUP_get_generator()
5161 EC_GROUP_get_order()
5162 EC_GROUP_get_cofactor()
5163 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5164 to methods, which would lead to unnecessary code duplication when
5165 adding different types of curves.
5166 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5168 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5169 arithmetic, and such that modified wNAFs are generated
5170 (which avoid length expansion in many cases).
5173 *) Add a function EC_GROUP_check_discriminant() (defined via
5174 EC_METHOD) that verifies that the curve discriminant is non-zero.
5176 Add a function EC_GROUP_check() that makes some sanity tests
5177 on a EC_GROUP, its generator and order. This includes
5178 EC_GROUP_check_discriminant().
5179 [Nils Larsch <nla@trustcenter.de>]
5181 *) Add ECDSA in new directory crypto/ecdsa/.
5183 Add applications 'openssl ecparam' and 'openssl ecdsa'
5184 (these are based on 'openssl dsaparam' and 'openssl dsa').
5186 ECDSA support is also included in various other files across the
5187 library. Most notably,
5188 - 'openssl req' now has a '-newkey ecdsa:file' option;
5189 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5190 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5191 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5192 them suitable for ECDSA where domain parameters must be
5193 extracted before the specific public key;
5194 - ECDSA engine support has been added.
5195 [Nils Larsch <nla@trustcenter.de>]
5197 *) Include some named elliptic curves, and add OIDs from X9.62,
5198 SECG, and WAP/WTLS. Each curve can be obtained from the new
5200 EC_GROUP_new_by_curve_name(),
5201 and the list of available named curves can be obtained with
5202 EC_get_builtin_curves().
5203 Also add a 'curve_name' member to EC_GROUP objects, which can be
5205 EC_GROUP_set_curve_name()
5206 EC_GROUP_get_curve_name()
5207 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5209 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5210 was actually never needed) and in BN_mul(). The removal in BN_mul()
5211 required a small change in bn_mul_part_recursive() and the addition
5212 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5213 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5214 bn_sub_words() and bn_add_words() except they take arrays with
5218 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5220 *) Cleanse PEM buffers before freeing them since they may contain
5222 [Benjamin Bennett <ben@psc.edu>]
5224 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5225 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5226 authentication-only ciphersuites.
5229 *) Since AES128 and AES256 share a single mask bit in the logic of
5230 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5231 kludge to work properly if AES128 is available and AES256 isn't.
5234 *) Expand security boundary to match 1.1.1 module.
5237 *) Remove redundant features: hash file source, editing of test vectors
5238 modify fipsld to use external fips_premain.c signature.
5241 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5242 run algorithm test programs.
5245 *) Make algorithm test programs more tolerant of whitespace.
5248 *) Have SSL/TLS server implementation tolerate "mismatched" record
5249 protocol version while receiving ClientHello even if the
5250 ClientHello is fragmented. (The server can't insist on the
5251 particular protocol version it has chosen before the ServerHello
5252 message has informed the client about his choice.)
5255 *) Load error codes if they are not already present instead of using a
5256 static variable. This allows them to be cleanly unloaded and reloaded.
5259 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5261 *) Introduce limits to prevent malicious keys being able to
5262 cause a denial of service. (CVE-2006-2940)
5263 [Steve Henson, Bodo Moeller]
5265 *) Fix ASN.1 parsing of certain invalid structures that can result
5266 in a denial of service. (CVE-2006-2937) [Steve Henson]
5268 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5269 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5271 *) Fix SSL client code which could crash if connecting to a
5272 malicious SSLv2 server. (CVE-2006-4343)
5273 [Tavis Ormandy and Will Drewry, Google Security Team]
5275 *) Change ciphersuite string processing so that an explicit
5276 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5277 will no longer include "AES128-SHA"), and any other similar
5278 ciphersuite (same bitmap) from *other* protocol versions (so that
5279 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5280 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5281 changes from 0.9.8b and 0.9.8d.
5284 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5286 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5287 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5289 *) Change the Unix randomness entropy gathering to use poll() when
5290 possible instead of select(), since the latter has some
5291 undesirable limitations.
5292 [Darryl Miles via Richard Levitte and Bodo Moeller]
5294 *) Disable rogue ciphersuites:
5296 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5297 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5298 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5300 The latter two were purportedly from
5301 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5304 Also deactive the remaining ciphersuites from
5305 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5306 unofficial, and the ID has long expired.
5309 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5310 dual-core machines) and other potential thread-safety issues.
5313 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5315 *) Adapt fipsld and the build system to link against the validated FIPS
5316 module in FIPS mode.
5319 *) Fixes for VC++ 2005 build under Windows.
5322 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5323 from a Windows bash shell such as MSYS. It is autodetected from the
5324 "config" script when run from a VC++ environment. Modify standard VC++
5325 build to use fipscanister.o from the GNU make build.
5328 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5330 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5331 The value now differs depending on if you build for FIPS or not.
5332 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5333 safely run with a non-FIPSed libcrypto, as it may crash because of
5334 the difference induced by this change.
5337 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5339 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5340 (part of SSL_OP_ALL). This option used to disable the
5341 countermeasure against man-in-the-middle protocol-version
5342 rollback in the SSL 2.0 server implementation, which is a bad
5343 idea. (CVE-2005-2969)
5345 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5346 for Information Security, National Institute of Advanced Industrial
5347 Science and Technology [AIST], Japan)]
5349 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5350 mainly for FIPS compliance and not fully integrated at this stage.
5353 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5354 the exponentiation using a fixed-length exponent. (Otherwise,
5355 the information leaked through timing could expose the secret key
5356 after many signatures; cf. Bleichenbacher's attack on DSA with
5360 *) Make a new fixed-window mod_exp implementation the default for
5361 RSA, DSA, and DH private-key operations so that the sequence of
5362 squares and multiplies and the memory access pattern are
5363 independent of the particular secret key. This will mitigate
5364 cache-timing and potential related attacks.
5366 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5367 and this is automatically used by BN_mod_exp_mont() if the new flag
5368 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5369 will use this BN flag for private exponents unless the flag
5370 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5371 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5373 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5375 *) Change the client implementation for SSLv23_method() and
5376 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5377 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5378 (Previously, the SSL 2.0 backwards compatible Client Hello
5379 message format would be used even with SSL_OP_NO_SSLv2.)
5382 *) Add support for smime-type MIME parameter in S/MIME messages which some
5386 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5387 a threadsafe manner. Modify rsa code to use new function and add calls
5388 to dsa and dh code (which had race conditions before).
5391 *) Include the fixed error library code in the C error file definitions
5392 instead of fixing them up at runtime. This keeps the error code
5393 structures constant.
5396 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5398 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5401 *) Fixes for newer kerberos headers. NB: the casts are needed because
5402 the 'length' field is signed on one version and unsigned on another
5403 with no (?) obvious way to tell the difference, without these VC++
5404 complains. Also the "definition" of FAR (blank) is no longer included
5405 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5406 some needed definitions.
5409 *) Undo Cygwin change.
5412 *) Added support for proxy certificates according to RFC 3820.
5413 Because they may be a security thread to unaware applications,
5414 they must be explicitly allowed in run-time. See
5415 docs/HOWTO/proxy_certificates.txt for further information.
5418 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5420 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5421 server and client random values. Previously
5422 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5423 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5425 This change has negligible security impact because:
5427 1. Server and client random values still have 24 bytes of pseudo random
5430 2. Server and client random values are sent in the clear in the initial
5433 3. The master secret is derived using the premaster secret (48 bytes in
5434 size for static RSA ciphersuites) as well as client server and random
5437 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5440 [Stephen Henson, reported by UK NISCC]
5442 *) Use Windows randomness collection on Cygwin.
5445 *) Fix hang in EGD/PRNGD query when communication socket is closed
5446 prematurely by EGD/PRNGD.
5447 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5449 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5452 *) Back-port of selected performance improvements from development
5453 branch, as well as improved support for PowerPC platforms.
5456 *) Add lots of checks for memory allocation failure, error codes to indicate
5457 failure and freeing up memory if a failure occurs.
5458 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5460 *) Add new -passin argument to dgst.
5463 *) Perform some character comparisons of different types in X509_NAME_cmp:
5464 this is needed for some certificates that reencode DNs into UTF8Strings
5465 (in violation of RFC3280) and can't or wont issue name rollover
5469 *) Make an explicit check during certificate validation to see that
5470 the CA setting in each certificate on the chain is correct. As a
5471 side effect always do the following basic checks on extensions,
5472 not just when there's an associated purpose to the check:
5474 - if there is an unhandled critical extension (unless the user
5475 has chosen to ignore this fault)
5476 - if the path length has been exceeded (if one is set at all)
5477 - that certain extensions fit the associated purpose (if one has
5481 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5483 *) Avoid a race condition when CRLs are checked in a multi threaded
5484 environment. This would happen due to the reordering of the revoked
5485 entries during signature checking and serial number lookup. Now the
5486 encoding is cached and the serial number sort performed under a lock.
5487 Add new STACK function sk_is_sorted().
5490 *) Add Delta CRL to the extension code.
5493 *) Various fixes to s3_pkt.c so alerts are sent properly.
5494 [David Holmes <d.holmes@f5.com>]
5496 *) Reduce the chances of duplicate issuer name and serial numbers (in
5497 violation of RFC3280) using the OpenSSL certificate creation utilities.
5498 This is done by creating a random 64 bit value for the initial serial
5499 number when a serial number file is created or when a self signed
5500 certificate is created using 'openssl req -x509'. The initial serial
5501 number file is created using 'openssl x509 -next_serial' in CA.pl
5502 rather than being initialized to 1.
5505 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5507 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5508 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5509 [Joe Orton, Steve Henson]
5511 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5513 [Joe Orton, Steve Henson]
5515 *) Make it possible to have multiple active certificates with the same
5516 subject in the CA index file. This is done only if the keyword
5517 'unique_subject' is set to 'no' in the main CA section (default
5518 if 'CA_default') of the configuration file. The value is saved
5519 with the database itself in a separate index attribute file,
5520 named like the index file with '.attr' appended to the name.
5523 *) X509 verify fixes. Disable broken certificate workarounds when
5524 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5525 keyUsage extension present. Don't accept CRLs with unhandled critical
5526 extensions: since verify currently doesn't process CRL extensions this
5527 rejects a CRL with *any* critical extensions. Add new verify error codes
5531 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5532 A clarification of RFC2560 will require the use of OCTET STRINGs and
5533 some implementations cannot handle the current raw format. Since OpenSSL
5534 copies and compares OCSP nonces as opaque blobs without any attempt at
5535 parsing them this should not create any compatibility issues.
5538 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5539 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5540 this HMAC (and other) operations are several times slower than OpenSSL
5544 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5545 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5547 *) Use the correct content when signing type "other".
5550 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5552 *) Fix various bugs revealed by running the NISCC test suite:
5554 Stop out of bounds reads in the ASN1 code when presented with
5555 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5557 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5559 If verify callback ignores invalid public key errors don't try to check
5560 certificate signature with the NULL public key.
5564 *) New -ignore_err option in ocsp application to stop the server
5565 exiting on the first error in a request.
5568 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5569 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5573 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5574 extra data after the compression methods not only for TLS 1.0
5575 but also for SSL 3.0 (as required by the specification).
5576 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5578 *) Change X509_certificate_type() to mark the key as exported/exportable
5579 when it's 512 *bits* long, not 512 bytes.
5582 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5583 blocks during encryption.
5586 *) Various fixes to base64 BIO and non blocking I/O. On write
5587 flushes were not handled properly if the BIO retried. On read
5588 data was not being buffered properly and had various logic bugs.
5589 This also affects blocking I/O when the data being decoded is a
5593 *) Various S/MIME bugfixes and compatibility changes:
5594 output correct application/pkcs7 MIME type if
5595 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5596 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5597 of files as .eml work). Correctly handle very long lines in MIME
5601 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5603 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5604 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5605 a protocol version number mismatch like a decryption error
5606 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5609 *) Turn on RSA blinding by default in the default implementation
5610 to avoid a timing attack. Applications that don't want it can call
5611 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5612 They would be ill-advised to do so in most cases.
5613 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5615 *) Change RSA blinding code so that it works when the PRNG is not
5616 seeded (in this case, the secret RSA exponent is abused as
5617 an unpredictable seed -- if it is not unpredictable, there
5618 is no point in blinding anyway). Make RSA blinding thread-safe
5619 by remembering the creator's thread ID in rsa->blinding and
5620 having all other threads use local one-time blinding factors
5621 (this requires more computation than sharing rsa->blinding, but
5622 avoids excessive locking; and if an RSA object is not shared
5623 between threads, blinding will still be very fast).
5626 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5627 ENGINE as defaults for all supported algorithms irrespective of
5628 the 'flags' parameter. 'flags' is now honoured, so applications
5629 should make sure they are passing it correctly.
5632 *) Target "mingw" now allows native Windows code to be generated in
5633 the Cygwin environment as well as with the MinGW compiler.
5636 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5638 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5639 via timing by performing a MAC computation even if incorrect
5640 block cipher padding has been found. This is a countermeasure
5641 against active attacks where the attacker has to distinguish
5642 between bad padding and a MAC verification error. (CVE-2003-0078)
5644 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5645 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5646 Martin Vuagnoux (EPFL, Ilion)]
5648 *) Make the no-err option work as intended. The intention with no-err
5649 is not to have the whole error stack handling routines removed from
5650 libcrypto, it's only intended to remove all the function name and
5651 reason texts, thereby removing some of the footprint that may not
5652 be interesting if those errors aren't displayed anyway.
5654 NOTE: it's still possible for any application or module to have it's
5655 own set of error texts inserted. The routines are there, just not
5656 used by default when no-err is given.
5659 *) Add support for FreeBSD on IA64.
5660 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5662 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5663 Kerberos function mit_des_cbc_cksum(). Before this change,
5664 the value returned by DES_cbc_cksum() was like the one from
5665 mit_des_cbc_cksum(), except the bytes were swapped.
5666 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5668 *) Allow an application to disable the automatic SSL chain building.
5669 Before this a rather primitive chain build was always performed in
5670 ssl3_output_cert_chain(): an application had no way to send the
5671 correct chain if the automatic operation produced an incorrect result.
5673 Now the chain builder is disabled if either:
5675 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5677 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5679 The reasoning behind this is that an application would not want the
5680 auto chain building to take place if extra chain certificates are
5681 present and it might also want a means of sending no additional
5682 certificates (for example the chain has two certificates and the
5686 *) Add the possibility to build without the ENGINE framework.
5687 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5689 *) Under Win32 gmtime() can return NULL: check return value in
5690 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5693 *) DSA routines: under certain error conditions uninitialized BN objects
5694 could be freed. Solution: make sure initialization is performed early
5695 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5696 Nils Larsch <nla@trustcenter.de> via PR#459)
5699 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5700 checked on reconnect on the client side, therefore session resumption
5701 could still fail with a "ssl session id is different" error. This
5702 behaviour is masked when SSL_OP_ALL is used due to
5703 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5704 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5705 followup to PR #377.
5708 *) IA-32 assembler support enhancements: unified ELF targets, support
5709 for SCO/Caldera platforms, fix for Cygwin shared build.
5712 *) Add support for FreeBSD on sparc64. As a consequence, support for
5713 FreeBSD on non-x86 processors is separate from x86 processors on
5714 the config script, much like the NetBSD support.
5715 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5717 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5719 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5722 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5723 code (06) was taken as the first octet of the session ID and the last
5724 octet was ignored consequently. As a result SSLv2 client side session
5725 caching could not have worked due to the session ID mismatch between
5727 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5731 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5732 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5736 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5737 seems that in spite of existing for more than a year, many application
5738 author have done nothing to provide the necessary callbacks, which
5739 means that this particular engine will not work properly anywhere.
5740 This is a very unfortunate situation which forces us, in the name
5741 of usability, to give the hw_ncipher.c a static lock, which is part
5743 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5744 appear in 0.9.8 or later. We EXPECT application authors to have
5745 dealt properly with this when 0.9.8 is released (unless we actually
5746 make such changes in the libcrypto locking code that changes will
5747 have to be made anyway).
5750 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5751 octets have been read, EOF or an error occurs. Without this change
5752 some truncated ASN1 structures will not produce an error.
5755 *) Disable Heimdal support, since it hasn't been fully implemented.
5756 Still give the possibility to force the use of Heimdal, but with
5757 warnings and a request that patches get sent to openssl-dev.
5760 *) Add the VC-CE target, introduce the WINCE sysname, and add
5761 INSTALL.WCE and appropriate conditionals to make it build.
5762 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5764 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5765 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5766 edit numbers of the version.
5767 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5769 *) Introduce safe string copy and catenation functions
5770 (BUF_strlcpy() and BUF_strlcat()).
5771 [Ben Laurie (CHATS) and Richard Levitte]
5773 *) Avoid using fixed-size buffers for one-line DNs.
5774 [Ben Laurie (CHATS)]
5776 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5777 resizing buffers containing secrets, and use where appropriate.
5778 [Ben Laurie (CHATS)]
5780 *) Avoid using fixed size buffers for configuration file location.
5781 [Ben Laurie (CHATS)]
5783 *) Avoid filename truncation for various CA files.
5784 [Ben Laurie (CHATS)]
5786 *) Use sizeof in preference to magic numbers.
5787 [Ben Laurie (CHATS)]
5789 *) Avoid filename truncation in cert requests.
5790 [Ben Laurie (CHATS)]
5792 *) Add assertions to check for (supposedly impossible) buffer
5794 [Ben Laurie (CHATS)]
5796 *) Don't cache truncated DNS entries in the local cache (this could
5797 potentially lead to a spoofing attack).
5798 [Ben Laurie (CHATS)]
5800 *) Fix various buffers to be large enough for hex/decimal
5801 representations in a platform independent manner.
5802 [Ben Laurie (CHATS)]
5804 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5805 resizing buffers containing secrets, and use where appropriate.
5806 [Ben Laurie (CHATS)]
5808 *) Add BIO_indent() to avoid much slightly worrying code to do
5810 [Ben Laurie (CHATS)]
5812 *) Convert sprintf()/BIO_puts() to BIO_printf().
5813 [Ben Laurie (CHATS)]
5815 *) buffer_gets() could terminate with the buffer only half
5817 [Ben Laurie (CHATS)]
5819 *) Add assertions to prevent user-supplied crypto functions from
5820 overflowing internal buffers by having large block sizes, etc.
5821 [Ben Laurie (CHATS)]
5823 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5825 [Ben Laurie (CHATS)]
5827 *) Eliminate unused copy of key in RC4.
5828 [Ben Laurie (CHATS)]
5830 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5831 [Ben Laurie (CHATS)]
5833 *) Fix off-by-one error in EGD path.
5834 [Ben Laurie (CHATS)]
5836 *) If RANDFILE path is too long, ignore instead of truncating.
5837 [Ben Laurie (CHATS)]
5839 *) Eliminate unused and incorrectly sized X.509 structure
5841 [Ben Laurie (CHATS)]
5843 *) Eliminate unused and dangerous function knumber().
5844 [Ben Laurie (CHATS)]
5846 *) Eliminate unused and dangerous structure, KSSL_ERR.
5847 [Ben Laurie (CHATS)]
5849 *) Protect against overlong session ID context length in an encoded
5850 session object. Since these are local, this does not appear to be
5852 [Ben Laurie (CHATS)]
5854 *) Change from security patch (see 0.9.6e below) that did not affect
5855 the 0.9.6 release series:
5857 Remote buffer overflow in SSL3 protocol - an attacker could
5858 supply an oversized master key in Kerberos-enabled versions.
5860 [Ben Laurie (CHATS)]
5862 *) Change the SSL kerb5 codes to match RFC 2712.
5865 *) Make -nameopt work fully for req and add -reqopt switch.
5866 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5868 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5869 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5871 *) Make sure tests can be performed even if the corresponding algorithms
5872 have been removed entirely. This was also the last step to make
5873 OpenSSL compilable with DJGPP under all reasonable conditions.
5874 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5876 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5877 to allow version independent disabling of normally unselected ciphers,
5878 which may be activated as a side-effect of selecting a single cipher.
5880 (E.g., cipher list string "RSA" enables ciphersuites that are left
5881 out of "ALL" because they do not provide symmetric encryption.
5882 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5883 [Lutz Jaenicke, Bodo Moeller]
5885 *) Add appropriate support for separate platform-dependent build
5886 directories. The recommended way to make a platform-dependent
5887 build directory is the following (tested on Linux), maybe with
5890 # Place yourself outside of the OpenSSL source tree. In
5891 # this example, the environment variable OPENSSL_SOURCE
5892 # is assumed to contain the absolute OpenSSL source directory.
5893 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5894 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5895 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5896 mkdir -p `dirname $F`
5897 ln -s $OPENSSL_SOURCE/$F $F
5900 To be absolutely sure not to disturb the source tree, a "make clean"
5901 is a good thing. If it isn't successful, don't worry about it,
5902 it probably means the source directory is very clean.
5905 *) Make sure any ENGINE control commands make local copies of string
5906 pointers passed to them whenever necessary. Otherwise it is possible
5907 the caller may have overwritten (or deallocated) the original string
5908 data when a later ENGINE operation tries to use the stored values.
5909 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5911 *) Improve diagnostics in file reading and command-line digests.
5912 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5914 *) Add AES modes CFB and OFB to the object database. Correct an
5915 error in AES-CFB decryption.
5918 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5919 allows existing EVP_CIPHER_CTX structures to be reused after
5920 calling EVP_*Final(). This behaviour is used by encryption
5921 BIOs and some applications. This has the side effect that
5922 applications must explicitly clean up cipher contexts with
5923 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5926 *) Check the values of dna and dnb in bn_mul_recursive before calling
5927 bn_mul_comba (a non zero value means the a or b arrays do not contain
5928 n2 elements) and fallback to bn_mul_normal if either is not zero.
5931 *) Fix escaping of non-ASCII characters when using the -subj option
5932 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5935 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5936 form for "surname", serialNumber has no short form.
5937 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5938 therefore remove "mail" short name for "internet 7".
5939 The OID for unique identifiers in X509 certificates is
5940 x500UniqueIdentifier, not uniqueIdentifier.
5941 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5944 *) Add an "init" command to the ENGINE config module and auto initialize
5945 ENGINEs. Without any "init" command the ENGINE will be initialized
5946 after all ctrl commands have been executed on it. If init=1 the
5947 ENGINE is initailized at that point (ctrls before that point are run
5948 on the uninitialized ENGINE and after on the initialized one). If
5949 init=0 then the ENGINE will not be iniatialized at all.
5952 *) Fix the 'app_verify_callback' interface so that the user-defined
5953 argument is actually passed to the callback: In the
5954 SSL_CTX_set_cert_verify_callback() prototype, the callback
5955 declaration has been changed from
5958 int (*cb)(X509_STORE_CTX *,void *);
5959 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5960 i=s->ctx->app_verify_callback(&ctx)
5961 has been changed into
5962 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5964 To update applications using SSL_CTX_set_cert_verify_callback(),
5965 a dummy argument can be added to their callback functions.
5966 [D. K. Smetters <smetters@parc.xerox.com>]
5968 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5969 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5971 *) Add and OPENSSL_LOAD_CONF define which will cause
5972 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5973 This allows older applications to transparently support certain
5974 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5975 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5976 load the config file and OPENSSL_add_all_algorithms_conf() which will
5977 always load it have also been added.
5980 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5981 Adjust NIDs and EVP layer.
5982 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5984 *) Config modules support in openssl utility.
5986 Most commands now load modules from the config file,
5987 though in a few (such as version) this isn't done
5988 because it couldn't be used for anything.
5990 In the case of ca and req the config file used is
5991 the same as the utility itself: that is the -config
5992 command line option can be used to specify an
5996 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5997 use "openssl_conf" if filename is NULL use default openssl config file.
6000 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6001 config section name. Add a new flag to tolerate a missing config file
6002 and move code to CONF_modules_load_file().
6005 *) Support for crypto accelerator cards from Accelerated Encryption
6006 Processing, www.aep.ie. (Use engine 'aep')
6007 The support was copied from 0.9.6c [engine] and adapted/corrected
6008 to work with the new engine framework.
6009 [AEP Inc. and Richard Levitte]
6011 *) Support for SureWare crypto accelerator cards from Baltimore
6012 Technologies. (Use engine 'sureware')
6013 The support was copied from 0.9.6c [engine] and adapted
6014 to work with the new engine framework.
6017 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6018 make the newer ENGINE framework commands for the CHIL engine work.
6019 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6021 *) Make it possible to produce shared libraries on ReliantUNIX.
6022 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6024 *) Add the configuration target debug-linux-ppro.
6025 Make 'openssl rsa' use the general key loading routines
6026 implemented in apps.c, and make those routines able to
6027 handle the key format FORMAT_NETSCAPE and the variant
6029 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6031 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6032 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6034 *) Add -keyform to rsautl, and document -engine.
6035 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6037 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6038 BIO_R_NO_SUCH_FILE error code rather than the generic
6039 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6042 *) Add new functions
6044 ERR_peek_last_error_line
6045 ERR_peek_last_error_line_data.
6046 These are similar to
6049 ERR_peek_error_line_data,
6050 but report on the latest error recorded rather than the first one
6051 still in the error queue.
6052 [Ben Laurie, Bodo Moeller]
6054 *) default_algorithms option in ENGINE config module. This allows things
6056 default_algorithms = ALL
6057 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6060 *) Preliminary ENGINE config module.
6063 *) New experimental application configuration code.
6066 *) Change the AES code to follow the same name structure as all other
6067 symmetric ciphers, and behave the same way. Move everything to
6068 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6069 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6071 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6072 [Ben Laurie and Theo de Raadt]
6074 *) Add option to output public keys in req command.
6075 [Massimiliano Pala madwolf@openca.org]
6077 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6078 (up to about 10% better than before for P-192 and P-224).
6081 *) New functions/macros
6083 SSL_CTX_set_msg_callback(ctx, cb)
6084 SSL_CTX_set_msg_callback_arg(ctx, arg)
6085 SSL_set_msg_callback(ssl, cb)
6086 SSL_set_msg_callback_arg(ssl, arg)
6088 to request calling a callback function
6090 void cb(int write_p, int version, int content_type,
6091 const void *buf, size_t len, SSL *ssl, void *arg)
6093 whenever a protocol message has been completely received
6094 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6095 protocol version according to which the SSL library interprets
6096 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6097 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6098 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6099 specification (change_cipher_spec(20), alert(21), handshake(22)).
6100 'buf' and 'len' point to the actual message, 'ssl' to the
6101 SSL object, and 'arg' is the application-defined value set by
6102 SSL[_CTX]_set_msg_callback_arg().
6104 'openssl s_client' and 'openssl s_server' have new '-msg' options
6105 to enable a callback that displays all protocol messages.
6108 *) Change the shared library support so shared libraries are built as
6109 soon as the corresponding static library is finished, and thereby get
6110 openssl and the test programs linked against the shared library.
6111 This still only happens when the keyword "shard" has been given to
6112 the configuration scripts.
6114 NOTE: shared library support is still an experimental thing, and
6115 backward binary compatibility is still not guaranteed.
6116 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6118 *) Add support for Subject Information Access extension.
6119 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6121 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6122 additional bytes when new memory had to be allocated, not just
6123 when reusing an existing buffer.
6126 *) New command line and configuration option 'utf8' for the req command.
6127 This allows field values to be specified as UTF8 strings.
6130 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6131 runs for the former and machine-readable output for the latter.
6134 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6135 of the e-mail address in the DN (i.e., it will go into a certificate
6136 extension only). The new configuration file option 'email_in_dn = no'
6137 has the same effect.
6138 [Massimiliano Pala madwolf@openca.org]
6140 *) Change all functions with names starting with des_ to be starting
6141 with DES_ instead. Add wrappers that are compatible with libdes,
6142 but are named _ossl_old_des_*. Finally, add macros that map the
6143 des_* symbols to the corresponding _ossl_old_des_* if libdes
6144 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6145 desired, the des_* symbols will be mapped to DES_*, with one
6148 Since we provide two compatibility mappings, the user needs to
6149 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6150 compatibility is desired. The default (i.e., when that macro
6151 isn't defined) is OpenSSL 0.9.6c compatibility.
6153 There are also macros that enable and disable the support of old
6154 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6155 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6156 are defined, the default will apply: to support the old des routines.
6158 In either case, one must include openssl/des.h to get the correct
6159 definitions. Do not try to just include openssl/des_old.h, that
6162 NOTE: This is a major break of an old API into a new one. Software
6163 authors are encouraged to switch to the DES_ style functions. Some
6164 time in the future, des_old.h and the libdes compatibility functions
6165 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6166 default), and then completely removed.
6169 *) Test for certificates which contain unsupported critical extensions.
6170 If such a certificate is found during a verify operation it is
6171 rejected by default: this behaviour can be overridden by either
6172 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6173 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6174 X509_supported_extension() has also been added which returns 1 if a
6175 particular extension is supported.
6178 *) Modify the behaviour of EVP cipher functions in similar way to digests
6179 to retain compatibility with existing code.
6182 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6183 compatibility with existing code. In particular the 'ctx' parameter does
6184 not have to be to be initialized before the call to EVP_DigestInit() and
6185 it is tidied up after a call to EVP_DigestFinal(). New function
6186 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6187 EVP_MD_CTX_copy() changed to not require the destination to be
6188 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6189 requires the destination to be valid.
6191 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6192 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6195 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6196 so that complete 'Handshake' protocol structures are kept in memory
6197 instead of overwriting 'msg_type' and 'length' with 'body' data.
6200 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6201 [Massimo Santin via Richard Levitte]
6203 *) Major restructuring to the underlying ENGINE code. This includes
6204 reduction of linker bloat, separation of pure "ENGINE" manipulation
6205 (initialisation, etc) from functionality dealing with implementations
6206 of specific crypto iterfaces. This change also introduces integrated
6207 support for symmetric ciphers and digest implementations - so ENGINEs
6208 can now accelerate these by providing EVP_CIPHER and EVP_MD
6209 implementations of their own. This is detailed in crypto/engine/README
6210 as it couldn't be adequately described here. However, there are a few
6211 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6212 were changed in the original introduction of ENGINE code have now
6213 reverted back - the hooking from this code to ENGINE is now a good
6214 deal more passive and at run-time, operations deal directly with
6215 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6216 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6217 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6218 they were not being used by the framework as there is no concept of a
6219 BIGNUM_METHOD and they could not be generalised to the new
6220 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6221 ENGINE_cpy() has been removed as it cannot be consistently defined in
6225 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6228 *) Change mkdef.pl to sort symbols that get the same entry number,
6229 and make sure the automatically generated functions ERR_load_*
6230 become part of libeay.num as well.
6233 *) New function SSL_renegotiate_pending(). This returns true once
6234 renegotiation has been requested (either SSL_renegotiate() call
6235 or HelloRequest/ClientHello received from the peer) and becomes
6236 false once a handshake has been completed.
6237 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6238 sends a HelloRequest, but does not ensure that a handshake takes
6239 place. SSL_renegotiate_pending() is useful for checking if the
6240 client has followed the request.)
6243 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6244 By default, clients may request session resumption even during
6245 renegotiation (if session ID contexts permit); with this option,
6246 session resumption is possible only in the first handshake.
6248 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6249 more bits available for options that should not be part of
6250 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6253 *) Add some demos for certificate and certificate request creation.
6256 *) Make maximum certificate chain size accepted from the peer application
6257 settable (SSL*_get/set_max_cert_list()), as proposed by
6258 "Douglas E. Engert" <deengert@anl.gov>.
6261 *) Add support for shared libraries for Unixware-7
6262 (Boyd Lynn Gerber <gerberb@zenez.com>).
6265 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6266 be done prior to destruction. Use this to unload error strings from
6267 ENGINEs that load their own error strings. NB: This adds two new API
6268 functions to "get" and "set" this destroy handler in an ENGINE.
6271 *) Alter all existing ENGINE implementations (except "openssl" and
6272 "openbsd") to dynamically instantiate their own error strings. This
6273 makes them more flexible to be built both as statically-linked ENGINEs
6274 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6275 Also, add stub code to each that makes building them as self-contained
6276 shared-libraries easier (see README.ENGINE).
6279 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6280 implementations into applications that are completely implemented in
6281 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6282 commands that can be used to configure what shared-library to load and
6283 to control aspects of the way it is handled. Also, made an update to
6284 the README.ENGINE file that brings its information up-to-date and
6285 provides some information and instructions on the "dynamic" ENGINE
6286 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6289 *) Make it possible to unload ranges of ERR strings with a new
6290 "ERR_unload_strings" function.
6293 *) Add a copy() function to EVP_MD.
6296 *) Make EVP_MD routines take a context pointer instead of just the
6297 md_data void pointer.
6300 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6301 that the digest can only process a single chunk of data
6302 (typically because it is provided by a piece of
6303 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6304 is only going to provide a single chunk of data, and hence the
6305 framework needn't accumulate the data for oneshot drivers.
6308 *) As with "ERR", make it possible to replace the underlying "ex_data"
6309 functions. This change also alters the storage and management of global
6310 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6311 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6312 index counters. The API functions that use this state have been changed
6313 to take a "class_index" rather than pointers to the class's local STACK
6314 and counter, and there is now an API function to dynamically create new
6315 classes. This centralisation allows us to (a) plug a lot of the
6316 thread-safety problems that existed, and (b) makes it possible to clean
6317 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6318 such data would previously have always leaked in application code and
6319 workarounds were in place to make the memory debugging turn a blind eye
6320 to it. Application code that doesn't use this new function will still
6321 leak as before, but their memory debugging output will announce it now
6322 rather than letting it slide.
6324 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6325 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6326 has a return value to indicate success or failure.
6329 *) Make it possible to replace the underlying "ERR" functions such that the
6330 global state (2 LHASH tables and 2 locks) is only used by the "default"
6331 implementation. This change also adds two functions to "get" and "set"
6332 the implementation prior to it being automatically set the first time
6333 any other ERR function takes place. Ie. an application can call "get",
6334 pass the return value to a module it has just loaded, and that module
6335 can call its own "set" function using that value. This means the
6336 module's "ERR" operations will use (and modify) the error state in the
6337 application and not in its own statically linked copy of OpenSSL code.
6340 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6341 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6342 the operation, and provides a more encapsulated way for external code
6343 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6344 to use these functions rather than manually incrementing the counts.
6346 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6349 *) Add EVP test program.
6352 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6355 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6356 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6357 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6358 These allow a CRL to be built without having to access X509_CRL fields
6359 directly. Modify 'ca' application to use new functions.
6362 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6363 bug workarounds. Rollback attack detection is a security feature.
6364 The problem will only arise on OpenSSL servers when TLSv1 is not
6365 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6366 Software authors not wanting to support TLSv1 will have special reasons
6367 for their choice and can explicitly enable this option.
6368 [Bodo Moeller, Lutz Jaenicke]
6370 *) Rationalise EVP so it can be extended: don't include a union of
6371 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6372 (similar to those existing for EVP_CIPHER_CTX).
6377 EVP_MD_CTX_init(&md); /* new function call */
6378 EVP_DigestInit(&md, EVP_sha1());
6379 EVP_DigestUpdate(&md, in, len);
6380 EVP_DigestFinal(&md, out, NULL);
6381 EVP_MD_CTX_cleanup(&md); /* new function call */
6385 *) Make DES key schedule conform to the usual scheme, as well as
6386 correcting its structure. This means that calls to DES functions
6387 now have to pass a pointer to a des_key_schedule instead of a
6388 plain des_key_schedule (which was actually always a pointer
6391 des_key_schedule ks;
6393 des_set_key_checked(..., &ks);
6394 des_ncbc_encrypt(..., &ks, ...);
6396 (Note that a later change renames 'des_...' into 'DES_...'.)
6399 *) Initial reduction of linker bloat: the use of some functions, such as
6400 PEM causes large amounts of unused functions to be linked in due to
6401 poor organisation. For example pem_all.c contains every PEM function
6402 which has a knock on effect of linking in large amounts of (unused)
6403 ASN1 code. Grouping together similar functions and splitting unrelated
6404 functions prevents this.
6407 *) Cleanup of EVP macros.
6410 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6411 correct _ecb suffix.
6414 *) Add initial OCSP responder support to ocsp application. The
6415 revocation information is handled using the text based index
6416 use by the ca application. The responder can either handle
6417 requests generated internally, supplied in files (for example
6418 via a CGI script) or using an internal minimal server.
6421 *) Add configuration choices to get zlib compression for TLS.
6424 *) Changes to Kerberos SSL for RFC 2712 compliance:
6425 1. Implemented real KerberosWrapper, instead of just using
6426 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6427 2. Implemented optional authenticator field of KerberosWrapper.
6429 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6430 and authenticator structs; see crypto/krb5/.
6432 Generalized Kerberos calls to support multiple Kerberos libraries.
6433 [Vern Staats <staatsvr@asc.hpc.mil>,
6434 Jeffrey Altman <jaltman@columbia.edu>
6435 via Richard Levitte]
6437 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6438 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6439 values for each of the key sizes rather than having just
6440 parameters (and 'speed' generating keys each time).
6443 *) Speed up EVP routines.
6446 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6447 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6448 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6449 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6451 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6452 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6453 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6456 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6458 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6461 *) Added the OS2-EMX target.
6462 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6464 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6465 to support NCONF routines in extension code. New function CONF_set_nconf()
6466 to allow functions which take an NCONF to also handle the old LHASH
6467 structure: this means that the old CONF compatible routines can be
6468 retained (in particular wrt extensions) without having to duplicate the
6469 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6472 *) Enhance the general user interface with mechanisms for inner control
6473 and with possibilities to have yes/no kind of prompts.
6476 *) Change all calls to low level digest routines in the library and
6477 applications to use EVP. Add missing calls to HMAC_cleanup() and
6478 don't assume HMAC_CTX can be copied using memcpy().
6479 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6481 *) Add the possibility to control engines through control names but with
6482 arbitrary arguments instead of just a string.
6483 Change the key loaders to take a UI_METHOD instead of a callback
6484 function pointer. NOTE: this breaks binary compatibility with earlier
6485 versions of OpenSSL [engine].
6486 Adapt the nCipher code for these new conditions and add a card insertion
6490 *) Enhance the general user interface with mechanisms to better support
6491 dialog box interfaces, application-defined prompts, the possibility
6492 to use defaults (for example default passwords from somewhere else)
6493 and interrupts/cancellations.
6496 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6497 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6500 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6501 tidy up some unnecessarily weird code in 'sk_new()').
6502 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6504 *) Change the key loading routines for ENGINEs to use the same kind
6505 callback (pem_password_cb) as all other routines that need this
6509 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6510 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6511 than this minimum value is recommended.
6514 *) New random seeder for OpenVMS, using the system process statistics
6515 that are easily reachable.
6518 *) Windows apparently can't transparently handle global
6519 variables defined in DLLs. Initialisations such as:
6521 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6523 wont compile. This is used by the any applications that need to
6524 declare their own ASN1 modules. This was fixed by adding the option
6525 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6526 needed for static libraries under Win32.
6529 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6530 setting of purpose and trust fields. New X509_STORE trust and
6531 purpose functions and tidy up setting in other SSL functions.
6534 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6535 structure. These are inherited by X509_STORE_CTX when it is
6536 initialised. This allows various defaults to be set in the
6537 X509_STORE structure (such as flags for CRL checking and custom
6538 purpose or trust settings) for functions which only use X509_STORE_CTX
6539 internally such as S/MIME.
6541 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6542 trust settings if they are not set in X509_STORE. This allows X509_STORE
6543 purposes and trust (in S/MIME for example) to override any set by default.
6545 Add command line options for CRL checking to smime, s_client and s_server
6549 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6550 are set then the CRL is looked up in the X509_STORE structure and
6551 its validity and signature checked, then if the certificate is found
6552 in the CRL the verify fails with a revoked error.
6554 Various new CRL related callbacks added to X509_STORE_CTX structure.
6556 Command line options added to 'verify' application to support this.
6558 This needs some additional work, such as being able to handle multiple
6559 CRLs with different times, extension based lookup (rather than just
6560 by subject name) and ultimately more complete V2 CRL extension
6564 *) Add a general user interface API (crypto/ui/). This is designed
6565 to replace things like des_read_password and friends (backward
6566 compatibility functions using this new API are provided).
6567 The purpose is to remove prompting functions from the DES code
6568 section as well as provide for prompting through dialog boxes in
6569 a window system and the like.
6572 *) Add "ex_data" support to ENGINE so implementations can add state at a
6573 per-structure level rather than having to store it globally.
6576 *) Make it possible for ENGINE structures to be copied when retrieved by
6577 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6578 This causes the "original" ENGINE structure to act like a template,
6579 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6580 operational state can be localised to each ENGINE structure, despite the
6581 fact they all share the same "methods". New ENGINE structures returned in
6582 this case have no functional references and the return value is the single
6583 structural reference. This matches the single structural reference returned
6584 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6588 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6589 needs to match any other type at all we need to manually clear the
6593 *) Changes to the "openssl engine" utility to include;
6594 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6595 about an ENGINE's available control commands.
6596 - executing control commands from command line arguments using the
6597 '-pre' and '-post' switches. '-post' is only used if '-t' is
6598 specified and the ENGINE is successfully initialised. The syntax for
6599 the individual commands are colon-separated, for example;
6600 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6603 *) New dynamic control command support for ENGINEs. ENGINEs can now
6604 declare their own commands (numbers), names (strings), descriptions,
6605 and input types for run-time discovery by calling applications. A
6606 subset of these commands are implicitly classed as "executable"
6607 depending on their input type, and only these can be invoked through
6608 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6609 can be based on user input, config files, etc). The distinction is
6610 that "executable" commands cannot return anything other than a boolean
6611 result and can only support numeric or string input, whereas some
6612 discoverable commands may only be for direct use through
6613 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6614 pointers, or other custom uses. The "executable" commands are to
6615 support parameterisations of ENGINE behaviour that can be
6616 unambiguously defined by ENGINEs and used consistently across any
6617 OpenSSL-based application. Commands have been added to all the
6618 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6619 control over shared-library paths without source code alterations.
6622 *) Changed all ENGINE implementations to dynamically allocate their
6623 ENGINEs rather than declaring them statically. Apart from this being
6624 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6625 this also allows the implementations to compile without using the
6626 internal engine_int.h header.
6629 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6630 'const' value. Any code that should be able to modify a RAND_METHOD
6631 should already have non-const pointers to it (ie. they should only
6632 modify their own ones).
6635 *) Made a variety of little tweaks to the ENGINE code.
6636 - "atalla" and "ubsec" string definitions were moved from header files
6637 to C code. "nuron" string definitions were placed in variables
6638 rather than hard-coded - allowing parameterisation of these values
6639 later on via ctrl() commands.
6640 - Removed unused "#if 0"'d code.
6641 - Fixed engine list iteration code so it uses ENGINE_free() to release
6642 structural references.
6643 - Constified the RAND_METHOD element of ENGINE structures.
6644 - Constified various get/set functions as appropriate and added
6645 missing functions (including a catch-all ENGINE_cpy that duplicates
6646 all ENGINE values onto a new ENGINE except reference counts/state).
6647 - Removed NULL parameter checks in get/set functions. Setting a method
6648 or function to NULL is a way of cancelling out a previously set
6649 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6650 and doesn't justify the extra error symbols and code.
6651 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6652 flags from engine_int.h to engine.h.
6653 - Changed prototypes for ENGINE handler functions (init(), finish(),
6654 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6657 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6658 to the algorithm using long division. The binary algorithm can be
6659 used only if the modulus is odd. On 32-bit systems, it is faster
6660 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6661 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6662 up to 450 bits. In 64-bit environments, the binary algorithm
6663 appears to be advantageous for much longer moduli; here we use it
6664 for moduli up to 2048 bits.
6667 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6668 could not support the combine flag in choice fields.
6671 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6672 extensions from a certificate request to the certificate.
6675 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6676 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6677 file: this allows the display of the certificate about to be
6678 signed to be customised, to allow certain fields to be included
6679 or excluded and extension details. The old system didn't display
6680 multicharacter strings properly, omitted fields not in the policy
6681 and couldn't display additional details such as extensions.
6684 *) Function EC_POINTs_mul for multiple scalar multiplication
6685 of an arbitrary number of elliptic curve points
6686 \sum scalars[i]*points[i],
6687 optionally including the generator defined for the EC_GROUP:
6688 scalar*generator + \sum scalars[i]*points[i].
6690 EC_POINT_mul is a simple wrapper function for the typical case
6691 that the point list has just one item (besides the optional
6695 *) First EC_METHODs for curves over GF(p):
6697 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6698 operations and provides various method functions that can also
6699 operate with faster implementations of modular arithmetic.
6701 EC_GFp_mont_method() reuses most functions that are part of
6702 EC_GFp_simple_method, but uses Montgomery arithmetic.
6704 [Bodo Moeller; point addition and point doubling
6705 implementation directly derived from source code provided by
6706 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6708 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6709 crypto/ec/ec_lib.c):
6711 Curves are EC_GROUP objects (with an optional group generator)
6712 based on EC_METHODs that are built into the library.
6714 Points are EC_POINT objects based on EC_GROUP objects.
6716 Most of the framework would be able to handle curves over arbitrary
6717 finite fields, but as there are no obvious types for fields other
6718 than GF(p), some functions are limited to that for now.
6721 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6722 that the file contains a complete HTTP response.
6725 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6726 change the def and num file printf format specifier from "%-40sXXX"
6727 to "%-39s XXX". The latter will always guarantee a space after the
6728 field while the former will cause them to run together if the field
6729 is 40 of more characters long.
6732 *) Constify the cipher and digest 'method' functions and structures
6733 and modify related functions to take constant EVP_MD and EVP_CIPHER
6737 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6738 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6741 *) Modify EVP_Digest*() routines so they now return values. Although the
6742 internal software routines can never fail additional hardware versions
6746 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6748 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6749 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6752 ERR_R_NESTED_ASN1_ERROR
6754 ERR_R_MISSING_ASN1_EOS
6755 were 4 .. 9, conflicting with
6756 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6758 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6759 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6761 Add new error code 'ERR_R_INTERNAL_ERROR'.
6764 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6768 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6769 sets the subject name for a new request or supersedes the
6770 subject name in a given request. Formats that can be parsed are
6771 'CN=Some Name, OU=myOU, C=IT'
6773 'CN=Some Name/OU=myOU/C=IT'.
6775 Add options '-batch' and '-verbose' to 'openssl req'.
6776 [Massimiliano Pala <madwolf@hackmasters.net>]
6778 *) Introduce the possibility to access global variables through
6779 functions on platform were that's the best way to handle exporting
6780 global variables in shared libraries. To enable this functionality,
6781 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6782 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6783 is normally done by Configure or something similar).
6785 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6786 in the source file (foo.c) like this:
6788 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6789 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6791 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6792 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6794 OPENSSL_DECLARE_GLOBAL(int,foo);
6795 #define foo OPENSSL_GLOBAL_REF(foo)
6796 OPENSSL_DECLARE_GLOBAL(double,bar);
6797 #define bar OPENSSL_GLOBAL_REF(bar)
6799 The #defines are very important, and therefore so is including the
6800 header file everywhere where the defined globals are used.
6802 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6803 of ASN.1 items, but that structure is a bit different.
6805 The largest change is in util/mkdef.pl which has been enhanced with
6806 better and easier to understand logic to choose which symbols should
6807 go into the Windows .def files as well as a number of fixes and code
6808 cleanup (among others, algorithm keywords are now sorted
6809 lexicographically to avoid constant rewrites).
6812 *) In BN_div() keep a copy of the sign of 'num' before writing the
6813 result to 'rm' because if rm==num the value will be overwritten
6814 and produce the wrong result if 'num' is negative: this caused
6815 problems with BN_mod() and BN_nnmod().
6818 *) Function OCSP_request_verify(). This checks the signature on an
6819 OCSP request and verifies the signer certificate. The signer
6820 certificate is just checked for a generic purpose and OCSP request
6824 *) Add OCSP_check_validity() function to check the validity of OCSP
6825 responses. OCSP responses are prepared in real time and may only
6826 be a few seconds old. Simply checking that the current time lies
6827 between thisUpdate and nextUpdate max reject otherwise valid responses
6828 caused by either OCSP responder or client clock inaccuracy. Instead
6829 we allow thisUpdate and nextUpdate to fall within a certain period of
6830 the current time. The age of the response can also optionally be
6831 checked. Two new options -validity_period and -status_age added to
6835 *) If signature or public key algorithm is unrecognized print out its
6836 OID rather that just UNKNOWN.
6839 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6840 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6841 ID to be generated from the issuer certificate alone which can then be
6842 passed to OCSP_id_issuer_cmp().
6845 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6846 ASN1 modules to export functions returning ASN1_ITEM pointers
6847 instead of the ASN1_ITEM structures themselves. This adds several
6848 new macros which allow the underlying ASN1 function/structure to
6849 be accessed transparently. As a result code should not use ASN1_ITEM
6850 references directly (such as &X509_it) but instead use the relevant
6851 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6852 use of the new ASN1 code on platforms where exporting structures
6853 is problematical (for example in shared libraries) but exporting
6854 functions returning pointers to structures is not.
6857 *) Add support for overriding the generation of SSL/TLS session IDs.
6858 These callbacks can be registered either in an SSL_CTX or per SSL.
6859 The purpose of this is to allow applications to control, if they wish,
6860 the arbitrary values chosen for use as session IDs, particularly as it
6861 can be useful for session caching in multiple-server environments. A
6862 command-line switch for testing this (and any client code that wishes
6863 to use such a feature) has been added to "s_server".
6864 [Geoff Thorpe, Lutz Jaenicke]
6866 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6867 of the form '#if defined(...) || defined(...) || ...' and
6868 '#if !defined(...) && !defined(...) && ...'. This also avoids
6869 the growing number of special cases it was previously handling.
6872 *) Make all configuration macros available for application by making
6873 sure they are available in opensslconf.h, by giving them names starting
6874 with "OPENSSL_" to avoid conflicts with other packages and by making
6875 sure e_os2.h will cover all platform-specific cases together with
6877 Additionally, it is now possible to define configuration/platform-
6878 specific names (called "system identities"). In the C code, these
6879 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6880 macro with the name beginning with "OPENSSL_SYS_", which is determined
6881 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6885 *) New option -set_serial to 'req' and 'x509' this allows the serial
6886 number to use to be specified on the command line. Previously self
6887 signed certificates were hard coded with serial number 0 and the
6888 CA options of 'x509' had to use a serial number in a file which was
6892 *) New options to 'ca' utility to support V2 CRL entry extensions.
6893 Currently CRL reason, invalidity date and hold instruction are
6894 supported. Add new CRL extensions to V3 code and some new objects.
6897 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6898 disable standard block padding (aka PKCS#5 padding) in the EVP
6899 API, which was previously mandatory. This means that the data is
6900 not padded in any way and so the total length much be a multiple
6901 of the block size, otherwise an error occurs.
6904 *) Initial (incomplete) OCSP SSL support.
6907 *) New function OCSP_parse_url(). This splits up a URL into its host,
6908 port and path components: primarily to parse OCSP URLs. New -url
6909 option to ocsp utility.
6912 *) New nonce behavior. The return value of OCSP_check_nonce() now
6913 reflects the various checks performed. Applications can decide
6914 whether to tolerate certain situations such as an absent nonce
6915 in a response when one was present in a request: the ocsp application
6916 just prints out a warning. New function OCSP_add1_basic_nonce()
6917 this is to allow responders to include a nonce in a response even if
6918 the request is nonce-less.
6921 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6922 skipped when using openssl x509 multiple times on a single input file,
6923 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6926 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6927 set string type: to handle setting ASN1_TIME structures. Fix ca
6928 utility to correctly initialize revocation date of CRLs.
6931 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6932 the clients preferred ciphersuites and rather use its own preferences.
6933 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6934 Internet Explorer by ensuring unchanged hash method during stepup.
6935 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6938 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6939 to aes and add a new 'exist' option to print out symbols that don't
6943 *) Additional options to ocsp utility to allow flags to be set and
6944 additional certificates supplied.
6947 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6948 OCSP client a number of certificate to only verify the response
6952 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6953 handle the new API. Currently only ECB, CBC modes supported. Add new
6956 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6957 Encryption Standard (AES) Ciphersuites for Transport Layer
6958 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6959 not enabled by default and were not part of the "ALL" ciphersuite
6960 alias because they were not yet official; they could be
6961 explicitly requested by specifying the "AESdraft" ciphersuite
6962 group alias. In the final release of OpenSSL 0.9.7, the group
6963 alias is called "AES" and is part of "ALL".)
6964 [Ben Laurie, Steve Henson, Bodo Moeller]
6966 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6967 request to response.
6970 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6971 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6972 extract information from a certificate request. OCSP_response_create()
6973 creates a response and optionally adds a basic response structure.
6974 OCSP_basic_add1_status() adds a complete single response to a basic
6975 response and returns the OCSP_SINGLERESP structure just added (to allow
6976 extensions to be included for example). OCSP_basic_add1_cert() adds a
6977 certificate to a basic response and OCSP_basic_sign() signs a basic
6978 response with various flags. New helper functions ASN1_TIME_check()
6979 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6980 (converts ASN1_TIME to GeneralizedTime).
6983 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6984 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6985 structure from a certificate. X509_pubkey_digest() digests the public_key
6986 contents: this is used in various key identifiers.
6989 *) Make sk_sort() tolerate a NULL argument.
6990 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6992 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6993 passed by the function are trusted implicitly. If any of them signed the
6994 response then it is assumed to be valid and is not verified.
6997 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6998 to data. This was previously part of the PKCS7 ASN1 code. This
6999 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7000 [Steve Henson, reported by Kenneth R. Robinette
7001 <support@securenetterm.com>]
7003 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7004 routines: without these tracing memory leaks is very painful.
7005 Fix leaks in PKCS12 and PKCS7 routines.
7008 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7009 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7010 effectively meant GeneralizedTime would never be used. Now it
7011 is initialised to -1 but X509_time_adj() now has to check the value
7012 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7013 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7014 [Steve Henson, reported by Kenneth R. Robinette
7015 <support@securenetterm.com>]
7017 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7018 result in a zero length in the ASN1_INTEGER structure which was
7019 not consistent with the structure when d2i_ASN1_INTEGER() was used
7020 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7021 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7022 where it did not print out a minus for negative ASN1_INTEGER.
7025 *) Add summary printout to ocsp utility. The various functions which
7026 convert status values to strings have been renamed to:
7027 OCSP_response_status_str(), OCSP_cert_status_str() and
7028 OCSP_crl_reason_str() and are no longer static. New options
7029 to verify nonce values and to disable verification. OCSP response
7030 printout format cleaned up.
7033 *) Add additional OCSP certificate checks. These are those specified
7034 in RFC2560. This consists of two separate checks: the CA of the
7035 certificate being checked must either be the OCSP signer certificate
7036 or the issuer of the OCSP signer certificate. In the latter case the
7037 OCSP signer certificate must contain the OCSP signing extended key
7038 usage. This check is performed by attempting to match the OCSP
7039 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7040 in the OCSP_CERTID structures of the response.
7043 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7044 and related routines. This uses the standard OpenSSL certificate
7045 verify routines to perform initial checks (just CA validity) and
7046 to obtain the certificate chain. Then additional checks will be
7047 performed on the chain. Currently the root CA is checked to see
7048 if it is explicitly trusted for OCSP signing. This is used to set
7049 a root CA as a global signing root: that is any certificate that
7050 chains to that CA is an acceptable OCSP signing certificate.
7053 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7054 extensions from a separate configuration file.
7055 As when reading extensions from the main configuration file,
7056 the '-extensions ...' option may be used for specifying the
7058 [Massimiliano Pala <madwolf@comune.modena.it>]
7060 *) New OCSP utility. Allows OCSP requests to be generated or
7061 read. The request can be sent to a responder and the output
7062 parsed, outputed or printed in text form. Not complete yet:
7063 still needs to check the OCSP response validity.
7066 *) New subcommands for 'openssl ca':
7067 'openssl ca -status <serial>' prints the status of the cert with
7068 the given serial number (according to the index file).
7069 'openssl ca -updatedb' updates the expiry status of certificates
7071 [Massimiliano Pala <madwolf@comune.modena.it>]
7073 *) New '-newreq-nodes' command option to CA.pl. This is like
7074 '-newreq', but calls 'openssl req' with the '-nodes' option
7075 so that the resulting key is not encrypted.
7076 [Damien Miller <djm@mindrot.org>]
7078 *) New configuration for the GNU Hurd.
7079 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7081 *) Initial code to implement OCSP basic response verify. This
7082 is currently incomplete. Currently just finds the signer's
7083 certificate and verifies the signature on the response.
7086 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7087 value of OPENSSLDIR. This is available via the new '-d' option
7088 to 'openssl version', and is also included in 'openssl version -a'.
7091 *) Allowing defining memory allocation callbacks that will be given
7092 file name and line number information in additional arguments
7093 (a const char* and an int). The basic functionality remains, as
7094 well as the original possibility to just replace malloc(),
7095 realloc() and free() by functions that do not know about these
7096 additional arguments. To register and find out the current
7097 settings for extended allocation functions, the following
7098 functions are provided:
7100 CRYPTO_set_mem_ex_functions
7101 CRYPTO_set_locked_mem_ex_functions
7102 CRYPTO_get_mem_ex_functions
7103 CRYPTO_get_locked_mem_ex_functions
7105 These work the same way as CRYPTO_set_mem_functions and friends.
7106 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7107 extended allocation function is enabled.
7108 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7109 a conventional allocation function is enabled.
7110 [Richard Levitte, Bodo Moeller]
7112 *) Finish off removing the remaining LHASH function pointer casts.
7113 There should no longer be any prototype-casting required when using
7114 the LHASH abstraction, and any casts that remain are "bugs". See
7115 the callback types and macros at the head of lhash.h for details
7116 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7119 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7120 If /dev/[u]random devices are not available or do not return enough
7121 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7123 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7124 /etc/entropy will be queried once each in this sequence, quering stops
7125 when enough entropy was collected without querying more sockets.
7128 *) Change the Unix RAND_poll() variant to be able to poll several
7129 random devices, as specified by DEVRANDOM, until a sufficient amount
7130 of data has been collected. We spend at most 10 ms on each file
7131 (select timeout) and read in non-blocking mode. DEVRANDOM now
7132 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7133 (previously it was just the string "/dev/urandom"), so on typical
7134 platforms the 10 ms delay will never occur.
7135 Also separate out the Unix variant to its own file, rand_unix.c.
7136 For VMS, there's a currently-empty rand_vms.c.
7139 *) Move OCSP client related routines to ocsp_cl.c. These
7140 provide utility functions which an application needing
7141 to issue a request to an OCSP responder and analyse the
7142 response will typically need: as opposed to those which an
7143 OCSP responder itself would need which will be added later.
7145 OCSP_request_sign() signs an OCSP request with an API similar
7146 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7147 response. OCSP_response_get1_basic() extracts basic response
7148 from response. OCSP_resp_find_status(): finds and extracts status
7149 information from an OCSP_CERTID structure (which will be created
7150 when the request structure is built). These are built from lower
7151 level functions which work on OCSP_SINGLERESP structures but
7152 wont normally be used unless the application wishes to examine
7153 extensions in the OCSP response for example.
7155 Replace nonce routines with a pair of functions.
7156 OCSP_request_add1_nonce() adds a nonce value and optionally
7157 generates a random value. OCSP_check_nonce() checks the
7158 validity of the nonce in an OCSP response.
7161 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7162 This doesn't copy the supplied OCSP_CERTID and avoids the
7163 need to free up the newly created id. Change return type
7164 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7165 This can then be used to add extensions to the request.
7166 Deleted OCSP_request_new(), since most of its functionality
7167 is now in OCSP_REQUEST_new() (and the case insensitive name
7168 clash) apart from the ability to set the request name which
7169 will be added elsewhere.
7172 *) Update OCSP API. Remove obsolete extensions argument from
7173 various functions. Extensions are now handled using the new
7174 OCSP extension code. New simple OCSP HTTP function which
7175 can be used to send requests and parse the response.
7178 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7179 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7180 uses the special reorder version of SET OF to sort the attributes
7181 and reorder them to match the encoded order. This resolves a long
7182 standing problem: a verify on a PKCS7 structure just after signing
7183 it used to fail because the attribute order did not match the
7184 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7185 it uses the received order. This is necessary to tolerate some broken
7186 software that does not order SET OF. This is handled by encoding
7187 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7188 to produce the required SET OF.
7191 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7192 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7193 files to get correct declarations of the ASN.1 item variables.
7196 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7197 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7198 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7199 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7200 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7201 ASN1_ITEM and no wrapper functions.
7204 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7205 replace the old function pointer based I/O routines. Change most of
7206 the *_d2i_bio() and *_d2i_fp() functions to use these.
7209 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7210 lines, recognice more "algorithms" that can be deselected, and make
7211 it complain about algorithm deselection that isn't recognised.
7214 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7215 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7216 to use new functions. Add NO_ASN1_OLD which can be set to remove
7217 some old style ASN1 functions: this can be used to determine if old
7218 code will still work when these eventually go away.
7221 *) New extension functions for OCSP structures, these follow the
7222 same conventions as certificates and CRLs.
7225 *) New function X509V3_add1_i2d(). This automatically encodes and
7226 adds an extension. Its behaviour can be customised with various
7227 flags to append, replace or delete. Various wrappers added for
7228 certifcates and CRLs.
7231 *) Fix to avoid calling the underlying ASN1 print routine when
7232 an extension cannot be parsed. Correct a typo in the
7233 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7236 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7237 entries for variables.
7240 *) Add functionality to apps/openssl.c for detecting locking
7241 problems: As the program is single-threaded, all we have
7242 to do is register a locking callback using an array for
7243 storing which locks are currently held by the program.
7246 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7247 SSL_get_ex_data_X509_STORE_idx(), which is used in
7248 ssl_verify_cert_chain() and thus can be called at any time
7249 during TLS/SSL handshakes so that thread-safety is essential.
7250 Unfortunately, the ex_data design is not at all suited
7251 for multi-threaded use, so it probably should be abolished.
7254 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7255 [Broadcom, tweaked and integrated by Geoff Thorpe]
7257 *) Move common extension printing code to new function
7258 X509V3_print_extensions(). Reorganise OCSP print routines and
7259 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7262 *) New function X509_signature_print() to remove duplication in some
7266 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7267 set (this was treated exactly the same as SET OF previously). This
7268 is used to reorder the STACK representing the structure to match the
7269 encoding. This will be used to get round a problem where a PKCS7
7270 structure which was signed could not be verified because the STACK
7271 order did not reflect the encoded order.
7274 *) Reimplement the OCSP ASN1 module using the new code.
7277 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7278 for its ASN1 operations. The old style function pointers still exist
7279 for now but they will eventually go away.
7282 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7283 completely replaces the old ASN1 functionality with a table driven
7284 encoder and decoder which interprets an ASN1_ITEM structure describing
7285 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7286 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7287 has also been converted to the new form.
7290 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7291 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7292 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7293 for negative moduli.
7296 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7297 of not touching the result's sign bit.
7300 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7304 *) Changed the LHASH code to use prototypes for callbacks, and created
7305 macros to declare and implement thin (optionally static) functions
7306 that provide type-safety and avoid function pointer casting for the
7307 type-specific callbacks.
7310 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7312 [Veers Staats <staatsvr@asc.hpc.mil>,
7313 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7315 *) Reformat the FAQ so the different questions and answers can be divided
7316 in sections depending on the subject.
7319 *) Have the zlib compression code load ZLIB.DLL dynamically under
7323 *) New function BN_mod_sqrt for computing square roots modulo a prime
7324 (using the probabilistic Tonelli-Shanks algorithm unless
7325 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7326 be handled deterministically).
7327 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7329 *) Make BN_mod_inverse faster by explicitly handling small quotients
7330 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7331 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7334 *) New function BN_kronecker.
7337 *) Fix BN_gcd so that it works on negative inputs; the result is
7338 positive unless both parameters are zero.
7339 Previously something reasonably close to an infinite loop was
7340 possible because numbers could be growing instead of shrinking
7341 in the implementation of Euclid's algorithm.
7344 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7345 sign of the number in question.
7347 Fix BN_is_word(a,w) to work correctly for w == 0.
7349 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7350 because its test if the absolute value of 'a' equals 'w'.
7351 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7352 it exists mostly for use in the implementations of BN_is_zero(),
7353 BN_is_one(), and BN_is_word().
7356 *) New function BN_swap.
7359 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7360 the exponentiation functions are more likely to produce reasonable
7361 results on negative inputs.
7364 *) Change BN_mod_mul so that the result is always non-negative.
7365 Previously, it could be negative if one of the factors was negative;
7366 I don't think anyone really wanted that behaviour.
7369 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7370 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7371 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7372 and add new functions:
7381 BN_mod_lshift1_quick
7385 These functions always generate non-negative results.
7387 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7388 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7390 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7391 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7392 be reduced modulo m.
7393 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7396 The following entry accidentally appeared in the CHANGES file
7397 distributed with OpenSSL 0.9.7. The modifications described in
7398 it do *not* apply to OpenSSL 0.9.7.
7400 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7401 was actually never needed) and in BN_mul(). The removal in BN_mul()
7402 required a small change in bn_mul_part_recursive() and the addition
7403 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7404 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7405 bn_sub_words() and bn_add_words() except they take arrays with
7410 *) In 'openssl passwd', verify passwords read from the terminal
7411 unless the '-salt' option is used (which usually means that
7412 verification would just waste user's time since the resulting
7413 hash is going to be compared with some given password hash)
7414 or the new '-noverify' option is used.
7416 This is an incompatible change, but it does not affect
7417 non-interactive use of 'openssl passwd' (passwords on the command
7418 line, '-stdin' option, '-in ...' option) and thus should not
7422 *) Remove all references to RSAref, since there's no more need for it.
7425 *) Make DSO load along a path given through an environment variable
7426 (SHLIB_PATH) with shl_load().
7429 *) Constify the ENGINE code as a result of BIGNUM constification.
7430 Also constify the RSA code and most things related to it. In a
7431 few places, most notable in the depth of the ASN.1 code, ugly
7432 casts back to non-const were required (to be solved at a later
7436 *) Make it so the openssl application has all engines loaded by default.
7439 *) Constify the BIGNUM routines a little more.
7442 *) Add the following functions:
7444 ENGINE_load_cswift()
7446 ENGINE_load_atalla()
7448 ENGINE_load_builtin_engines()
7450 That way, an application can itself choose if external engines that
7451 are built-in in OpenSSL shall ever be used or not. The benefit is
7452 that applications won't have to be linked with libdl or other dso
7453 libraries unless it's really needed.
7455 Changed 'openssl engine' to load all engines on demand.
7456 Changed the engine header files to avoid the duplication of some
7457 declarations (they differed!).
7460 *) 'openssl engine' can now list capabilities.
7463 *) Better error reporting in 'openssl engine'.
7466 *) Never call load_dh_param(NULL) in s_server.
7469 *) Add engine application. It can currently list engines by name and
7470 identity, and test if they are actually available.
7473 *) Improve RPM specification file by forcing symbolic linking and making
7474 sure the installed documentation is also owned by root.root.
7475 [Damien Miller <djm@mindrot.org>]
7477 *) Give the OpenSSL applications more possibilities to make use of
7478 keys (public as well as private) handled by engines.
7481 *) Add OCSP code that comes from CertCo.
7484 *) Add VMS support for the Rijndael code.
7487 *) Added untested support for Nuron crypto accelerator.
7490 *) Add support for external cryptographic devices. This code was
7491 previously distributed separately as the "engine" branch.
7492 [Geoff Thorpe, Richard Levitte]
7494 *) Rework the filename-translation in the DSO code. It is now possible to
7495 have far greater control over how a "name" is turned into a filename
7496 depending on the operating environment and any oddities about the
7497 different shared library filenames on each system.
7500 *) Support threads on FreeBSD-elf in Configure.
7503 *) Fix for SHA1 assembly problem with MASM: it produces
7504 warnings about corrupt line number information when assembling
7505 with debugging information. This is caused by the overlapping
7507 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7510 NCONF_get_number() has no error checking at all. As a replacement,
7511 NCONF_get_number_e() is defined (_e for "error checking") and is
7512 promoted strongly. The old NCONF_get_number is kept around for
7513 binary backward compatibility.
7514 Make it possible for methods to load from something other than a BIO,
7515 by providing a function pointer that is given a name instead of a BIO.
7516 For example, this could be used to load configuration data from an
7520 *) Fix for non blocking accept BIOs. Added new I/O special reason
7521 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7522 with non blocking I/O was not possible because no retry code was
7523 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7527 *) Added the beginnings of Rijndael support.
7530 *) Fix for bug in DirectoryString mask setting. Add support for
7531 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7532 to allow certificate printing to more controllable, additional
7533 'certopt' option to 'x509' to allow new printing options to be
7537 *) Clean old EAY MD5 hack from e_os.h.
7540 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7542 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7543 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7544 [Joe Orton, Steve Henson]
7546 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7548 *) Fix additional bug revealed by the NISCC test suite:
7550 Stop bug triggering large recursion when presented with
7551 certain ASN.1 tags (CVE-2003-0851)
7554 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7556 *) Fix various bugs revealed by running the NISCC test suite:
7558 Stop out of bounds reads in the ASN1 code when presented with
7559 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7561 If verify callback ignores invalid public key errors don't try to check
7562 certificate signature with the NULL public key.
7566 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7567 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7571 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7572 extra data after the compression methods not only for TLS 1.0
7573 but also for SSL 3.0 (as required by the specification).
7574 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7576 *) Change X509_certificate_type() to mark the key as exported/exportable
7577 when it's 512 *bits* long, not 512 bytes.
7580 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7582 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7583 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7584 a protocol version number mismatch like a decryption error
7585 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7588 *) Turn on RSA blinding by default in the default implementation
7589 to avoid a timing attack. Applications that don't want it can call
7590 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7591 They would be ill-advised to do so in most cases.
7592 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7594 *) Change RSA blinding code so that it works when the PRNG is not
7595 seeded (in this case, the secret RSA exponent is abused as
7596 an unpredictable seed -- if it is not unpredictable, there
7597 is no point in blinding anyway). Make RSA blinding thread-safe
7598 by remembering the creator's thread ID in rsa->blinding and
7599 having all other threads use local one-time blinding factors
7600 (this requires more computation than sharing rsa->blinding, but
7601 avoids excessive locking; and if an RSA object is not shared
7602 between threads, blinding will still be very fast).
7605 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7607 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7608 via timing by performing a MAC computation even if incorrrect
7609 block cipher padding has been found. This is a countermeasure
7610 against active attacks where the attacker has to distinguish
7611 between bad padding and a MAC verification error. (CVE-2003-0078)
7613 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7614 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7615 Martin Vuagnoux (EPFL, Ilion)]
7617 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7619 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7620 memory from it's contents. This is done with a counter that will
7621 place alternating values in each byte. This can be used to solve
7622 two issues: 1) the removal of calls to memset() by highly optimizing
7623 compilers, and 2) cleansing with other values than 0, since those can
7624 be read through on certain media, for example a swap space on disk.
7627 *) Bugfix: client side session caching did not work with external caching,
7628 because the session->cipher setting was not restored when reloading
7629 from the external cache. This problem was masked, when
7630 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7631 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7634 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7635 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7636 [Zeev Lieber <zeev-l@yahoo.com>]
7638 *) Undo an undocumented change introduced in 0.9.6e which caused
7639 repeated calls to OpenSSL_add_all_ciphers() and
7640 OpenSSL_add_all_digests() to be ignored, even after calling
7644 *) Change the default configuration reader to deal with last line not
7645 being properly terminated.
7648 *) Change X509_NAME_cmp() so it applies the special rules on handling
7649 DN values that are of type PrintableString, as well as RDNs of type
7650 emailAddress where the value has the type ia5String.
7651 [stefank@valicert.com via Richard Levitte]
7653 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7654 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7655 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7656 the bitwise-OR of the two for use by the majority of applications
7657 wanting this behaviour, and update the docs. The documented
7658 behaviour and actual behaviour were inconsistent and had been
7659 changing anyway, so this is more a bug-fix than a behavioural
7661 [Geoff Thorpe, diagnosed by Nadav Har'El]
7663 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7664 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7667 *) Fix initialization code race conditions in
7668 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7669 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7670 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7671 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7672 ssl2_get_cipher_by_char(),
7673 ssl3_get_cipher_by_char().
7674 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7676 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7677 the cached sessions are flushed, as the remove_cb() might use ex_data
7678 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7679 (see [openssl.org #212]).
7680 [Geoff Thorpe, Lutz Jaenicke]
7682 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7683 length, instead of the encoding length to d2i_ASN1_OBJECT.
7686 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7688 *) [In 0.9.6g-engine release:]
7689 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7690 [Lynn Gazis <lgazis@rainbow.com>]
7692 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7694 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7695 and get fix the header length calculation.
7696 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7697 Alon Kantor <alonk@checkpoint.com> (and others),
7700 *) Use proper error handling instead of 'assertions' in buffer
7701 overflow checks added in 0.9.6e. This prevents DoS (the
7702 assertions could call abort()).
7703 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7705 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7707 *) Add various sanity checks to asn1_get_length() to reject
7708 the ASN1 length bytes if they exceed sizeof(long), will appear
7709 negative or the content length exceeds the length of the
7711 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7713 *) Fix cipher selection routines: ciphers without encryption had no flags
7714 for the cipher strength set and where therefore not handled correctly
7715 by the selection routines (PR #130).
7718 *) Fix EVP_dsa_sha macro.
7722 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7723 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7724 that was added in OpenSSL 0.9.6d.
7726 As the countermeasure turned out to be incompatible with some
7727 broken SSL implementations, the new option is part of SSL_OP_ALL.
7728 SSL_OP_ALL is usually employed when compatibility with weird SSL
7729 implementations is desired (e.g. '-bugs' option to 's_client' and
7730 's_server'), so the new option is automatically set in many
7734 *) Changes in security patch:
7736 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7737 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7738 Air Force Materiel Command, USAF, under agreement number
7741 *) Add various sanity checks to asn1_get_length() to reject
7742 the ASN1 length bytes if they exceed sizeof(long), will appear
7743 negative or the content length exceeds the length of the
7744 supplied buffer. (CVE-2002-0659)
7745 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7747 *) Assertions for various potential buffer overflows, not known to
7749 [Ben Laurie (CHATS)]
7751 *) Various temporary buffers to hold ASCII versions of integers were
7752 too small for 64 bit platforms. (CVE-2002-0655)
7753 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7755 *) Remote buffer overflow in SSL3 protocol - an attacker could
7756 supply an oversized session ID to a client. (CVE-2002-0656)
7757 [Ben Laurie (CHATS)]
7759 *) Remote buffer overflow in SSL2 protocol - an attacker could
7760 supply an oversized client master key. (CVE-2002-0656)
7761 [Ben Laurie (CHATS)]
7763 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7765 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7766 encoded as NULL) with id-dsa-with-sha1.
7767 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7769 *) Check various X509_...() return values in apps/req.c.
7770 [Nils Larsch <nla@trustcenter.de>]
7772 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7773 an end-of-file condition would erronously be flagged, when the CRLF
7774 was just at the end of a processed block. The bug was discovered when
7775 processing data through a buffering memory BIO handing the data to a
7776 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7777 <ptsekov@syntrex.com> and Nedelcho Stanev.
7780 *) Implement a countermeasure against a vulnerability recently found
7781 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7782 before application data chunks to avoid the use of known IVs
7783 with data potentially chosen by the attacker.
7786 *) Fix length checks in ssl3_get_client_hello().
7789 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7790 to prevent ssl3_read_internal() from incorrectly assuming that
7791 ssl3_read_bytes() found application data while handshake
7792 processing was enabled when in fact s->s3->in_read_app_data was
7793 merely automatically cleared during the initial handshake.
7794 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7796 *) Fix object definitions for Private and Enterprise: they were not
7797 recognized in their shortname (=lowercase) representation. Extend
7798 obj_dat.pl to issue an error when using undefined keywords instead
7799 of silently ignoring the problem (Svenning Sorensen
7800 <sss@sss.dnsalias.net>).
7803 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7804 generators, i.e. generators other than 2 and 5. (Previously, the
7805 code did not properly initialise the 'add' and 'rem' values to
7806 BN_generate_prime().)
7808 In the new general case, we do not insist that 'generator' is
7809 actually a primitive root: This requirement is rather pointless;
7810 a generator of the order-q subgroup is just as good, if not
7814 *) Map new X509 verification errors to alerts. Discovered and submitted by
7815 Tom Wu <tom@arcot.com>.
7818 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7819 returning non-zero before the data has been completely received
7820 when using non-blocking I/O.
7821 [Bodo Moeller; problem pointed out by John Hughes]
7823 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7824 [Ben Laurie, Lutz Jaenicke]
7826 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7827 Yoram Zahavi <YoramZ@gilian.com>).
7830 *) Add information about CygWin 1.3 and on, and preserve proper
7831 configuration for the versions before that.
7832 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7834 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7835 check whether we deal with a copy of a session and do not delete from
7836 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7837 <izhar@checkpoint.com>.
7840 *) Do not store session data into the internal session cache, if it
7841 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7842 flag is set). Proposed by Aslam <aslam@funk.com>.
7845 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7849 *) [In 0.9.6d-engine release:]
7850 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7851 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7853 *) Add the configuration target linux-s390x.
7854 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7856 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7857 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7858 variable as an indication that a ClientHello message has been
7859 received. As the flag value will be lost between multiple
7860 invocations of ssl3_accept when using non-blocking I/O, the
7861 function may not be aware that a handshake has actually taken
7862 place, thus preventing a new session from being added to the
7865 To avoid this problem, we now set s->new_session to 2 instead of
7866 using a local variable.
7867 [Lutz Jaenicke, Bodo Moeller]
7869 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7870 if the SSL_R_LENGTH_MISMATCH error is detected.
7871 [Geoff Thorpe, Bodo Moeller]
7873 *) New 'shared_ldflag' column in Configure platform table.
7876 *) Fix EVP_CIPHER_mode macro.
7877 ["Dan S. Camper" <dan@bti.net>]
7879 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7880 type, we must throw them away by setting rr->length to 0.
7881 [D P Chang <dpc@qualys.com>]
7883 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7885 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7886 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7887 worked incorrectly for those cases where range = 10..._2 and
7888 3*range is two bits longer than range.)
7891 *) Only add signing time to PKCS7 structures if it is not already
7895 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7896 OBJ_ld_ce should be OBJ_id_ce.
7897 Also some ip-pda OIDs in crypto/objects/objects.txt were
7898 incorrect (cf. RFC 3039).
7899 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7901 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7902 returns early because it has nothing to do.
7903 [Andy Schneider <andy.schneider@bjss.co.uk>]
7905 *) [In 0.9.6c-engine release:]
7906 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7907 [Andy Schneider <andy.schneider@bjss.co.uk>]
7909 *) [In 0.9.6c-engine release:]
7910 Add support for Cryptographic Appliance's keyserver technology.
7911 (Use engine 'keyclient')
7912 [Cryptographic Appliances and Geoff Thorpe]
7914 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7915 is called via tools/c89.sh because arguments have to be
7916 rearranged (all '-L' options must appear before the first object
7918 [Richard Shapiro <rshapiro@abinitio.com>]
7920 *) [In 0.9.6c-engine release:]
7921 Add support for Broadcom crypto accelerator cards, backported
7923 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7925 *) [In 0.9.6c-engine release:]
7926 Add support for SureWare crypto accelerator cards from
7927 Baltimore Technologies. (Use engine 'sureware')
7928 [Baltimore Technologies and Mark Cox]
7930 *) [In 0.9.6c-engine release:]
7931 Add support for crypto accelerator cards from Accelerated
7932 Encryption Processing, www.aep.ie. (Use engine 'aep')
7933 [AEP Inc. and Mark Cox]
7935 *) Add a configuration entry for gcc on UnixWare.
7936 [Gary Benson <gbenson@redhat.com>]
7938 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7939 messages are stored in a single piece (fixed-length part and
7940 variable-length part combined) and fix various bugs found on the way.
7943 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7944 instead. BIO_gethostbyname() does not know what timeouts are
7945 appropriate, so entries would stay in cache even when they have
7947 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7949 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7950 faced with a pathologically small ClientHello fragment that does
7951 not contain client_version: Instead of aborting with an error,
7952 simply choose the highest available protocol version (i.e.,
7953 TLS 1.0 unless it is disabled). In practice, ClientHello
7954 messages are never sent like this, but this change gives us
7955 strictly correct behaviour at least for TLS.
7958 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7959 never resets s->method to s->ctx->method when called from within
7960 one of the SSL handshake functions.
7961 [Bodo Moeller; problem pointed out by Niko Baric]
7963 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7964 (sent using the client's version number) if client_version is
7965 smaller than the protocol version in use. Also change
7966 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7967 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7968 the client will at least see that alert.
7971 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7975 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7976 client receives HelloRequest while in a handshake.
7977 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7979 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7980 should end in 'break', not 'goto end' which circumvents various
7981 cleanups done in state SSL_ST_OK. But session related stuff
7982 must be disabled for SSL_ST_OK in the case that we just sent a
7985 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7986 before just sending a HelloRequest.
7987 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7989 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7990 reveal whether illegal block cipher padding was found or a MAC
7991 verification error occurred. (Neither SSLerr() codes nor alerts
7992 are directly visible to potential attackers, but the information
7993 may leak via logfiles.)
7995 Similar changes are not required for the SSL 2.0 implementation
7996 because the number of padding bytes is sent in clear for SSL 2.0,
7997 and the extra bytes are just ignored. However ssl/s2_pkt.c
7998 failed to verify that the purported number of padding bytes is in
8002 *) Add OpenUNIX-8 support including shared libraries
8003 (Boyd Lynn Gerber <gerberb@zenez.com>).
8006 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8007 'wristwatch attack' using huge encoding parameters (cf.
8008 James H. Manger's CRYPTO 2001 paper). Note that the
8009 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8010 encoding parameters and hence was not vulnerable.
8013 *) BN_sqr() bug fix.
8014 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8016 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8017 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8018 followed by modular reduction.
8019 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8021 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8022 equivalent based on BN_pseudo_rand() instead of BN_rand().
8025 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8026 This function was broken, as the check for a new client hello message
8027 to handle SGC did not allow these large messages.
8028 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8031 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8034 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8035 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8038 *) Rework the configuration and shared library support for Tru64 Unix.
8039 The configuration part makes use of modern compiler features and
8040 still retains old compiler behavior for those that run older versions
8041 of the OS. The shared library support part includes a variant that
8042 uses the RPATH feature, and is available through the special
8043 configuration target "alpha-cc-rpath", which will never be selected
8045 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8047 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8048 with the same message size as in ssl3_get_certificate_request().
8049 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8050 messages might inadvertently be reject as too long.
8051 [Petr Lampa <lampa@fee.vutbr.cz>]
8053 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8056 *) Modified SSL library such that the verify_callback that has been set
8057 specificly for an SSL object with SSL_set_verify() is actually being
8058 used. Before the change, a verify_callback set with this function was
8059 ignored and the verify_callback() set in the SSL_CTX at the time of
8060 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8061 to allow the necessary settings.
8064 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8065 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8066 done automatically (in contradiction to the requirements of the C
8067 standard). This made problems when used from OpenSSH.
8070 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8071 dh->length and always used
8073 BN_rand_range(priv_key, dh->p).
8075 BN_rand_range() is not necessary for Diffie-Hellman, and this
8076 specific range makes Diffie-Hellman unnecessarily inefficient if
8077 dh->length (recommended exponent length) is much smaller than the
8078 length of dh->p. We could use BN_rand_range() if the order of
8079 the subgroup was stored in the DH structure, but we only have
8084 BN_rand(priv_key, l, ...)
8086 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8092 RSA_eay_public_encrypt
8093 RSA_eay_private_decrypt
8094 RSA_eay_private_encrypt (signing)
8095 RSA_eay_public_decrypt (signature verification)
8097 (default implementations for RSA_public_encrypt,
8098 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8099 always reject numbers >= n.
8102 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8103 to synchronize access to 'locking_thread'. This is necessary on
8104 systems where access to 'locking_thread' (an 'unsigned long'
8105 variable) is not atomic.
8108 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8109 *before* setting the 'crypto_lock_rand' flag. The previous code had
8110 a race condition if 0 is a valid thread ID.
8111 [Travis Vitek <vitek@roguewave.com>]
8113 *) Add support for shared libraries under Irix.
8114 [Albert Chin-A-Young <china@thewrittenword.com>]
8116 *) Add configuration option to build on Linux on both big-endian and
8118 [Ralf Baechle <ralf@uni-koblenz.de>]
8120 *) Add the possibility to create shared libraries on HP-UX.
8123 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8125 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8126 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8127 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8128 PRNG state recovery was possible based on the output of
8129 one PRNG request appropriately sized to gain knowledge on
8130 'md' followed by enough consecutive 1-byte PRNG requests
8131 to traverse all of 'state'.
8133 1. When updating 'md_local' (the current thread's copy of 'md')
8134 during PRNG output generation, hash all of the previous
8135 'md_local' value, not just the half used for PRNG output.
8137 2. Make the number of bytes from 'state' included into the hash
8138 independent from the number of PRNG bytes requested.
8140 The first measure alone would be sufficient to avoid
8141 Markku-Juhani's attack. (Actually it had never occurred
8142 to me that the half of 'md_local' used for chaining was the
8143 half from which PRNG output bytes were taken -- I had always
8144 assumed that the secret half would be used.) The second
8145 measure makes sure that additional data from 'state' is never
8146 mixed into 'md_local' in small portions; this heuristically
8147 further strengthens the PRNG.
8150 *) Fix crypto/bn/asm/mips3.s.
8153 *) When only the key is given to "enc", the IV is undefined. Print out
8154 an error message in this case.
8157 *) Handle special case when X509_NAME is empty in X509 printing routines.
8160 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8161 positive and less than q.
8164 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8165 used: it isn't thread safe and the add_lock_callback should handle
8167 [Paul Rose <Paul.Rose@bridge.com>]
8169 *) Verify that incoming data obeys the block size in
8170 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8174 [Ulf Möller, Bodo Möller]
8176 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8177 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8178 when fixing the server behaviour for backwards-compatible 'client
8179 hello' messages. (Note that the attack is impractical against
8180 SSL 3.0 and TLS 1.0 anyway because length and version checking
8181 means that the probability of guessing a valid ciphertext is
8182 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8185 Before 0.9.5, the countermeasure (hide the error by generating a
8186 random 'decryption result') did not work properly because
8187 ERR_clear_error() was missing, meaning that SSL_get_error() would
8188 detect the supposedly ignored error.
8190 Both problems are now fixed.
8193 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8194 (previously it was 1024).
8197 *) Fix for compatibility mode trust settings: ignore trust settings
8198 unless some valid trust or reject settings are present.
8201 *) Fix for blowfish EVP: its a variable length cipher.
8204 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8205 parameters in DSA public key structures and return an error in the
8206 DSA routines if parameters are absent.
8209 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8210 in the current directory if neither $RANDFILE nor $HOME was set.
8211 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8212 caused some confusion to Windows users who haven't defined $HOME.
8213 Thus RAND_file_name() is changed again: e_os.h can define a
8214 DEFAULT_HOME, which will be used if $HOME is not set.
8215 For Windows, we use "C:"; on other platforms, we still require
8216 environment variables.
8218 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8219 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8220 having multiple threads call RAND_poll() concurrently.
8223 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8224 combination of a flag and a thread ID variable.
8225 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8226 flag), *other* threads can enter ssleay_add_bytes without obeying
8227 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8228 that they do not hold after the first thread unsets add_do_not_lock).
8231 *) Change bctest again: '-x' expressions are not available in all
8235 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8237 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8238 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8240 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8241 the default extension for executables, if any. Also, make the perl
8242 scripts that use symlink() to test if it really exists and use "cp"
8243 if it doesn't. All this made OpenSSL compilable and installable in
8247 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8248 If SEQUENCE is length is indefinite just set c->slen to the total
8249 amount of data available.
8250 [Steve Henson, reported by shige@FreeBSD.org]
8251 [This change does not apply to 0.9.7.]
8253 *) Change bctest to avoid here-documents inside command substitution
8254 (workaround for FreeBSD /bin/sh bug).
8255 For compatibility with Ultrix, avoid shell functions (introduced
8256 in the bctest version that searches along $PATH).
8259 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8260 with des_encrypt() defined on some operating systems, like Solaris
8264 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8265 On the Importance of Eliminating Errors in Cryptographic
8266 Computations, J. Cryptology 14 (2001) 2, 101-119,
8267 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8270 *) MIPS assembler BIGNUM division bug fix.
8273 *) Disabled incorrect Alpha assembler code.
8276 *) Fix PKCS#7 decode routines so they correctly update the length
8277 after reading an EOC for the EXPLICIT tag.
8279 [This change does not apply to 0.9.7.]
8281 *) Fix bug in PKCS#12 key generation routines. This was triggered
8282 if a 3DES key was generated with a 0 initial byte. Include
8283 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8284 (but broken) behaviour.
8287 *) Enhance bctest to search for a working bc along $PATH and print
8289 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8291 *) Fix memory leaks in err.c: free err_data string if necessary;
8292 don't write to the wrong index in ERR_set_error_data.
8295 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8299 *) Replace rdtsc with _emit statements for VC++ version 5.
8300 [Jeremy Cooper <jeremy@baymoo.org>]
8302 *) Make it possible to reuse SSLv2 sessions.
8305 *) In copy_email() check for >= 0 as a return value for
8306 X509_NAME_get_index_by_NID() since 0 is a valid index.
8307 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8309 *) Avoid coredump with unsupported or invalid public keys by checking if
8310 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8311 PKCS7_verify() fails with non detached data.
8314 *) Don't use getenv in library functions when run as setuid/setgid.
8315 New function OPENSSL_issetugid().
8318 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8319 due to incorrect handling of multi-threading:
8321 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8323 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8325 3. Count how many times MemCheck_off() has been called so that
8326 nested use can be treated correctly. This also avoids
8327 inband-signalling in the previous code (which relied on the
8328 assumption that thread ID 0 is impossible).
8331 *) Add "-rand" option also to s_client and s_server.
8334 *) Fix CPU detection on Irix 6.x.
8335 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8336 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8338 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8341 [This change does not apply to 0.9.7.]
8343 *) Use the cached encoding of an X509_NAME structure rather than
8344 copying it. This is apparently the reason for the libsafe "errors"
8345 but the code is actually correct.
8348 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8349 Bleichenbacher's DSA attack.
8350 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8351 to be set and top=0 forces the highest bit to be set; top=-1 is new
8352 and leaves the highest bit random.
8353 [Ulf Moeller, Bodo Moeller]
8355 *) In the NCONF_...-based implementations for CONF_... queries
8356 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8357 a temporary CONF structure with the data component set to NULL
8358 (which gives segmentation faults in lh_retrieve).
8359 Instead, use NULL for the CONF pointer in CONF_get_string and
8360 CONF_get_number (which may use environment variables) and directly
8361 return NULL from CONF_get_section.
8364 *) Fix potential buffer overrun for EBCDIC.
8367 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8368 keyUsage if basicConstraints absent for a CA.
8371 *) Make SMIME_write_PKCS7() write mail header values with a format that
8372 is more generally accepted (no spaces before the semicolon), since
8373 some programs can't parse those values properly otherwise. Also make
8374 sure BIO's that break lines after each write do not create invalid
8378 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8379 macros previously used would not encode an empty SEQUENCE OF
8380 and break the signature.
8382 [This change does not apply to 0.9.7.]
8384 *) Zero the premaster secret after deriving the master secret in
8388 *) Add some EVP_add_digest_alias registrations (as found in
8389 OpenSSL_add_all_digests()) to SSL_library_init()
8390 aka OpenSSL_add_ssl_algorithms(). This provides improved
8391 compatibility with peers using X.509 certificates
8392 with unconventional AlgorithmIdentifier OIDs.
8395 *) Fix for Irix with NO_ASM.
8396 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8398 *) ./config script fixes.
8399 [Ulf Moeller, Richard Levitte]
8401 *) Fix 'openssl passwd -1'.
8404 *) Change PKCS12_key_gen_asc() so it can cope with non null
8405 terminated strings whose length is passed in the passlen
8406 parameter, for example from PEM callbacks. This was done
8407 by adding an extra length parameter to asc2uni().
8408 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8410 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8411 call failed, free the DSA structure.
8414 *) Fix to uni2asc() to cope with zero length Unicode strings.
8415 These are present in some PKCS#12 files.
8418 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8419 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8420 when writing a 32767 byte record.
8421 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8423 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8424 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8426 (RSA objects have a reference count access to which is protected
8427 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8428 so they are meant to be shared between threads.)
8429 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8430 "Reddie, Steven" <Steven.Reddie@ca.com>]
8432 *) Fix a deadlock in CRYPTO_mem_leaks().
8435 *) Use better test patterns in bntest.
8438 *) rand_win.c fix for Borland C.
8441 *) BN_rshift bugfix for n == 0.
8444 *) Add a 'bctest' script that checks for some known 'bc' bugs
8445 so that 'make test' does not abort just because 'bc' is broken.
8448 *) Store verify_result within SSL_SESSION also for client side to
8449 avoid potential security hole. (Re-used sessions on the client side
8450 always resulted in verify_result==X509_V_OK, not using the original
8451 result of the server certificate verification.)
8454 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8455 SSL3_RT_APPLICATION_DATA, return 0.
8456 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8460 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8461 releases, have been re-implemented by renaming the previous
8462 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8463 and ssl3_read_internal, respectively, and adding 'peek' parameters
8464 to them. The new ssl[23]_{read,peek} functions are calls to
8465 ssl[23]_read_internal with the 'peek' flag set appropriately.
8466 A 'peek' parameter has also been added to ssl3_read_bytes, which
8467 does the actual work for ssl3_read_internal.
8470 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8471 the method-specific "init()" handler. Also clean up ex_data after
8472 calling the method-specific "finish()" handler. Previously, this was
8473 happening the other way round.
8476 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8477 The previous value, 12, was not always sufficient for BN_mod_exp().
8480 *) Make sure that shared libraries get the internal name engine with
8481 the full version number and not just 0. This should mark the
8482 shared libraries as not backward compatible. Of course, this should
8483 be changed again when we can guarantee backward binary compatibility.
8486 *) Fix typo in get_cert_by_subject() in by_dir.c
8487 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8489 *) Rework the system to generate shared libraries:
8491 - Make note of the expected extension for the shared libraries and
8492 if there is a need for symbolic links from for example libcrypto.so.0
8493 to libcrypto.so.0.9.7. There is extended info in Configure for
8496 - Make as few rebuilds of the shared libraries as possible.
8498 - Still avoid linking the OpenSSL programs with the shared libraries.
8500 - When installing, install the shared libraries separately from the
8504 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8506 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8507 and not in SSL_clear because the latter is also used by the
8508 accept/connect functions; previously, the settings made by
8509 SSL_set_read_ahead would be lost during the handshake.
8510 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8512 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8513 Previously, it would create entries for disabled algorithms no
8517 *) Added several new manual pages for SSL_* function.
8520 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8522 *) In ssl23_get_client_hello, generate an error message when faced
8523 with an initial SSL 3.0/TLS record that is too small to contain the
8524 first two bytes of the ClientHello message, i.e. client_version.
8525 (Note that this is a pathologic case that probably has never happened
8526 in real life.) The previous approach was to use the version number
8527 from the record header as a substitute; but our protocol choice
8528 should not depend on that one because it is not authenticated
8529 by the Finished messages.
8532 *) More robust randomness gathering functions for Windows.
8533 [Jeffrey Altman <jaltman@columbia.edu>]
8535 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8536 not set then we don't setup the error code for issuer check errors
8537 to avoid possibly overwriting other errors which the callback does
8538 handle. If an application does set the flag then we assume it knows
8539 what it is doing and can handle the new informational codes
8543 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8544 a general "ANY" type, as such it should be able to decode anything
8545 including tagged types. However it didn't check the class so it would
8546 wrongly interpret tagged types in the same way as their universal
8547 counterpart and unknown types were just rejected. Changed so that the
8548 tagged and unknown types are handled in the same way as a SEQUENCE:
8549 that is the encoding is stored intact. There is also a new type
8550 "V_ASN1_OTHER" which is used when the class is not universal, in this
8551 case we have no idea what the actual type is so we just lump them all
8555 *) On VMS, stdout may very well lead to a file that is written to
8556 in a record-oriented fashion. That means that every write() will
8557 write a separate record, which will be read separately by the
8558 programs trying to read from it. This can be very confusing.
8560 The solution is to put a BIO filter in the way that will buffer
8561 text until a linefeed is reached, and then write everything a
8562 line at a time, so every record written will be an actual line,
8563 not chunks of lines and not (usually doesn't happen, but I've
8564 seen it once) several lines in one record. BIO_f_linebuffer() is
8567 Currently, it's a VMS-only method, because that's where it has
8568 been tested well enough.
8571 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8572 it can return incorrect results.
8573 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8574 but it was in 0.9.6-beta[12].)
8577 *) Disable the check for content being present when verifying detached
8578 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8579 include zero length content when signing messages.
8582 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8583 BIO_ctrl (for BIO pairs).
8586 *) Add DSO method for VMS.
8589 *) Bug fix: Montgomery multiplication could produce results with the
8593 *) Add RPM specification openssl.spec and modify it to build three
8594 packages. The default package contains applications, application
8595 documentation and run-time libraries. The devel package contains
8596 include files, static libraries and function documentation. The
8597 doc package contains the contents of the doc directory. The original
8598 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8601 *) Add a large number of documentation files for many SSL routines.
8602 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8604 *) Add a configuration entry for Sony News 4.
8605 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8607 *) Don't set the two most significant bits to one when generating a
8608 random number < q in the DSA library.
8611 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8612 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8613 the underlying transport is blocking) if a handshake took place.
8614 (The default behaviour is needed by applications such as s_client
8615 and s_server that use select() to determine when to use SSL_read;
8616 but for applications that know in advance when to expect data, it
8617 just makes things more complicated.)
8620 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8624 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8625 work better on such systems.
8626 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8628 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8629 Update PKCS12_parse() so it copies the friendlyName and the
8630 keyid to the certificates aux info.
8633 *) Fix bug in PKCS7_verify() which caused an infinite loop
8634 if there was more than one signature.
8635 [Sven Uszpelkat <su@celocom.de>]
8637 *) Major change in util/mkdef.pl to include extra information
8638 about each symbol, as well as presenting variables as well
8639 as functions. This change means that there's n more need
8640 to rebuild the .num files when some algorithms are excluded.
8643 *) Allow the verify time to be set by an application,
8644 rather than always using the current time.
8647 *) Phase 2 verify code reorganisation. The certificate
8648 verify code now looks up an issuer certificate by a
8649 number of criteria: subject name, authority key id
8650 and key usage. It also verifies self signed certificates
8651 by the same criteria. The main comparison function is
8652 X509_check_issued() which performs these checks.
8654 Lot of changes were necessary in order to support this
8655 without completely rewriting the lookup code.
8657 Authority and subject key identifier are now cached.
8659 The LHASH 'certs' is X509_STORE has now been replaced
8660 by a STACK_OF(X509_OBJECT). This is mainly because an
8661 LHASH can't store or retrieve multiple objects with
8662 the same hash value.
8664 As a result various functions (which were all internal
8665 use only) have changed to handle the new X509_STORE
8666 structure. This will break anything that messed round
8667 with X509_STORE internally.
8669 The functions X509_STORE_add_cert() now checks for an
8670 exact match, rather than just subject name.
8672 The X509_STORE API doesn't directly support the retrieval
8673 of multiple certificates matching a given criteria, however
8674 this can be worked round by performing a lookup first
8675 (which will fill the cache with candidate certificates)
8676 and then examining the cache for matches. This is probably
8677 the best we can do without throwing out X509_LOOKUP
8678 entirely (maybe later...).
8680 The X509_VERIFY_CTX structure has been enhanced considerably.
8682 All certificate lookup operations now go via a get_issuer()
8683 callback. Although this currently uses an X509_STORE it
8684 can be replaced by custom lookups. This is a simple way
8685 to bypass the X509_STORE hackery necessary to make this
8686 work and makes it possible to use more efficient techniques
8687 in future. A very simple version which uses a simple
8688 STACK for its trusted certificate store is also provided
8689 using X509_STORE_CTX_trusted_stack().
8691 The verify_cb() and verify() callbacks now have equivalents
8692 in the X509_STORE_CTX structure.
8694 X509_STORE_CTX also has a 'flags' field which can be used
8695 to customise the verify behaviour.
8698 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8699 excludes S/MIME capabilities.
8702 *) When a certificate request is read in keep a copy of the
8703 original encoding of the signed data and use it when outputing
8704 again. Signatures then use the original encoding rather than
8705 a decoded, encoded version which may cause problems if the
8706 request is improperly encoded.
8709 *) For consistency with other BIO_puts implementations, call
8710 buffer_write(b, ...) directly in buffer_puts instead of calling
8713 In BIO_puts, increment b->num_write as in BIO_write.
8714 [Peter.Sylvester@EdelWeb.fr]
8716 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8717 BN_zero, we may not return a BIGNUM with an array consisting of
8721 *) Avoid calling abort() from within the library when problems are
8722 detected, except if preprocessor symbols have been defined
8723 (such as REF_CHECK, BN_DEBUG etc.).
8726 *) New openssl application 'rsautl'. This utility can be
8727 used for low level RSA operations. DER public key
8728 BIO/fp routines also added.
8731 *) New Configure entry and patches for compiling on QNX 4.
8732 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8734 *) A demo state-machine implementation was sponsored by
8735 Nuron (http://www.nuron.com/) and is now available in
8736 demos/state_machine.
8739 *) New options added to the 'dgst' utility for signature
8740 generation and verification.
8743 *) Unrecognized PKCS#7 content types are now handled via a
8744 catch all ASN1_TYPE structure. This allows unsupported
8745 types to be stored as a "blob" and an application can
8746 encode and decode it manually.
8749 *) Fix various signed/unsigned issues to make a_strex.c
8751 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8753 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8754 length if passed a buffer. ASN1_INTEGER_to_BN failed
8755 if passed a NULL BN and its argument was negative.
8756 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8758 *) Modification to PKCS#7 encoding routines to output definite
8759 length encoding. Since currently the whole structures are in
8760 memory there's not real point in using indefinite length
8761 constructed encoding. However if OpenSSL is compiled with
8762 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8765 *) Added BIO_vprintf() and BIO_vsnprintf().
8768 *) Added more prefixes to parse for in the the strings written
8769 through a logging bio, to cover all the levels that are available
8770 through syslog. The prefixes are now:
8772 PANIC, EMERG, EMR => LOG_EMERG
8773 ALERT, ALR => LOG_ALERT
8774 CRIT, CRI => LOG_CRIT
8775 ERROR, ERR => LOG_ERR
8776 WARNING, WARN, WAR => LOG_WARNING
8777 NOTICE, NOTE, NOT => LOG_NOTICE
8778 INFO, INF => LOG_INFO
8779 DEBUG, DBG => LOG_DEBUG
8781 and as before, if none of those prefixes are present at the
8782 beginning of the string, LOG_ERR is chosen.
8784 On Win32, the LOG_* levels are mapped according to this:
8786 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8787 LOG_WARNING => EVENTLOG_WARNING_TYPE
8788 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8792 *) Made it possible to reconfigure with just the configuration
8793 argument "reconf" or "reconfigure". The command line arguments
8794 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8795 and are retrieved from there when reconfiguring.
8799 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8801 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8804 *) The obj_dat.pl script was messing up the sorting of object
8805 names. The reason was that it compared the quoted version
8806 of strings as a result "OCSP" > "OCSP Signing" because
8807 " > SPACE. Changed script to store unquoted versions of
8808 names and add quotes on output. It was also omitting some
8809 names from the lookup table if they were given a default
8810 value (that is if SN is missing it is given the same
8811 value as LN and vice versa), these are now added on the
8812 grounds that if an object has a name we should be able to
8813 look it up. Finally added warning output when duplicate
8814 short or long names are found.
8817 *) Changes needed for Tandem NSK.
8818 [Scott Uroff <scott@xypro.com>]
8820 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8821 RSA_padding_check_SSLv23(), special padding was never detected
8822 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8823 version rollback attacks was not effective.
8825 In s23_clnt.c, don't use special rollback-attack detection padding
8826 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8827 client; similarly, in s23_srvr.c, don't do the rollback check if
8828 SSL 2.0 is the only protocol enabled in the server.
8831 *) Make it possible to get hexdumps of unprintable data with 'openssl
8832 asn1parse'. By implication, the functions ASN1_parse_dump() and
8833 BIO_dump_indent() are added.
8836 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8837 these print out strings and name structures based on various
8838 flags including RFC2253 support and proper handling of
8839 multibyte characters. Added options to the 'x509' utility
8840 to allow the various flags to be set.
8843 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8844 Also change the functions X509_cmp_current_time() and
8845 X509_gmtime_adj() work with an ASN1_TIME structure,
8846 this will enable certificates using GeneralizedTime in validity
8847 dates to be checked.
8850 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8851 negative public key encodings) on by default,
8852 NO_NEG_PUBKEY_BUG can be set to disable it.
8855 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8856 content octets. An i2c_ASN1_OBJECT is unnecessary because
8857 the encoding can be trivially obtained from the structure.
8860 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8861 not read locks (CRYPTO_r_[un]lock).
8864 *) A first attempt at creating official support for shared
8865 libraries through configuration. I've kept it so the
8866 default is static libraries only, and the OpenSSL programs
8867 are always statically linked for now, but there are
8868 preparations for dynamic linking in place.
8869 This has been tested on Linux and Tru64.
8872 *) Randomness polling function for Win9x, as described in:
8873 Peter Gutmann, Software Generation of Practically Strong
8877 *) Fix so PRNG is seeded in req if using an already existing
8881 *) New options to smime application. -inform and -outform
8882 allow alternative formats for the S/MIME message including
8883 PEM and DER. The -content option allows the content to be
8884 specified separately. This should allow things like Netscape
8885 form signing output easier to verify.
8888 *) Fix the ASN1 encoding of tags using the 'long form'.
8891 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8892 STRING types. These convert content octets to and from the
8893 underlying type. The actual tag and length octets are
8894 already assumed to have been read in and checked. These
8895 are needed because all other string types have virtually
8896 identical handling apart from the tag. By having versions
8897 of the ASN1 functions that just operate on content octets
8898 IMPLICIT tagging can be handled properly. It also allows
8899 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8900 and ASN1_INTEGER are identical apart from the tag.
8903 *) Change the handling of OID objects as follows:
8905 - New object identifiers are inserted in objects.txt, following
8906 the syntax given in objects.README.
8907 - objects.pl is used to process obj_mac.num and create a new
8909 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8912 This is currently kind of a hack, and the perl code in objects.pl
8913 isn't very elegant, but it works as I intended. The simplest way
8914 to check that it worked correctly is to look in obj_dat.h and
8915 check the array nid_objs and make sure the objects haven't moved
8916 around (this is important!). Additions are OK, as well as
8917 consistent name changes.
8920 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8923 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8924 The given file adds to whatever has already been seeded into the
8925 random pool through the RANDFILE configuration file option or
8926 environment variable, or the default random state file.
8929 *) mkstack.pl now sorts each macro group into lexical order.
8930 Previously the output order depended on the order the files
8931 appeared in the directory, resulting in needless rewriting
8935 *) Patches to make OpenSSL compile under Win32 again. Mostly
8936 work arounds for the VC++ problem that it treats func() as
8937 func(void). Also stripped out the parts of mkdef.pl that
8938 added extra typesafe functions: these no longer exist.
8941 *) Reorganisation of the stack code. The macros are now all
8942 collected in safestack.h . Each macro is defined in terms of
8943 a "stack macro" of the form SKM_<name>(type, a, b). The
8944 DEBUG_SAFESTACK is now handled in terms of function casts,
8945 this has the advantage of retaining type safety without the
8946 use of additional functions. If DEBUG_SAFESTACK is not defined
8947 then the non typesafe macros are used instead. Also modified the
8948 mkstack.pl script to handle the new form. Needs testing to see
8949 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8950 the default if no major problems. Similar behaviour for ASN1_SET_OF
8951 and PKCS12_STACK_OF.
8954 *) When some versions of IIS use the 'NET' form of private key the
8955 key derivation algorithm is different. Normally MD5(password) is
8956 used as a 128 bit RC4 key. In the modified case
8957 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8958 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8959 as the old Netscape_RSA functions except they have an additional
8960 'sgckey' parameter which uses the modified algorithm. Also added
8961 an -sgckey command line option to the rsa utility. Thanks to
8962 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8963 algorithm to openssl-dev.
8966 *) The evp_local.h macros were using 'c.##kname' which resulted in
8967 invalid expansion on some systems (SCO 5.0.5 for example).
8968 Corrected to 'c.kname'.
8969 [Phillip Porch <root@theporch.com>]
8971 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8972 a STACK of email addresses from a certificate or request, these look
8973 in the subject name and the subject alternative name extensions and
8974 omit any duplicate addresses.
8977 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8978 This makes DSA verification about 2 % faster.
8981 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8982 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8983 plus overhead for 1024 bit moduli).
8984 This makes exponentiations about 0.5 % faster for 1024 bit
8985 exponents (as measured by "openssl speed rsa2048").
8988 *) Rename memory handling macros to avoid conflicts with other
8990 Malloc => OPENSSL_malloc
8991 Malloc_locked => OPENSSL_malloc_locked
8992 Realloc => OPENSSL_realloc
8993 Free => OPENSSL_free
8996 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8997 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9000 *) CygWin32 support.
9001 [John Jarvie <jjarvie@newsguy.com>]
9003 *) The type-safe stack code has been rejigged. It is now only compiled
9004 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9005 by default all type-specific stack functions are "#define"d back to
9006 standard stack functions. This results in more streamlined output
9007 but retains the type-safety checking possibilities of the original
9011 *) The STACK code has been cleaned up, and certain type declarations
9012 that didn't make a lot of sense have been brought in line. This has
9013 also involved a cleanup of sorts in safestack.h to more correctly
9014 map type-safe stack functions onto their plain stack counterparts.
9015 This work has also resulted in a variety of "const"ifications of
9016 lots of the code, especially "_cmp" operations which should normally
9017 be prototyped with "const" parameters anyway.
9020 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9021 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9022 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9023 where all of 'md' is used each time the PRNG is used, but 'state'
9024 is used only indexed by a cyclic counter. As entropy may not be
9025 well distributed from the beginning, 'md' is important as a
9026 chaining variable. However, the output function chains only half
9027 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9028 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9029 in all of 'state' being rewritten, with the new values depending
9030 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9033 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9034 the handshake is continued after ssl_verify_cert_chain();
9035 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9036 can lead to 'unexplainable' connection aborts later.
9037 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9039 *) Major EVP API cipher revision.
9040 Add hooks for extra EVP features. This allows various cipher
9041 parameters to be set in the EVP interface. Support added for variable
9042 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9043 setting of RC2 and RC5 parameters.
9045 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9048 Remove lots of duplicated code from the EVP library. For example *every*
9049 cipher init() function handles the 'iv' in the same way according to the
9050 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9051 for CFB and OFB modes they zero ctx->num.
9053 New functionality allows removal of S/MIME code RC2 hack.
9055 Most of the routines have the same form and so can be declared in terms
9058 By shifting this to the top level EVP_CipherInit() it can be removed from
9059 all individual ciphers. If the cipher wants to handle IVs or keys
9060 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9063 Change lots of functions like EVP_EncryptUpdate() to now return a
9064 value: although software versions of the algorithms cannot fail
9065 any installed hardware versions can.
9068 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9069 this option is set, tolerate broken clients that send the negotiated
9070 protocol version number instead of the requested protocol version
9074 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9075 i.e. non-zero for export ciphersuites, zero otherwise.
9076 Previous versions had this flag inverted, inconsistent with
9077 rsa_tmp_cb (..._TMP_RSA_CB).
9078 [Bodo Moeller; problem reported by Amit Chopra]
9080 *) Add missing DSA library text string. Work around for some IIS
9081 key files with invalid SEQUENCE encoding.
9084 *) Add a document (doc/standards.txt) that list all kinds of standards
9085 and so on that are implemented in OpenSSL.
9088 *) Enhance c_rehash script. Old version would mishandle certificates
9089 with the same subject name hash and wouldn't handle CRLs at all.
9090 Added -fingerprint option to crl utility, to support new c_rehash
9094 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9097 *) Fix for SSL server purpose checking. Server checking was
9098 rejecting certificates which had extended key usage present
9099 but no ssl client purpose.
9100 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9102 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9103 is a little unclear about how a blank password is handled.
9104 Since the password in encoded as a BMPString with terminating
9105 double NULL a zero length password would end up as just the
9106 double NULL. However no password at all is different and is
9107 handled differently in the PKCS#12 key generation code. NS
9108 treats a blank password as zero length. MSIE treats it as no
9109 password on export: but it will try both on import. We now do
9110 the same: PKCS12_parse() tries zero length and no password if
9111 the password is set to "" or NULL (NULL is now a valid password:
9112 it wasn't before) as does the pkcs12 application.
9115 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9116 perror when PEM_read_bio_X509_REQ fails, the error message must
9117 be obtained from the error queue.
9120 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9121 it in ERR_remove_state if appropriate, and change ERR_get_state
9122 accordingly to avoid race conditions (this is necessary because
9123 thread_hash is no longer constant once set).
9126 *) Bugfix for linux-elf makefile.one.
9129 *) RSA_get_default_method() will now cause a default
9130 RSA_METHOD to be chosen if one doesn't exist already.
9131 Previously this was only set during a call to RSA_new()
9132 or RSA_new_method(NULL) meaning it was possible for
9133 RSA_get_default_method() to return NULL.
9136 *) Added native name translation to the existing DSO code
9137 that will convert (if the flag to do so is set) filenames
9138 that are sufficiently small and have no path information
9139 into a canonical native form. Eg. "blah" converted to
9140 "libblah.so" or "blah.dll" etc.
9143 *) New function ERR_error_string_n(e, buf, len) which is like
9144 ERR_error_string(e, buf), but writes at most 'len' bytes
9145 including the 0 terminator. For ERR_error_string_n, 'buf'
9147 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9149 *) CONF library reworked to become more general. A new CONF
9150 configuration file reader "class" is implemented as well as a
9151 new functions (NCONF_*, for "New CONF") to handle it. The now
9152 old CONF_* functions are still there, but are reimplemented to
9153 work in terms of the new functions. Also, a set of functions
9154 to handle the internal storage of the configuration data is
9155 provided to make it easier to write new configuration file
9156 reader "classes" (I can definitely see something reading a
9157 configuration file in XML format, for example), called _CONF_*,
9158 or "the configuration storage API"...
9160 The new configuration file reading functions are:
9162 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9163 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9165 NCONF_default, NCONF_WIN32
9167 NCONF_dump_fp, NCONF_dump_bio
9169 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9170 NCONF_new creates a new CONF object. This works in the same way
9171 as other interfaces in OpenSSL, like the BIO interface.
9172 NCONF_dump_* dump the internal storage of the configuration file,
9173 which is useful for debugging. All other functions take the same
9174 arguments as the old CONF_* functions wth the exception of the
9175 first that must be a `CONF *' instead of a `LHASH *'.
9177 To make it easer to use the new classes with the old CONF_* functions,
9178 the function CONF_set_default_method is provided.
9181 *) Add '-tls1' option to 'openssl ciphers', which was already
9182 mentioned in the documentation but had not been implemented.
9183 (This option is not yet really useful because even the additional
9184 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9187 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9188 OpenSSL-based applications) load shared libraries and bind to
9189 them in a portable way.
9190 [Geoff Thorpe, with contributions from Richard Levitte]
9192 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9194 *) Make sure _lrotl and _lrotr are only used with MSVC.
9196 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9197 (the default implementation of RAND_status).
9199 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9200 to '-clrext' (= clear extensions), as intended and documented.
9201 [Bodo Moeller; inconsistency pointed out by Michael Attili
9204 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9205 was larger than the MD block size.
9206 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9208 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9209 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9210 using the passed key: if the passed key was a private key the result
9211 of X509_print(), for example, would be to print out all the private key
9215 *) des_quad_cksum() byte order bug fix.
9216 [Ulf Möller, using the problem description in krb4-0.9.7, where
9217 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9219 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9221 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9223 *) For easily testing in shell scripts whether some command
9224 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9225 returns with exit code 0 iff no command of the given name is available.
9226 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9227 the output goes to stdout and nothing is printed to stderr.
9228 Additional arguments are always ignored.
9230 Since for each cipher there is a command of the same name,
9231 the 'no-cipher' compilation switches can be tested this way.
9233 ('openssl no-XXX' is not able to detect pseudo-commands such
9234 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9237 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9240 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9241 is set; it will be thrown away anyway because each handshake creates
9243 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9244 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9245 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9246 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9249 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9250 'Q' and 'R' lose their special meanings (quit/renegotiate).
9251 This is part of what -quiet does; unlike -quiet, -ign_eof
9252 does not suppress any output.
9255 *) Add compatibility options to the purpose and trust code. The
9256 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9257 accepts a certificate or CA, this was the previous behaviour,
9258 with all the associated security issues.
9260 X509_TRUST_COMPAT is the old trust behaviour: only and
9261 automatically trust self signed roots in certificate store. A
9262 new trust setting X509_TRUST_DEFAULT is used to specify that
9263 a purpose has no associated trust setting and it should instead
9264 use the value in the default purpose.
9267 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9268 and fix a memory leak.
9271 *) In util/mkerr.pl (which implements 'make errors'), preserve
9272 reason strings from the previous version of the .c file, as
9273 the default to have only downcase letters (and digits) in
9274 automatically generated reasons codes is not always appropriate.
9277 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9278 using strerror. Previously, ERR_reason_error_string() returned
9279 library names as reason strings for SYSerr; but SYSerr is a special
9280 case where small numbers are errno values, not library numbers.
9283 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9284 converts DSA parameters into DH parameters. (When creating parameters,
9285 DSA_generate_parameters is used.)
9288 *) Include 'length' (recommended exponent length) in C code generated
9289 by 'openssl dhparam -C'.
9292 *) The second argument to set_label in perlasm was already being used
9293 so couldn't be used as a "file scope" flag. Moved to third argument
9297 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9298 instead of RAND_bytes for encryption IVs and salts.
9301 *) Include RAND_status() into RAND_METHOD instead of implementing
9302 it only for md_rand.c Otherwise replacing the PRNG by calling
9303 RAND_set_rand_method would be impossible.
9306 *) Don't let DSA_generate_key() enter an infinite loop if the random
9307 number generation fails.
9310 *) New 'rand' application for creating pseudo-random output.
9313 *) Added configuration support for Linux/IA64
9314 [Rolf Haberrecker <rolf@suse.de>]
9316 *) Assembler module support for Mingw32.
9319 *) Shared library support for HPUX (in shlib/).
9320 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9322 *) Shared library support for Solaris gcc.
9323 [Lutz Behnke <behnke@trustcenter.de>]
9325 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9327 *) PKCS7_encrypt() was adding text MIME headers twice because they
9328 were added manually and by SMIME_crlf_copy().
9331 *) In bntest.c don't call BN_rand with zero bits argument.
9332 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9334 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9335 case was implemented. This caused BN_div_recp() to fail occasionally.
9338 *) Add an optional second argument to the set_label() in the perl
9339 assembly language builder. If this argument exists and is set
9340 to 1 it signals that the assembler should use a symbol whose
9341 scope is the entire file, not just the current function. This
9342 is needed with MASM which uses the format label:: for this scope.
9343 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9345 *) Change the ASN1 types so they are typedefs by default. Before
9346 almost all types were #define'd to ASN1_STRING which was causing
9347 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9351 *) Change names of new functions to the new get1/get0 naming
9352 convention: After 'get1', the caller owns a reference count
9353 and has to call ..._free; 'get0' returns a pointer to some
9354 data structure without incrementing reference counters.
9355 (Some of the existing 'get' functions increment a reference
9356 counter, some don't.)
9357 Similarly, 'set1' and 'add1' functions increase reference
9358 counters or duplicate objects.
9361 *) Allow for the possibility of temp RSA key generation failure:
9362 the code used to assume it always worked and crashed on failure.
9365 *) Fix potential buffer overrun problem in BIO_printf().
9366 [Ulf Möller, using public domain code by Patrick Powell; problem
9367 pointed out by David Sacerdote <das33@cornell.edu>]
9369 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9370 RAND_egd() and RAND_status(). In the command line application,
9371 the EGD socket can be specified like a seed file using RANDFILE
9375 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9376 Some CAs (e.g. Verisign) distribute certificates in this form.
9379 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9380 list to exclude them. This means that no special compilation option
9381 is needed to use anonymous DH: it just needs to be included in the
9385 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9386 EVP_MD_type. The old functionality is available in a new macro called
9387 EVP_MD_md(). Change code that uses it and update docs.
9390 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9391 where the 'void *' argument is replaced by a function pointer argument.
9392 Previously 'void *' was abused to point to functions, which works on
9393 many platforms, but is not correct. As these functions are usually
9394 called by macros defined in OpenSSL header files, most source code
9395 should work without changes.
9398 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9399 sections with information on -D... compiler switches used for
9400 compiling the library so that applications can see them. To enable
9401 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9402 must be defined. E.g.,
9403 #define OPENSSL_ALGORITHM_DEFINES
9404 #include <openssl/opensslconf.h>
9405 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9406 [Richard Levitte, Ulf and Bodo Möller]
9408 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9412 *) Change the 'other' type in certificate aux info to a STACK_OF
9413 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9414 the required ASN1 format: arbitrary types determined by an OID.
9417 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9418 argument to 'req'. This is not because the function is newer or
9419 better than others it just uses the work 'NEW' in the certificate
9420 request header lines. Some software needs this.
9423 *) Reorganise password command line arguments: now passwords can be
9424 obtained from various sources. Delete the PEM_cb function and make
9425 it the default behaviour: i.e. if the callback is NULL and the
9426 usrdata argument is not NULL interpret it as a null terminated pass
9427 phrase. If usrdata and the callback are NULL then the pass phrase
9428 is prompted for as usual.
9431 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9432 the support is automatically enabled. The resulting binaries will
9433 autodetect the card and use it if present.
9434 [Ben Laurie and Compaq Inc.]
9436 *) Work around for Netscape hang bug. This sends certificate request
9437 and server done in one record. Since this is perfectly legal in the
9438 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9439 the bugs/SSLv3 entry for more info.
9442 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9445 *) Add -rand argument to smime and pkcs12 applications and read/write
9449 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9452 *) Add command line password options to the remaining applications.
9455 *) Bug fix for BN_div_recp() for numerators with an even number of
9459 *) More tests in bntest.c, and changed test_bn output.
9462 *) ./config recognizes MacOS X now.
9465 *) Bug fix for BN_div() when the first words of num and divsor are
9466 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9469 *) Add support for various broken PKCS#8 formats, and command line
9470 options to produce them.
9473 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9474 get temporary BIGNUMs from a BN_CTX.
9477 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9481 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9482 include a #define from the old name to the new. The original intent
9483 was that statically linked binaries could for example just call
9484 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9485 link with digests. This never worked because SSLeay_add_all_digests()
9486 and SSLeay_add_all_ciphers() were in the same source file so calling
9487 one would link with the other. They are now in separate source files.
9490 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9493 *) Use a less unusual form of the Miller-Rabin primality test (it used
9494 a binary algorithm for exponentiation integrated into the Miller-Rabin
9495 loop, our standard modexp algorithms are faster).
9498 *) Support for the EBCDIC character set completed.
9499 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9501 *) Source code cleanups: use const where appropriate, eliminate casts,
9502 use void * instead of char * in lhash.
9505 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9506 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9507 this the server could overwrite ephemeral keys that the client
9511 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9512 using 50 iterations of the Rabin-Miller test.
9514 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9515 iterations of the Rabin-Miller test as required by the appendix
9516 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9517 As BN_is_prime_fasttest includes trial division, DSA parameter
9518 generation becomes much faster.
9520 This implies a change for the callback functions in DSA_is_prime
9521 and DSA_generate_parameters: The callback function is called once
9522 for each positive witness in the Rabin-Miller test, not just
9523 occasionally in the inner loop; and the parameters to the
9524 callback function now provide an iteration count for the outer
9525 loop rather than for the current invocation of the inner loop.
9526 DSA_generate_parameters additionally can call the callback
9527 function with an 'iteration count' of -1, meaning that a
9528 candidate has passed the trial division test (when q is generated
9529 from an application-provided seed, trial division is skipped).
9532 *) New function BN_is_prime_fasttest that optionally does trial
9533 division before starting the Rabin-Miller test and has
9534 an additional BN_CTX * argument (whereas BN_is_prime always
9535 has to allocate at least one BN_CTX).
9536 'callback(1, -1, cb_arg)' is called when a number has passed the
9537 trial division stage.
9540 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9544 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9547 *) New function BN_pseudo_rand().
9550 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9551 bignum version of BN_from_montgomery() with the working code from
9552 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9556 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9557 made it impossible to use the same SSL_SESSION data structure in
9558 SSL2 clients in multiple threads.
9561 *) The return value of RAND_load_file() no longer counts bytes obtained
9562 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9563 to seed the PRNG (previously an explicit byte count was required).
9564 [Ulf Möller, Bodo Möller]
9566 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9567 used (char *) instead of (void *) and had casts all over the place.
9570 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9573 *) Retain source code compatibility for BN_prime_checks macro:
9574 BN_is_prime(..., BN_prime_checks, ...) now uses
9575 BN_prime_checks_for_size to determine the appropriate number of
9576 Rabin-Miller iterations.
9579 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9580 DH_CHECK_P_NOT_SAFE_PRIME.
9581 (Check if this is true? OpenPGP calls them "strong".)
9584 *) Merge the functionality of "dh" and "gendh" programs into a new program
9585 "dhparam". The old programs are retained for now but will handle DH keys
9586 (instead of parameters) in future.
9589 *) Make the ciphers, s_server and s_client programs check the return values
9590 when a new cipher list is set.
9593 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9594 ciphers. Before when the 56bit ciphers were enabled the sorting was
9597 The syntax for the cipher sorting has been extended to support sorting by
9598 cipher-strength (using the strength_bits hard coded in the tables).
9599 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9601 Fix a bug in the cipher-command parser: when supplying a cipher command
9602 string with an "undefined" symbol (neither command nor alphanumeric
9603 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9604 an error is flagged.
9606 Due to the strength-sorting extension, the code of the
9607 ssl_create_cipher_list() function was completely rearranged. I hope that
9608 the readability was also increased :-)
9609 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9611 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9612 for the first serial number and places 2 in the serial number file. This
9613 avoids problems when the root CA is created with serial number zero and
9614 the first user certificate has the same issuer name and serial number
9618 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9619 the new code. Add documentation for this stuff.
9622 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9623 X509_*() to X509at_*() on the grounds that they don't handle X509
9624 structures and behave in an analogous way to the X509v3 functions:
9625 they shouldn't be called directly but wrapper functions should be used
9628 So we also now have some wrapper functions that call the X509at functions
9629 when passed certificate requests. (TO DO: similar things can be done with
9630 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9631 things. Some of these need some d2i or i2d and print functionality
9632 because they handle more complex structures.)
9635 *) Add missing #ifndefs that caused missing symbols when building libssl
9636 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9637 NO_RSA in ssl/s2*.c.
9638 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9640 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9641 has a return value which indicates the quality of the random data
9642 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9643 error queue. New function RAND_pseudo_bytes() generates output that is
9644 guaranteed to be unique but not unpredictable. RAND_add is like
9645 RAND_seed, but takes an extra argument for an entropy estimate
9646 (RAND_seed always assumes full entropy).
9649 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9650 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9651 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9652 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9653 false-positive rate of at most 2^-80 for random input.
9656 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9659 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9660 in the 0.9.5 release), this returns the chain
9661 from an X509_CTX structure with a dup of the stack and all
9662 the X509 reference counts upped: so the stack will exist
9663 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9666 Also make SSL_SESSION_print() print out the verify return
9670 *) Add manpage for the pkcs12 command. Also change the default
9671 behaviour so MAC iteration counts are used unless the new
9672 -nomaciter option is used. This improves file security and
9673 only older versions of MSIE (4.0 for example) need it.
9676 *) Honor the no-xxx Configure options when creating .DEF files.
9679 *) Add PKCS#10 attributes to field table: challengePassword,
9680 unstructuredName and unstructuredAddress. These are taken from
9681 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9682 international characters are used.
9684 More changes to X509_ATTRIBUTE code: allow the setting of types
9685 based on strings. Remove the 'loc' parameter when adding
9686 attributes because these will be a SET OF encoding which is sorted
9690 *) Initial changes to the 'req' utility to allow request generation
9691 automation. This will allow an application to just generate a template
9692 file containing all the field values and have req construct the
9695 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9696 used all over the place including certificate requests and PKCS#7
9697 structures. They are currently handled manually where necessary with
9698 some primitive wrappers for PKCS#7. The new functions behave in a
9699 manner analogous to the X509 extension functions: they allow
9700 attributes to be looked up by NID and added.
9702 Later something similar to the X509V3 code would be desirable to
9703 automatically handle the encoding, decoding and printing of the
9704 more complex types. The string types like challengePassword can
9705 be handled by the string table functions.
9707 Also modified the multi byte string table handling. Now there is
9708 a 'global mask' which masks out certain types. The table itself
9709 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9710 is useful when for example there is only one permissible type
9711 (as in countryName) and using the mask might result in no valid
9715 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9716 SSL_get_peer_finished to allow applications to obtain the latest
9717 Finished messages sent to the peer or expected from the peer,
9718 respectively. (SSL_get_peer_finished is usually the Finished message
9719 actually received from the peer, otherwise the protocol will be aborted.)
9721 As the Finished message are message digests of the complete handshake
9722 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9723 be used for external authentication procedures when the authentication
9724 provided by SSL/TLS is not desired or is not enough.
9727 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9728 the host supports BWX extension and if Compaq C is present on the
9729 $PATH. Just exploiting of the BWX extension results in 20-30%
9730 performance kick for some algorithms, e.g. DES and RC4 to mention
9731 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9735 *) Add support for MS "fast SGC". This is arguably a violation of the
9736 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9737 weak crypto and after checking the certificate is SGC a second one
9738 with strong crypto. MS SGC stops the first handshake after receiving
9739 the server certificate message and sends a second client hello. Since
9740 a server will typically do all the time consuming operations before
9741 expecting any further messages from the client (server key exchange
9742 is the most expensive) there is little difference between the two.
9744 To get OpenSSL to support MS SGC we have to permit a second client
9745 hello message after we have sent server done. In addition we have to
9746 reset the MAC if we do get this second client hello.
9749 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9750 if a DER encoded private key is RSA or DSA traditional format. Changed
9751 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9752 format DER encoded private key. Newer code should use PKCS#8 format which
9753 has the key type encoded in the ASN1 structure. Added DER private key
9754 support to pkcs8 application.
9757 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9758 ciphersuites has been selected (as required by the SSL 3/TLS 1
9759 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9760 is set, we interpret this as a request to violate the specification
9761 (the worst that can happen is a handshake failure, and 'correct'
9762 behaviour would result in a handshake failure anyway).
9765 *) In SSL_CTX_add_session, take into account that there might be multiple
9766 SSL_SESSION structures with the same session ID (e.g. when two threads
9767 concurrently obtain them from an external cache).
9768 The internal cache can handle only one SSL_SESSION with a given ID,
9769 so if there's a conflict, we now throw out the old one to achieve
9773 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9774 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9775 some routines that use cipher OIDs: some ciphers do not have OIDs
9776 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9780 *) Simplify the trust setting structure and code. Now we just have
9781 two sequences of OIDs for trusted and rejected settings. These will
9782 typically have values the same as the extended key usage extension
9783 and any application specific purposes.
9785 The trust checking code now has a default behaviour: it will just
9786 check for an object with the same NID as the passed id. Functions can
9787 be provided to override either the default behaviour or the behaviour
9788 for a given id. SSL client, server and email already have functions
9789 in place for compatibility: they check the NID and also return "trusted"
9790 if the certificate is self signed.
9793 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9794 traditional format into an EVP_PKEY structure.
9797 *) Add a password callback function PEM_cb() which either prompts for
9798 a password if usr_data is NULL or otherwise assumes it is a null
9799 terminated password. Allow passwords to be passed on command line
9800 environment or config files in a few more utilities.
9803 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9804 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9805 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9806 Update documentation.
9809 *) Support for ASN1 "NULL" type. This could be handled before by using
9810 ASN1_TYPE but there wasn't any function that would try to read a NULL
9811 and produce an error if it couldn't. For compatibility we also have
9812 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9813 don't allocate anything because they don't need to.
9816 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9818 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9820 *) Rebuild of the memory allocation routines used by OpenSSL code and
9821 possibly others as well. The purpose is to make an interface that
9822 provide hooks so anyone can build a separate set of allocation and
9823 deallocation routines to be used by OpenSSL, for example memory
9824 pool implementations, or something else, which was previously hard
9825 since Malloc(), Realloc() and Free() were defined as macros having
9826 the values malloc, realloc and free, respectively (except for Win32
9827 compilations). The same is provided for memory debugging code.
9828 OpenSSL already comes with functionality to find memory leaks, but
9829 this gives people a chance to debug other memory problems.
9831 With these changes, a new set of functions and macros have appeared:
9833 CRYPTO_set_mem_debug_functions() [F]
9834 CRYPTO_get_mem_debug_functions() [F]
9835 CRYPTO_dbg_set_options() [F]
9836 CRYPTO_dbg_get_options() [F]
9837 CRYPTO_malloc_debug_init() [M]
9839 The memory debug functions are NULL by default, unless the library
9840 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9841 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9842 gives the standard debugging functions that come with OpenSSL) or
9843 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9844 provided by the library user) must be used. When the standard
9845 debugging functions are used, CRYPTO_dbg_set_options can be used to
9846 request additional information:
9847 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9848 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9850 Also, things like CRYPTO_set_mem_functions will always give the
9851 expected result (the new set of functions is used for allocation
9852 and deallocation) at all times, regardless of platform and compiler
9855 To finish it up, some functions that were never use in any other
9856 way than through macros have a new API and new semantic:
9859 CRYPTO_dbg_realloc()
9862 All macros of value have retained their old syntax.
9863 [Richard Levitte and Bodo Moeller]
9865 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9866 ordering of SMIMECapabilities wasn't in "strength order" and there
9867 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9871 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9872 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9873 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9875 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9876 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9877 functionality to handle multipart/signed properly) and a utility
9878 called 'smime' to call all this stuff. This is based on code I
9879 originally wrote for Celo who have kindly allowed it to be
9880 included in OpenSSL.
9883 *) Add variants des_set_key_checked and des_set_key_unchecked of
9884 des_set_key (aka des_key_sched). Global variable des_check_key
9885 decides which of these is called by des_set_key; this way
9886 des_check_key behaves as it always did, but applications and
9887 the library itself, which was buggy for des_check_key == 1,
9888 have a cleaner way to pick the version they need.
9891 *) New function PKCS12_newpass() which changes the password of a
9895 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9896 dynamic mix. In both cases the ids can be used as an index into the
9897 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9898 functions so they accept a list of the field values and the
9899 application doesn't need to directly manipulate the X509_TRUST
9903 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9907 *) Modify the way the V3 extension code looks up extensions. This now
9908 works in a similar way to the object code: we have some "standard"
9909 extensions in a static table which is searched with OBJ_bsearch()
9910 and the application can add dynamic ones if needed. The file
9911 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9912 updated whenever a new extension is added to the core code and kept
9913 in ext_nid order. There is a simple program 'tabtest.c' which checks
9914 this. New extensions are not added too often so this file can readily
9915 be maintained manually.
9917 There are two big advantages in doing things this way. The extensions
9918 can be looked up immediately and no longer need to be "added" using
9919 X509V3_add_standard_extensions(): this function now does nothing.
9920 [Side note: I get *lots* of email saying the extension code doesn't
9921 work because people forget to call this function]
9922 Also no dynamic allocation is done unless new extensions are added:
9923 so if we don't add custom extensions there is no need to call
9924 X509V3_EXT_cleanup().
9927 *) Modify enc utility's salting as follows: make salting the default. Add a
9928 magic header, so unsalted files fail gracefully instead of just decrypting
9929 to garbage. This is because not salting is a big security hole, so people
9930 should be discouraged from doing it.
9933 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9934 digest to be passed on the command line but it only used this
9935 parameter when signing a certificate. Modified so all relevant
9936 operations are affected by the digest parameter including the
9937 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9938 DSA key was used because it didn't fix the digest.
9941 *) Initial certificate chain verify code. Currently tests the untrusted
9942 certificates for consistency with the verify purpose (which is set
9943 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9945 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9946 this is because it will reject chains with invalid extensions whereas
9947 every previous version of OpenSSL and SSLeay made no checks at all.
9949 Trust code: checks the root CA for the relevant trust settings. Trust
9950 settings have an initial value consistent with the verify purpose: e.g.
9951 if the verify purpose is for SSL client use it expects the CA to be
9952 trusted for SSL client use. However the default value can be changed to
9953 permit custom trust settings: one example of this would be to only trust
9954 certificates from a specific "secure" set of CAs.
9956 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9957 which should be used for version portability: especially since the
9958 verify structure is likely to change more often now.
9960 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9961 to set them. If not set then assume SSL clients will verify SSL servers
9964 Two new options to the verify program: -untrusted allows a set of
9965 untrusted certificates to be passed in and -purpose which sets the
9966 intended purpose of the certificate. If a purpose is set then the
9967 new chain verify code is used to check extension consistency.
9970 *) Support for the authority information access extension.
9973 *) Modify RSA and DSA PEM read routines to transparently handle
9974 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9975 public keys in a format compatible with certificate
9976 SubjectPublicKeyInfo structures. Unfortunately there were already
9977 functions called *_PublicKey_* which used various odd formats so
9978 these are retained for compatibility: however the DSA variants were
9979 never in a public release so they have been deleted. Changed dsa/rsa
9980 utilities to handle the new format: note no releases ever handled public
9981 keys so we should be OK.
9983 The primary motivation for this change is to avoid the same fiasco
9984 that dogs private keys: there are several incompatible private key
9985 formats some of which are standard and some OpenSSL specific and
9986 require various evil hacks to allow partial transparent handling and
9987 even then it doesn't work with DER formats. Given the option anything
9988 other than PKCS#8 should be dumped: but the other formats have to
9989 stay in the name of compatibility.
9991 With public keys and the benefit of hindsight one standard format
9992 is used which works with EVP_PKEY, RSA or DSA structures: though
9993 it clearly returns an error if you try to read the wrong kind of key.
9995 Added a -pubkey option to the 'x509' utility to output the public key.
9996 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9997 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9998 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9999 that do the same as the EVP_PKEY_assign_*() except they up the
10000 reference count of the added key (they don't "swallow" the
10004 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10005 CRLs would fail if the file contained no certificates or no CRLs:
10006 added a new function to read in both types and return the number
10007 read: this means that if none are read it will be an error. The
10008 DER versions of the certificate and CRL reader would always fail
10009 because it isn't possible to mix certificates and CRLs in DER format
10010 without choking one or the other routine. Changed this to just read
10011 a certificate: this is the best we can do. Also modified the code
10012 in apps/verify.c to take notice of return codes: it was previously
10013 attempting to read in certificates from NULL pointers and ignoring
10014 any errors: this is one reason why the cert and CRL reader seemed
10015 to work. It doesn't check return codes from the default certificate
10016 routines: these may well fail if the certificates aren't installed.
10019 *) Code to support otherName option in GeneralName.
10022 *) First update to verify code. Change the verify utility
10023 so it warns if it is passed a self signed certificate:
10024 for consistency with the normal behaviour. X509_verify
10025 has been modified to it will now verify a self signed
10026 certificate if *exactly* the same certificate appears
10027 in the store: it was previously impossible to trust a
10028 single self signed certificate. This means that:
10029 openssl verify ss.pem
10030 now gives a warning about a self signed certificate but
10031 openssl verify -CAfile ss.pem ss.pem
10035 *) For servers, store verify_result in SSL_SESSION data structure
10036 (and add it to external session representation).
10037 This is needed when client certificate verifications fails,
10038 but an application-provided verification callback (set by
10039 SSL_CTX_set_cert_verify_callback) allows accepting the session
10040 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10041 but returns 1): When the session is reused, we have to set
10042 ssl->verify_result to the appropriate error code to avoid
10044 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10046 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10047 case in PKCS7_dataInit() where the signed PKCS7 structure
10048 didn't contain any existing data because it was being created.
10049 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10051 *) Add a salt to the key derivation routines in enc.c. This
10052 forms the first 8 bytes of the encrypted file. Also add a
10053 -S option to allow a salt to be input on the command line.
10056 *) New function X509_cmp(). Oddly enough there wasn't a function
10057 to compare two certificates. We do this by working out the SHA1
10058 hash and comparing that. X509_cmp() will be needed by the trust
10062 *) SSL_get1_session() is like SSL_get_session(), but increments
10063 the reference count in the SSL_SESSION returned.
10064 [Geoff Thorpe <geoff@eu.c2.net>]
10066 *) Fix for 'req': it was adding a null to request attributes.
10067 Also change the X509_LOOKUP and X509_INFO code to handle
10068 certificate auxiliary information.
10071 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10075 *) Add the possibility to add extra information to the memory leak
10076 detecting output, to form tracebacks, showing from where each
10077 allocation was originated: CRYPTO_push_info("constant string") adds
10078 the string plus current file name and line number to a per-thread
10079 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10080 is like calling CYRPTO_pop_info() until the stack is empty.
10081 Also updated memory leak detection code to be multi-thread-safe.
10084 *) Add options -text and -noout to pkcs7 utility and delete the
10085 encryption options which never did anything. Update docs.
10088 *) Add options to some of the utilities to allow the pass phrase
10089 to be included on either the command line (not recommended on
10090 OSes like Unix) or read from the environment. Update the
10091 manpages and fix a few bugs.
10094 *) Add a few manpages for some of the openssl commands.
10097 *) Fix the -revoke option in ca. It was freeing up memory twice,
10098 leaking and not finding already revoked certificates.
10101 *) Extensive changes to support certificate auxiliary information.
10102 This involves the use of X509_CERT_AUX structure and X509_AUX
10103 functions. An X509_AUX function such as PEM_read_X509_AUX()
10104 can still read in a certificate file in the usual way but it
10105 will also read in any additional "auxiliary information". By
10106 doing things this way a fair degree of compatibility can be
10107 retained: existing certificates can have this information added
10108 using the new 'x509' options.
10110 Current auxiliary information includes an "alias" and some trust
10111 settings. The trust settings will ultimately be used in enhanced
10112 certificate chain verification routines: currently a certificate
10113 can only be trusted if it is self signed and then it is trusted
10117 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10118 The problem was that one of the replacement routines had not been working
10119 since SSLeay releases. For now the offending routine has been replaced
10120 with non-optimised assembler. Even so, this now gives around 95%
10121 performance improvement for 1024 bit RSA signs.
10124 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10125 handling. Most clients have the effective key size in bits equal to
10126 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10127 A few however don't do this and instead use the size of the decrypted key
10128 to determine the RC2 key length and the AlgorithmIdentifier to determine
10129 the effective key length. In this case the effective key length can still
10130 be 40 bits but the key length can be 168 bits for example. This is fixed
10131 by manually forcing an RC2 key into the EVP_PKEY structure because the
10132 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10133 the key length and effective key length are equal.
10136 *) Add a bunch of functions that should simplify the creation of
10137 X509_NAME structures. Now you should be able to do:
10138 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10139 and have it automatically work out the correct field type and fill in
10140 the structures. The more adventurous can try:
10141 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10142 and it will (hopefully) work out the correct multibyte encoding.
10145 *) Change the 'req' utility to use the new field handling and multibyte
10146 copy routines. Before the DN field creation was handled in an ad hoc
10147 way in req, ca, and x509 which was rather broken and didn't support
10148 BMPStrings or UTF8Strings. Since some software doesn't implement
10149 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10150 using the dirstring_type option. See the new comment in the default
10151 openssl.cnf for more info.
10154 *) Make crypto/rand/md_rand.c more robust:
10155 - Assure unique random numbers after fork().
10156 - Make sure that concurrent threads access the global counter and
10157 md serializably so that we never lose entropy in them
10158 or use exactly the same state in multiple threads.
10159 Access to the large state is not always serializable because
10160 the additional locking could be a performance killer, and
10161 md should be large enough anyway.
10164 *) New file apps/app_rand.c with commonly needed functionality
10165 for handling the random seed file.
10167 Use the random seed file in some applications that previously did not:
10169 dsaparam -genkey (which also ignored its '-rand' option),
10172 x509 (when signing).
10173 Except on systems with /dev/urandom, it is crucial to have a random
10174 seed file at least for key creation, DSA signing, and for DH exchanges;
10175 for RSA signatures we could do without one.
10177 gendh and gendsa (unlike genrsa) used to read only the first byte
10178 of each file listed in the '-rand' option. The function as previously
10179 found in genrsa is now in app_rand.c and is used by all programs
10180 that support '-rand'.
10183 *) In RAND_write_file, use mode 0600 for creating files;
10184 don't just chmod when it may be too late.
10187 *) Report an error from X509_STORE_load_locations
10188 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10191 *) New function ASN1_mbstring_copy() this copies a string in either
10192 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10193 into an ASN1_STRING type. A mask of permissible types is passed
10194 and it chooses the "minimal" type to use or an error if not type
10198 *) Add function equivalents to the various macros in asn1.h. The old
10199 macros are retained with an M_ prefix. Code inside the library can
10200 use the M_ macros. External code (including the openssl utility)
10201 should *NOT* in order to be "shared library friendly".
10204 *) Add various functions that can check a certificate's extensions
10205 to see if it usable for various purposes such as SSL client,
10206 server or S/MIME and CAs of these types. This is currently
10207 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10208 verification. Also added a -purpose flag to x509 utility to
10209 print out all the purposes.
10212 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10216 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10217 for, obtain and decode and extension and obtain its critical flag.
10218 This allows all the necessary extension code to be handled in a
10219 single function call.
10222 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10223 platforms. See crypto/rc4/rc4_enc.c for further details.
10226 *) New -noout option to asn1parse. This causes no output to be produced
10227 its main use is when combined with -strparse and -out to extract data
10228 from a file (which may not be in ASN.1 format).
10231 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10232 when producing the local key id.
10233 [Richard Levitte <levitte@stacken.kth.se>]
10235 *) New option -dhparam in s_server. This allows a DH parameter file to be
10236 stated explicitly. If it is not stated then it tries the first server
10237 certificate file. The previous behaviour hard coded the filename
10241 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10242 a public key to be input or output. For example:
10243 openssl rsa -in key.pem -pubout -out pubkey.pem
10244 Also added necessary DSA public key functions to handle this.
10247 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10248 in the message. This was handled by allowing
10249 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10250 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10252 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10253 to the end of the strings whereas this didn't. This would cause problems
10254 if strings read with d2i_ASN1_bytes() were later modified.
10255 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10257 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10258 data and it contains EOF it will end up returning an error. This is
10259 caused by input 46 bytes long. The cause is due to the way base64
10260 BIOs find the start of base64 encoded data. They do this by trying a
10261 trial decode on each line until they find one that works. When they
10262 do a flag is set and it starts again knowing it can pass all the
10263 data directly through the decoder. Unfortunately it doesn't reset
10264 the context it uses. This means that if EOF is reached an attempt
10265 is made to pass two EOFs through the context and this causes the
10266 resulting error. This can also cause other problems as well. As is
10267 usual with these problems it takes *ages* to find and the fix is
10268 trivial: move one line.
10269 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10271 *) Ugly workaround to get s_client and s_server working under Windows. The
10272 old code wouldn't work because it needed to select() on sockets and the
10273 tty (for keypresses and to see if data could be written). Win32 only
10274 supports select() on sockets so we select() with a 1s timeout on the
10275 sockets and then see if any characters are waiting to be read, if none
10276 are present then we retry, we also assume we can always write data to
10277 the tty. This isn't nice because the code then blocks until we've
10278 received a complete line of data and it is effectively polling the
10279 keyboard at 1s intervals: however it's quite a bit better than not
10280 working at all :-) A dedicated Windows application might handle this
10281 with an event loop for example.
10284 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10285 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10286 will be called when RSA_sign() and RSA_verify() are used. This is useful
10287 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10288 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10289 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10290 This necessitated the support of an extra signature type NID_md5_sha1
10291 for SSL signatures and modifications to the SSL library to use it instead
10292 of calling RSA_public_decrypt() and RSA_private_encrypt().
10295 *) Add new -verify -CAfile and -CApath options to the crl program, these
10296 will lookup a CRL issuers certificate and verify the signature in a
10297 similar way to the verify program. Tidy up the crl program so it
10298 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10299 less strict. It will now permit CRL extensions even if it is not
10300 a V2 CRL: this will allow it to tolerate some broken CRLs.
10303 *) Initialize all non-automatic variables each time one of the openssl
10304 sub-programs is started (this is necessary as they may be started
10305 multiple times from the "OpenSSL>" prompt).
10306 [Lennart Bang, Bodo Moeller]
10308 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10309 removing all other RSA functionality (this is what NO_RSA does). This
10310 is so (for example) those in the US can disable those operations covered
10311 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10315 *) Non-copying interface to BIO pairs.
10316 (still largely untested)
10319 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10320 ASCII string. This was handled independently in various places before.
10323 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10324 UTF8 strings a character at a time.
10327 *) Use client_version from client hello to select the protocol
10328 (s23_srvr.c) and for RSA client key exchange verification
10329 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10332 *) Add various utility functions to handle SPKACs, these were previously
10333 handled by poking round in the structure internals. Added new function
10334 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10335 print, verify and generate SPKACs. Based on an original idea from
10336 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10339 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10342 *) Allow the config file extension section to be overwritten on the
10343 command line. Based on an original idea from Massimiliano Pala
10344 <madwolf@comune.modena.it>. The new option is called -extensions
10345 and can be applied to ca, req and x509. Also -reqexts to override
10346 the request extensions in req and -crlexts to override the crl extensions
10350 *) Add new feature to the SPKAC handling in ca. Now you can include
10351 the same field multiple times by preceding it by "XXXX." for example:
10354 this is the same syntax as used in the req config file.
10357 *) Allow certificate extensions to be added to certificate requests. These
10358 are specified in a 'req_extensions' option of the req section of the
10359 config file. They can be printed out with the -text option to req but
10360 are otherwise ignored at present.
10363 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10364 data read consists of only the final block it would not decrypted because
10365 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10366 A misplaced 'break' also meant the decrypted final block might not be
10367 copied until the next read.
10370 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10371 a few extra parameters to the DH structure: these will be useful if
10372 for example we want the value of 'q' or implement X9.42 DH.
10375 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10376 provides hooks that allow the default DSA functions or functions on a
10377 "per key" basis to be replaced. This allows hardware acceleration and
10378 hardware key storage to be handled without major modification to the
10379 library. Also added low level modexp hooks and CRYPTO_EX structure and
10380 associated functions.
10383 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10384 as "read only": it can't be written to and the buffer it points to will
10385 not be freed. Reading from a read only BIO is much more efficient than
10386 a normal memory BIO. This was added because there are several times when
10387 an area of memory needs to be read from a BIO. The previous method was
10388 to create a memory BIO and write the data to it, this results in two
10389 copies of the data and an O(n^2) reading algorithm. There is a new
10390 function BIO_new_mem_buf() which creates a read only memory BIO from
10391 an area of memory. Also modified the PKCS#7 routines to use read only
10395 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10396 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10397 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10398 but a retry condition occurred while trying to read the rest.
10401 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10402 NID_pkcs7_encrypted by default: this was wrong since this should almost
10403 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10404 the encrypted data type: this is a more sensible place to put it and it
10405 allows the PKCS#12 code to be tidied up that duplicated this
10409 *) Changed obj_dat.pl script so it takes its input and output files on
10410 the command line. This should avoid shell escape redirection problems
10414 *) Initial support for certificate extension requests, these are included
10415 in things like Xenroll certificate requests. Included functions to allow
10416 extensions to be obtained and added.
10419 *) -crlf option to s_client and s_server for sending newlines as
10420 CRLF (as required by many protocols).
10423 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10425 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10426 [Ralf S. Engelschall]
10428 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10429 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10431 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10435 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10436 DH parameters/keys (q is lost during that conversion, but the resulting
10437 DH parameters contain its length).
10439 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10440 much faster than DH_generate_parameters (which creates parameters
10441 where p = 2*q + 1), and also the smaller q makes DH computations
10442 much more efficient (160-bit exponentiation instead of 1024-bit
10443 exponentiation); so this provides a convenient way to support DHE
10444 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10445 utter importance to use
10446 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10448 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10449 when such DH parameters are used, because otherwise small subgroup
10450 attacks may become possible!
10453 *) Avoid memory leak in i2d_DHparams.
10456 *) Allow the -k option to be used more than once in the enc program:
10457 this allows the same encrypted message to be read by multiple recipients.
10460 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10461 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10462 it will always use the numerical form of the OID, even if it has a short
10466 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10467 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10468 otherwise bn_mod_exp was called. In the case of hardware keys for example
10469 no private key components need be present and it might store extra data
10470 in the RSA structure, which cannot be accessed from bn_mod_exp.
10471 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10472 private key operations.
10475 *) Added support for SPARC Linux.
10478 *) pem_password_cb function type incompatibly changed from
10479 typedef int pem_password_cb(char *buf, int size, int rwflag);
10481 ....(char *buf, int size, int rwflag, void *userdata);
10482 so that applications can pass data to their callbacks:
10483 The PEM[_ASN1]_{read,write}... functions and macros now take an
10484 additional void * argument, which is just handed through whenever
10485 the password callback is called.
10486 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10488 New function SSL_CTX_set_default_passwd_cb_userdata.
10490 Compatibility note: As many C implementations push function arguments
10491 onto the stack in reverse order, the new library version is likely to
10492 interoperate with programs that have been compiled with the old
10493 pem_password_cb definition (PEM_whatever takes some data that
10494 happens to be on the stack as its last argument, and the callback
10495 just ignores this garbage); but there is no guarantee whatsoever that
10498 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10499 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10500 problems not only on Windows, but also on some Unix platforms.
10501 To avoid problematic command lines, these definitions are now in an
10502 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10503 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10506 *) MIPS III/IV assembler module is reimplemented.
10509 *) More DES library cleanups: remove references to srand/rand and
10510 delete an unused file.
10513 *) Add support for the the free Netwide assembler (NASM) under Win32,
10514 since not many people have MASM (ml) and it can be hard to obtain.
10515 This is currently experimental but it seems to work OK and pass all
10516 the tests. Check out INSTALL.W32 for info.
10519 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10520 without temporary keys kept an extra copy of the server key,
10521 and connections with temporary keys did not free everything in case
10525 *) New function RSA_check_key and new openssl rsa option -check
10526 for verifying the consistency of RSA keys.
10527 [Ulf Moeller, Bodo Moeller]
10529 *) Various changes to make Win32 compile work:
10530 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10531 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10532 comparison" warnings.
10533 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10536 *) Add a debugging option to PKCS#5 v2 key generation function: when
10537 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10538 derived keys are printed to stderr.
10541 *) Copy the flags in ASN1_STRING_dup().
10542 [Roman E. Pavlov <pre@mo.msk.ru>]
10544 *) The x509 application mishandled signing requests containing DSA
10545 keys when the signing key was also DSA and the parameters didn't match.
10547 It was supposed to omit the parameters when they matched the signing key:
10548 the verifying software was then supposed to automatically use the CA's
10549 parameters if they were absent from the end user certificate.
10551 Omitting parameters is no longer recommended. The test was also
10552 the wrong way round! This was probably due to unusual behaviour in
10553 EVP_cmp_parameters() which returns 1 if the parameters match.
10554 This meant that parameters were omitted when they *didn't* match and
10555 the certificate was useless. Certificates signed with 'ca' didn't have
10557 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10559 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10560 The interface is as follows:
10561 Applications can use
10562 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10563 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10564 "off" is now the default.
10565 The library internally uses
10566 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10567 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10568 to disable memory-checking temporarily.
10570 Some inconsistent states that previously were possible (and were
10571 even the default) are now avoided.
10573 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10574 with each memory chunk allocated; this is occasionally more helpful
10575 than just having a counter.
10577 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10579 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10583 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10584 which largely parallels "options", but is for changing API behaviour,
10585 whereas "options" are about protocol behaviour.
10586 Initial "mode" flags are:
10588 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10589 a single record has been written.
10590 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10591 retries use the same buffer location.
10592 (But all of the contents must be
10596 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10599 *) Fix problems with no-hmac etc.
10600 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10602 *) New functions RSA_get_default_method(), RSA_set_method() and
10603 RSA_get_method(). These allows replacement of RSA_METHODs without having
10604 to mess around with the internals of an RSA structure.
10607 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10608 Also really enable memory leak checks in openssl.c and in some
10610 [Chad C. Mulligan, Bodo Moeller]
10612 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10613 up the length of negative integers. This has now been simplified to just
10614 store the length when it is first determined and use it later, rather
10615 than trying to keep track of where data is copied and updating it to
10617 [Steve Henson, reported by Brien Wheeler
10618 <bwheeler@authentica-security.com>]
10620 *) Add a new function PKCS7_signatureVerify. This allows the verification
10621 of a PKCS#7 signature but with the signing certificate passed to the
10622 function itself. This contrasts with PKCS7_dataVerify which assumes the
10623 certificate is present in the PKCS#7 structure. This isn't always the
10624 case: certificates can be omitted from a PKCS#7 structure and be
10625 distributed by "out of band" means (such as a certificate database).
10628 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10629 function prototypes in pem.h, also change util/mkdef.pl to add the
10630 necessary function names.
10633 *) mk1mf.pl (used by Windows builds) did not properly read the
10634 options set by Configure in the top level Makefile, and Configure
10635 was not even able to write more than one option correctly.
10636 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10639 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10640 file to be loaded from a BIO or FILE pointer. The BIO version will
10641 for example allow memory BIOs to contain config info.
10644 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10645 Whoever hopes to achieve shared-library compatibility across versions
10646 must use this, not the compile-time macro.
10647 (Exercise 0.9.4: Which is the minimum library version required by
10649 Note: All this applies only to multi-threaded programs, others don't
10653 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10654 through a BIO pair triggered the default case, i.e.
10655 SSLerr(...,SSL_R_UNKNOWN_STATE).
10658 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10659 can use the SSL library even if none of the specific BIOs is
10663 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10664 for the encoded length.
10665 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10667 *) Add initial documentation of the X509V3 functions.
10670 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10671 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10672 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10673 secure PKCS#8 private key format with a high iteration count.
10676 *) Fix determination of Perl interpreter: A perl or perl5
10677 _directory_ in $PATH was also accepted as the interpreter.
10678 [Ralf S. Engelschall]
10680 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10681 wrong with it but it was very old and did things like calling
10682 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10683 unusual formatting.
10686 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10687 to use the new extension code.
10690 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10691 with macros. This should make it easier to change their form, add extra
10692 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10696 *) Add to configuration table a new entry that can specify an alternative
10697 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10698 according to Mark Crispin <MRC@Panda.COM>.
10702 *) DES CBC did not update the IV. Weird.
10705 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10706 Changing the behaviour of the former might break existing programs --
10707 where IV updating is needed, des_ncbc_encrypt can be used.
10710 *) When bntest is run from "make test" it drives bc to check its
10711 calculations, as well as internally checking them. If an internal check
10712 fails, it needs to cause bc to give a non-zero result or make test carries
10713 on without noticing the failure. Fixed.
10716 *) DES library cleanups.
10719 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10720 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10721 ciphers. NOTE: although the key derivation function has been verified
10722 against some published test vectors it has not been extensively tested
10723 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10727 *) Instead of "mkdir -p", which is not fully portable, use new
10728 Perl script "util/mkdir-p.pl".
10731 *) Rewrite the way password based encryption (PBE) is handled. It used to
10732 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10733 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10734 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10735 the 'parameter' field of the AlgorithmIdentifier is passed to the
10736 underlying key generation function so it must do its own ASN1 parsing.
10737 This has also changed the EVP_PBE_CipherInit() function which now has a
10738 'parameter' argument instead of literal salt and iteration count values
10739 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10742 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10743 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10744 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10745 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10746 value was just used as a "magic string" and not used directly its
10747 value doesn't matter.
10750 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10754 *) "linux-sparc64" configuration (ultrapenguin).
10755 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10756 "linux-sparc" configuration.
10757 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10759 *) config now generates no-xxx options for missing ciphers.
10762 *) Support the EBCDIC character set (work in progress).
10763 File ebcdic.c not yet included because it has a different license.
10764 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10766 *) Support BS2000/OSD-POSIX.
10767 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10769 *) Make callbacks for key generation use void * instead of char *.
10772 *) Make S/MIME samples compile (not yet tested).
10775 *) Additional typesafe stacks.
10778 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10782 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10784 *) New configuration variant "sco5-gcc".
10786 *) Updated some demos.
10787 [Sean O Riordain, Wade Scholine]
10789 *) Add missing BIO_free at exit of pkcs12 application.
10792 *) Fix memory leak in conf.c.
10795 *) Updates for Win32 to assembler version of MD5.
10798 *) Set #! path to perl in apps/der_chop to where we found it
10799 instead of using a fixed path.
10802 *) SHA library changes for irix64-mips4-cc.
10805 *) Improvements for VMS support.
10809 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10811 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10812 This also avoids the problems with SC4.2 and unpatched SC5.
10813 [Andy Polyakov <appro@fy.chalmers.se>]
10815 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10816 These are required because of the typesafe stack would otherwise break
10817 existing code. If old code used a structure member which used to be STACK
10818 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10819 sk_num or sk_value it would produce an error because the num, data members
10820 are not present in STACK_OF. Now it just produces a warning. sk_set
10821 replaces the old method of assigning a value to sk_value
10822 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10823 that does this will no longer work (and should use sk_set instead) but
10824 this could be regarded as a "questionable" behaviour anyway.
10827 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10828 correctly handle encrypted S/MIME data.
10831 *) Change type of various DES function arguments from des_cblock
10832 (which means, in function argument declarations, pointer to char)
10833 to des_cblock * (meaning pointer to array with 8 char elements),
10834 which allows the compiler to do more typechecking; it was like
10835 that back in SSLeay, but with lots of ugly casts.
10837 Introduce new type const_des_cblock.
10840 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10841 problems: find RecipientInfo structure that matches recipient certificate
10842 and initialise the ASN1 structures properly based on passed cipher.
10845 *) Belatedly make the BN tests actually check the results.
10848 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10849 to and from BNs: it was completely broken. New compilation option
10850 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10851 key elements as negative integers.
10854 *) Reorganize and speed up MD5.
10855 [Andy Polyakov <appro@fy.chalmers.se>]
10858 [Richard Levitte <richard@levitte.org>]
10860 *) New option -out to asn1parse to allow the parsed structure to be
10861 output to a file. This is most useful when combined with the -strparse
10862 option to examine the output of things like OCTET STRINGS.
10865 *) Make SSL library a little more fool-proof by not requiring any longer
10866 that SSL_set_{accept,connect}_state be called before
10867 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10868 in many applications because usually everything *appeared* to work as
10869 intended anyway -- now it really works as intended).
10872 *) Move openssl.cnf out of lib/.
10875 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10876 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10877 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10878 [Ralf S. Engelschall]
10880 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10881 handle PKCS#7 enveloped data properly.
10882 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10884 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10885 copying pointers. The cert_st handling is changed by this in
10886 various ways (and thus what used to be known as ctx->default_cert
10887 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10888 any longer when s->cert does not give us what we need).
10889 ssl_cert_instantiate becomes obsolete by this change.
10890 As soon as we've got the new code right (possibly it already is?),
10891 we have solved a couple of bugs of the earlier code where s->cert
10892 was used as if it could not have been shared with other SSL structures.
10894 Note that using the SSL API in certain dirty ways now will result
10895 in different behaviour than observed with earlier library versions:
10896 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10897 does not influence s as it used to.
10899 In order to clean up things more thoroughly, inside SSL_SESSION
10900 we don't use CERT any longer, but a new structure SESS_CERT
10901 that holds per-session data (if available); currently, this is
10902 the peer's certificate chain and, for clients, the server's certificate
10903 and temporary key. CERT holds only those values that can have
10904 meaningful defaults in an SSL_CTX.
10907 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10908 from the internal representation. Various PKCS#7 fixes: remove some
10909 evil casts and set the enc_dig_alg field properly based on the signing
10913 *) Allow PKCS#12 password to be set from the command line or the
10914 environment. Let 'ca' get its config file name from the environment
10915 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10919 *) Allow certificate policies extension to use an IA5STRING for the
10920 organization field. This is contrary to the PKIX definition but
10921 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10925 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10926 without disallowing inline assembler and the like for non-pedantic builds.
10929 *) Support Borland C++ builder.
10930 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10932 *) Support Mingw32.
10935 *) SHA-1 cleanups and performance enhancements.
10936 [Andy Polyakov <appro@fy.chalmers.se>]
10938 *) Sparc v8plus assembler for the bignum library.
10939 [Andy Polyakov <appro@fy.chalmers.se>]
10941 *) Accept any -xxx and +xxx compiler options in Configure.
10944 *) Update HPUX configuration.
10947 *) Add missing sk_<type>_unshift() function to safestack.h
10948 [Ralf S. Engelschall]
10950 *) New function SSL_CTX_use_certificate_chain_file that sets the
10951 "extra_cert"s in addition to the certificate. (This makes sense
10952 only for "PEM" format files, as chains as a whole are not
10956 *) Support verify_depth from the SSL API.
10957 x509_vfy.c had what can be considered an off-by-one-error:
10958 Its depth (which was not part of the external interface)
10959 was actually counting the number of certificates in a chain;
10960 now it really counts the depth.
10963 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10964 instead of X509err, which often resulted in confusing error
10965 messages since the error codes are not globally unique
10966 (e.g. an alleged error in ssl3_accept when a certificate
10967 didn't match the private key).
10969 *) New function SSL_CTX_set_session_id_context that allows to set a default
10970 value (so that you don't need SSL_set_session_id_context for each
10971 connection using the SSL_CTX).
10974 *) OAEP decoding bug fix.
10977 *) Support INSTALL_PREFIX for package builders, as proposed by
10981 *) New Configure options "threads" and "no-threads". For systems
10982 where the proper compiler options are known (currently Solaris
10983 and Linux), "threads" is the default.
10986 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10989 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10990 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10991 such as /usr/local/bin.
10994 *) "make linux-shared" to build shared libraries.
10995 [Niels Poppe <niels@netbox.org>]
10997 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11000 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11001 extension adding in x509 utility.
11004 *) Remove NOPROTO sections and error code comments.
11007 *) Partial rewrite of the DEF file generator to now parse the ANSI
11011 *) New Configure options --prefix=DIR and --openssldir=DIR.
11014 *) Complete rewrite of the error code script(s). It is all now handled
11015 by one script at the top level which handles error code gathering,
11016 header rewriting and C source file generation. It should be much better
11017 than the old method: it now uses a modified version of Ulf's parser to
11018 read the ANSI prototypes in all header files (thus the old K&R definitions
11019 aren't needed for error creation any more) and do a better job of
11020 translating function codes into names. The old 'ASN1 error code imbedded
11021 in a comment' is no longer necessary and it doesn't use .err files which
11022 have now been deleted. Also the error code call doesn't have to appear all
11023 on one line (which resulted in some large lines...).
11026 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11029 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11030 0 (which usually indicates a closed connection), but continue reading.
11033 *) Fix some race conditions.
11036 *) Add support for CRL distribution points extension. Add Certificate
11037 Policies and CRL distribution points documentation.
11040 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11043 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11044 8 of keying material. Merlin has also confirmed interop with this fix
11045 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11046 [Merlin Hughes <merlin@baltimore.ie>]
11048 *) Fix lots of warnings.
11049 [Richard Levitte <levitte@stacken.kth.se>]
11051 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11052 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11053 [Richard Levitte <levitte@stacken.kth.se>]
11055 *) Fix problems with sizeof(long) == 8.
11056 [Andy Polyakov <appro@fy.chalmers.se>]
11058 *) Change functions to ANSI C.
11061 *) Fix typos in error codes.
11062 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11064 *) Remove defunct assembler files from Configure.
11067 *) SPARC v8 assembler BIGNUM implementation.
11068 [Andy Polyakov <appro@fy.chalmers.se>]
11070 *) Support for Certificate Policies extension: both print and set.
11071 Various additions to support the r2i method this uses.
11074 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11075 return a const string when you are expecting an allocated buffer.
11078 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11079 types DirectoryString and DisplayText.
11082 *) Add code to allow r2i extensions to access the configuration database,
11083 add an LHASH database driver and add several ctx helper functions.
11086 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11087 fail when they extended the size of a BIGNUM.
11090 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11091 support typesafe stack.
11094 *) Fix typo in SSL_[gs]et_options().
11095 [Nils Frostberg <nils@medcom.se>]
11097 *) Delete various functions and files that belonged to the (now obsolete)
11098 old X509V3 handling code.
11101 *) New Configure option "rsaref".
11104 *) Don't auto-generate pem.h.
11107 *) Introduce type-safe ASN.1 SETs.
11110 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11111 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11113 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11114 that links with OpenSSL (well at least cause lots of warnings), but fear
11115 not: the conversion is trivial, and it eliminates loads of evil casts. A
11116 few STACKed things have been converted already. Feel free to convert more.
11117 In the fullness of time, I'll do away with the STACK type altogether.
11120 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11121 specified in <certfile> by updating the entry in the index.txt file.
11122 This way one no longer has to edit the index.txt file manually for
11123 revoking a certificate. The -revoke option does the gory details now.
11124 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11126 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11127 `-text' option at all and this way the `-noout -text' combination was
11128 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11129 [Ralf S. Engelschall]
11131 *) Make sure a corresponding plain text error message exists for the
11132 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11133 verify callback function determined that a certificate was revoked.
11134 [Ralf S. Engelschall]
11136 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11137 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11138 all available cipers including rc5, which was forgotten until now.
11139 In order to let the testing shell script know which algorithms
11140 are available, a new (up to now undocumented) command
11141 "openssl list-cipher-commands" is used.
11144 *) Bugfix: s_client occasionally would sleep in select() when
11145 it should have checked SSL_pending() first.
11148 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11149 the raw DSA values prior to ASN.1 encoding.
11152 *) Tweaks to Configure
11153 [Niels Poppe <niels@netbox.org>]
11155 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11159 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11162 *) New config option to avoid instructions that are illegal on the 80386.
11163 The default code is faster, but requires at least a 486.
11166 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11167 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11168 same as SSL2_VERSION anyway.
11171 *) New "-showcerts" option for s_client.
11174 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11175 application. Various cleanups and fixes.
11178 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11179 modify error routines to work internally. Add error codes and PBE init
11180 to library startup routines.
11183 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11184 packing functions to asn1 and evp. Changed function names and error
11185 codes along the way.
11188 *) PKCS12 integration: and so it begins... First of several patches to
11189 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11190 objects to objects.h
11193 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11194 and display support for Thawte strong extranet extension.
11197 *) Add LinuxPPC support.
11198 [Jeff Dubrule <igor@pobox.org>]
11200 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11201 bn_div_words in alpha.s.
11202 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11204 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11205 OAEP isn't supported when OpenSSL is built with RSAref.
11206 [Ulf Moeller <ulf@fitug.de>]
11208 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11209 so they no longer are missing under -DNOPROTO.
11210 [Soren S. Jorvang <soren@t.dk>]
11213 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11215 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11216 doesn't work when the session is reused. Coming soon!
11219 *) Fix a security hole, that allows sessions to be reused in the wrong
11220 context thus bypassing client cert protection! All software that uses
11221 client certs and session caches in multiple contexts NEEDS PATCHING to
11222 allow session reuse! A fuller solution is in the works.
11223 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11225 *) Some more source tree cleanups (removed obsolete files
11226 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11227 permission on "config" script to be executable) and a fix for the INSTALL
11229 [Ulf Moeller <ulf@fitug.de>]
11231 *) Remove some legacy and erroneous uses of malloc, free instead of
11233 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11235 *) Make rsa_oaep_test return non-zero on error.
11236 [Ulf Moeller <ulf@fitug.de>]
11238 *) Add support for native Solaris shared libraries. Configure
11239 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11240 if someone would make that last step automatic.
11241 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11243 *) ctx_size was not built with the right compiler during "make links". Fixed.
11246 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11247 except NULL ciphers". This means the default cipher list will no longer
11248 enable NULL ciphers. They need to be specifically enabled e.g. with
11249 the string "DEFAULT:eNULL".
11252 *) Fix to RSA private encryption routines: if p < q then it would
11253 occasionally produce an invalid result. This will only happen with
11254 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11257 *) Be less restrictive and allow also `perl util/perlpath.pl
11258 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11259 because this way one can also use an interpreter named `perl5' (which is
11260 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11261 installed as `perl').
11262 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11264 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11265 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11267 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11268 advapi32.lib to Win32 build and change the pem test comparison
11269 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11270 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11271 and crypto/des/ede_cbcm_enc.c.
11274 *) DES quad checksum was broken on big-endian architectures. Fixed.
11277 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11278 Win32 test batch file so it (might) work again. The Win32 test batch file
11279 is horrible: I feel ill....
11282 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11283 in e_os.h. Audit of header files to check ANSI and non ANSI
11284 sections: 10 functions were absent from non ANSI section and not exported
11285 from Windows DLLs. Fixed up libeay.num for new functions.
11288 *) Make `openssl version' output lines consistent.
11289 [Ralf S. Engelschall]
11291 *) Fix Win32 symbol export lists for BIO functions: Added
11292 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11293 to ms/libeay{16,32}.def.
11294 [Ralf S. Engelschall]
11296 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11297 fine under Unix and passes some trivial tests I've now added. But the
11298 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11299 added to make sure no one expects that this stuff really works in the
11300 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11301 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11303 [Ralf S. Engelschall]
11305 *) Fix the generation of two part addresses in perl.
11306 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11308 *) Add config entry for Linux on MIPS.
11309 [John Tobey <jtobey@channel1.com>]
11311 *) Make links whenever Configure is run, unless we are on Windoze.
11314 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11315 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11319 *) Add a useful kludge to allow package maintainers to specify compiler and
11320 other platforms details on the command line without having to patch the
11321 Configure script everytime: One now can use ``perl Configure
11322 <id>:<details>'', i.e. platform ids are allowed to have details appended
11323 to them (separated by colons). This is treated as there would be a static
11324 pre-configured entry in Configure's %table under key <id> with value
11325 <details> and ``perl Configure <id>'' is called. So, when you want to
11326 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11327 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11328 now, which overrides the FreeBSD-elf entry on-the-fly.
11329 [Ralf S. Engelschall]
11331 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11334 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11335 on the `perl Configure ...' command line. This way one can compile
11336 OpenSSL libraries with Position Independent Code (PIC) which is needed
11337 for linking it into DSOs.
11338 [Ralf S. Engelschall]
11340 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11344 *) Cleaned up the LICENSE document: The official contact for any license
11345 questions now is the OpenSSL core team under openssl-core@openssl.org.
11346 And add a paragraph about the dual-license situation to make sure people
11347 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11348 to the OpenSSL toolkit.
11349 [Ralf S. Engelschall]
11351 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11352 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11353 Additionally cleaned up the `make links' target: Remove unnecessary
11354 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11355 to speed processing and no longer clutter the display with confusing
11356 stuff. Instead only the actually done links are displayed.
11357 [Ralf S. Engelschall]
11359 *) Permit null encryption ciphersuites, used for authentication only. It used
11360 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11361 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11365 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11366 signed attributes when verifying signatures (this would break them),
11367 the detached data encoding was wrong and public keys obtained using
11368 X509_get_pubkey() weren't freed.
11371 *) Add text documentation for the BUFFER functions. Also added a work around
11372 to a Win95 console bug. This was triggered by the password read stuff: the
11373 last character typed gets carried over to the next fread(). If you were
11374 generating a new cert request using 'req' for example then the last
11375 character of the passphrase would be CR which would then enter the first
11379 *) Added the new `Includes OpenSSL Cryptography Software' button as
11380 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11381 button and can be used by applications based on OpenSSL to show the
11382 relationship to the OpenSSL project.
11383 [Ralf S. Engelschall]
11385 *) Remove confusing variables in function signatures in files
11386 ssl/ssl_lib.c and ssl/ssl.h.
11387 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11389 *) Don't install bss_file.c under PREFIX/include/
11390 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11392 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11393 functions that return function pointers and has support for NT specific
11394 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11395 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11396 unsigned to signed types: this was killing the Win32 compile.
11399 *) Add new certificate file to stack functions,
11400 SSL_add_dir_cert_subjects_to_stack() and
11401 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11402 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11403 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11404 This means that Apache-SSL and similar packages don't have to mess around
11405 to add as many CAs as they want to the preferred list.
11408 *) Experiment with doxygen documentation. Currently only partially applied to
11410 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11411 openssl.doxy as the configuration file.
11414 *) Get rid of remaining C++-style comments which strict C compilers hate.
11415 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11417 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11418 compiled in by default: it has problems with large keys.
11421 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11422 DH private keys and/or callback functions which directly correspond to
11423 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11424 is needed for applications which have to configure certificates on a
11425 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11427 For the RSA certificate situation is makes no difference, but
11428 for the DSA certificate situation this fixes the "no shared cipher"
11429 problem where the OpenSSL cipher selection procedure failed because the
11430 temporary keys were not overtaken from the context and the API provided
11431 no way to reconfigure them.
11432 The new functions now let applications reconfigure the stuff and they
11433 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11434 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11435 non-public-API function ssl_cert_instantiate() is used as a helper
11436 function and also to reduce code redundancy inside ssl_rsa.c.
11437 [Ralf S. Engelschall]
11439 *) Move s_server -dcert and -dkey options out of the undocumented feature
11440 area because they are useful for the DSA situation and should be
11441 recognized by the users.
11442 [Ralf S. Engelschall]
11444 *) Fix the cipher decision scheme for export ciphers: the export bits are
11445 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11446 SSL_EXP_MASK. So, the original variable has to be used instead of the
11447 already masked variable.
11448 [Richard Levitte <levitte@stacken.kth.se>]
11450 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11451 [Richard Levitte <levitte@stacken.kth.se>]
11453 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11454 from `int' to `unsigned int' because it's a length and initialized by
11455 EVP_DigestFinal() which expects an `unsigned int *'.
11456 [Richard Levitte <levitte@stacken.kth.se>]
11458 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11459 script. Instead use the usual Shell->Perl transition trick.
11460 [Ralf S. Engelschall]
11462 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11463 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11464 -noout -modulus' as it's already the case for `openssl rsa -noout
11465 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11466 currently the public key is printed (a decision which was already done by
11467 `openssl dsa -modulus' in the past) which serves a similar purpose.
11468 Additionally the NO_RSA no longer completely removes the whole -modulus
11469 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11471 [Ralf S. Engelschall]
11473 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11474 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11475 [Arne Ansper <arne@ats.cyber.ee>]
11477 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11478 to be added. Now both 'req' and 'ca' can use new objects defined in the
11482 *) Add cool BIO that does syslog (or event log on NT).
11483 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11485 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11486 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11487 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11488 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11491 *) Add preliminary config info for new extension code.
11494 *) Make RSA_NO_PADDING really use no padding.
11495 [Ulf Moeller <ulf@fitug.de>]
11497 *) Generate errors when private/public key check is done.
11500 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11501 for some CRL extensions and new objects added.
11504 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11505 key usage extension and fuller support for authority key id.
11508 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11509 padding method for RSA, which is recommended for new applications in PKCS
11510 #1 v2.0 (RFC 2437, October 1998).
11511 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11512 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11513 against Bleichbacher's attack on RSA.
11514 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11517 *) Updates to the new SSL compression code
11518 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11520 *) Fix so that the version number in the master secret, when passed
11521 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11522 (because the server will not accept higher), that the version number
11523 is 0x03,0x01, not 0x03,0x00
11524 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11526 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11527 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11528 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11531 *) Support for RAW extensions where an arbitrary extension can be
11532 created by including its DER encoding. See apps/openssl.cnf for
11536 *) Make sure latest Perl versions don't interpret some generated C array
11537 code as Perl array code in the crypto/err/err_genc.pl script.
11538 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11540 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11541 not many people have the assembler. Various Win32 compilation fixes and
11542 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11543 build instructions.
11546 *) Modify configure script 'Configure' to automatically create crypto/date.h
11547 file under Win32 and also build pem.h from pem.org. New script
11548 util/mkfiles.pl to create the MINFO file on environments that can't do a
11549 'make files': perl util/mkfiles.pl >MINFO should work.
11552 *) Major rework of DES function declarations, in the pursuit of correctness
11553 and purity. As a result, many evil casts evaporated, and some weirdness,
11554 too. You may find this causes warnings in your code. Zapping your evil
11555 casts will probably fix them. Mostly.
11558 *) Fix for a typo in asn1.h. Bug fix to object creation script
11559 obj_dat.pl. It considered a zero in an object definition to mean
11560 "end of object": none of the objects in objects.h have any zeros
11561 so it wasn't spotted.
11562 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11564 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11565 Masking (CBCM). In the absence of test vectors, the best I have been able
11566 to do is check that the decrypt undoes the encrypt, so far. Send me test
11567 vectors if you have them.
11570 *) Correct calculation of key length for export ciphers (too much space was
11571 allocated for null ciphers). This has not been tested!
11574 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11575 message is now correct (it understands "crypto" and "ssl" on its
11576 command line). There is also now an "update" option. This will update
11577 the util/ssleay.num and util/libeay.num files with any new functions.
11579 perl util/mkdef.pl crypto ssl update
11580 it will update them.
11583 *) Overhauled the Perl interface (perl/*):
11584 - ported BN stuff to OpenSSL's different BN library
11585 - made the perl/ source tree CVS-aware
11586 - renamed the package from SSLeay to OpenSSL (the files still contain
11587 their history because I've copied them in the repository)
11588 - removed obsolete files (the test scripts will be replaced
11589 by better Test::Harness variants in the future)
11590 [Ralf S. Engelschall]
11592 *) First cut for a very conservative source tree cleanup:
11593 1. merge various obsolete readme texts into doc/ssleay.txt
11594 where we collect the old documents and readme texts.
11595 2. remove the first part of files where I'm already sure that we no
11596 longer need them because of three reasons: either they are just temporary
11597 files which were left by Eric or they are preserved original files where
11598 I've verified that the diff is also available in the CVS via "cvs diff
11599 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11600 the crypto/md/ stuff).
11601 [Ralf S. Engelschall]
11603 *) More extension code. Incomplete support for subject and issuer alt
11604 name, issuer and authority key id. Change the i2v function parameters
11605 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11606 what that's for :-) Fix to ASN1 macro which messed up
11607 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11610 *) Preliminary support for ENUMERATED type. This is largely copied from the
11614 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11615 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11617 *) Make sure `make rehash' target really finds the `openssl' program.
11618 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11620 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11621 like to hear about it if this slows down other processors.
11624 *) Add CygWin32 platform information to Configure script.
11625 [Alan Batie <batie@aahz.jf.intel.com>]
11627 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11628 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11630 *) New program nseq to manipulate netscape certificate sequences
11633 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11637 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11638 but the BN code had some problems that would cause failures when
11639 doing certificate verification and some other functions.
11640 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11642 *) Add ASN1 and PEM code to support netscape certificate sequences.
11645 *) Add ASN1 and PEM code to support netscape certificate sequences.
11648 *) Add several PKIX and private extended key usage OIDs.
11651 *) Modify the 'ca' program to handle the new extension code. Modify
11652 openssl.cnf for new extension format, add comments.
11655 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11656 and add a sample to openssl.cnf so req -x509 now adds appropriate
11660 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11661 error code, add initial support to X509_print() and x509 application.
11664 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11665 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11666 stuff is currently isolated and isn't even compiled yet.
11669 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11670 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11671 Removed the versions check from X509 routines when loading extensions:
11672 this allows certain broken certificates that don't set the version
11673 properly to be processed.
11676 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11677 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11678 can still be regenerated with "make depend".
11681 *) Spelling mistake in C version of CAST-128.
11682 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11684 *) Changes to the error generation code. The perl script err-code.pl
11685 now reads in the old error codes and retains the old numbers, only
11686 adding new ones if necessary. It also only changes the .err files if new
11687 codes are added. The makefiles have been modified to only insert errors
11688 when needed (to avoid needlessly modifying header files). This is done
11689 by only inserting errors if the .err file is newer than the auto generated
11690 C file. To rebuild all the error codes from scratch (the old behaviour)
11691 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11692 or delete all the .err files.
11695 *) CAST-128 was incorrectly implemented for short keys. The C version has
11696 been fixed, but is untested. The assembler versions are also fixed, but
11697 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11698 to regenerate it if needed.
11699 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11700 Hagino <itojun@kame.net>]
11702 *) File was opened incorrectly in randfile.c.
11703 [Ulf Möller <ulf@fitug.de>]
11705 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11706 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11707 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11708 al: it's just almost always a UTCTime. Note this patch adds new error
11709 codes so do a "make errors" if there are problems.
11712 *) Correct Linux 1 recognition in config.
11713 [Ulf Möller <ulf@fitug.de>]
11715 *) Remove pointless MD5 hash when using DSA keys in ca.
11716 [Anonymous <nobody@replay.com>]
11718 *) Generate an error if given an empty string as a cert directory. Also
11719 generate an error if handed NULL (previously returned 0 to indicate an
11720 error, but didn't set one).
11721 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11723 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11726 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11727 parameters. This was causing a warning which killed off the Win32 compile.
11730 *) Remove C++ style comments from crypto/bn/bn_local.h.
11731 [Neil Costigan <neil.costigan@celocom.com>]
11733 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11734 based on a text string, looking up short and long names and finally
11735 "dot" format. The "dot" format stuff didn't work. Added new function
11736 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11737 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11738 OID is not part of the table.
11741 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11742 X509_LOOKUP_by_alias().
11745 *) Sort openssl functions by name.
11748 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11749 encryption from sample DSA keys (in case anyone is interested the password
11753 *) Make _all_ *_free functions accept a NULL pointer.
11754 [Frans Heymans <fheymans@isaserver.be>]
11756 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11758 [Anonymous <nobody@replay.com>]
11760 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11761 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11763 *) Don't blow it for numeric -newkey arguments to apps/req.
11764 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11766 *) Temp key "for export" tests were wrong in s3_srvr.c.
11767 [Anonymous <nobody@replay.com>]
11769 *) Add prototype for temp key callback functions
11770 SSL_CTX_set_tmp_{rsa,dh}_callback().
11773 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11774 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11777 *) X509_name_add_entry() freed the wrong thing after an error.
11778 [Arne Ansper <arne@ats.cyber.ee>]
11780 *) rsa_eay.c would attempt to free a NULL context.
11781 [Arne Ansper <arne@ats.cyber.ee>]
11783 *) BIO_s_socket() had a broken should_retry() on Windoze.
11784 [Arne Ansper <arne@ats.cyber.ee>]
11786 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11787 [Arne Ansper <arne@ats.cyber.ee>]
11789 *) Make sure the already existing X509_STORE->depth variable is initialized
11790 in X509_STORE_new(), but document the fact that this variable is still
11791 unused in the certificate verification process.
11792 [Ralf S. Engelschall]
11794 *) Fix the various library and apps files to free up pkeys obtained from
11795 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11798 *) Fix reference counting in X509_PUBKEY_get(). This makes
11799 demos/maurice/example2.c work, amongst others, probably.
11800 [Steve Henson and Ben Laurie]
11802 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11803 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11804 are no longer created. This way we have a single and consistent command
11805 line interface `openssl <command>', similar to `cvs <command>'.
11806 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11808 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11809 BIT STRING wrapper always have zero unused bits.
11812 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11815 *) Make the top-level INSTALL documentation easier to understand.
11818 *) Makefiles updated to exit if an error occurs in a sub-directory
11819 make (including if user presses ^C) [Paul Sutton]
11821 *) Make Montgomery context stuff explicit in RSA data structure.
11824 *) Fix build order of pem and err to allow for generated pem.h.
11827 *) Fix renumbering bug in X509_NAME_delete_entry().
11830 *) Enhanced the err-ins.pl script so it makes the error library number
11831 global and can add a library name. This is needed for external ASN1 and
11832 other error libraries.
11835 *) Fixed sk_insert which never worked properly.
11838 *) Fix ASN1 macros so they can handle indefinite length constructed
11839 EXPLICIT tags. Some non standard certificates use these: they can now
11843 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11844 into a single doc/ssleay.txt bundle. This way the information is still
11845 preserved but no longer messes up this directory. Now it's new room for
11846 the new set of documentation files.
11847 [Ralf S. Engelschall]
11849 *) SETs were incorrectly DER encoded. This was a major pain, because they
11850 shared code with SEQUENCEs, which aren't coded the same. This means that
11851 almost everything to do with SETs or SEQUENCEs has either changed name or
11852 number of arguments.
11853 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11855 *) Fix test data to work with the above.
11858 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11859 was already fixed by Eric for 0.9.1 it seems.
11860 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11862 *) Autodetect FreeBSD3.
11865 *) Fix various bugs in Configure. This affects the following platforms:
11869 unixware-2.0-pentium
11873 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11874 before they are needed.
11877 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11881 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11883 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11884 changed SSLeay to OpenSSL in version strings.
11885 [Ralf S. Engelschall]
11887 *) Some fixups to the top-level documents.
11890 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11891 because the symlink to include/ was missing.
11892 [Ralf S. Engelschall]
11894 *) Incorporated the popular no-RSA/DSA-only patches
11895 which allow to compile a RSA-free SSLeay.
11896 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11898 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11899 when "ssleay" is still not found.
11900 [Ralf S. Engelschall]
11902 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11903 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11905 *) Updated the README file.
11906 [Ralf S. Engelschall]
11908 *) Added various .cvsignore files in the CVS repository subdirs
11909 to make a "cvs update" really silent.
11910 [Ralf S. Engelschall]
11912 *) Recompiled the error-definition header files and added
11913 missing symbols to the Win32 linker tables.
11914 [Ralf S. Engelschall]
11916 *) Cleaned up the top-level documents;
11917 o new files: CHANGES and LICENSE
11918 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11919 o merged COPYRIGHT into LICENSE
11920 o removed obsolete TODO file
11921 o renamed MICROSOFT to INSTALL.W32
11922 [Ralf S. Engelschall]
11924 *) Removed dummy files from the 0.9.1b source tree:
11925 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11926 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11927 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11928 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11929 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11930 [Ralf S. Engelschall]
11932 *) Added various platform portability fixes.
11935 *) The Genesis of the OpenSSL rpject:
11936 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11937 Young and Tim J. Hudson created while they were working for C2Net until
11939 [The OpenSSL Project]
11942 Changes between 0.9.0b and 0.9.1b [not released]
11944 *) Updated a few CA certificates under certs/
11947 *) Changed some BIGNUM api stuff.
11950 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11951 DGUX x86, Linux Alpha, etc.
11954 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11955 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11959 *) Add -strparse option to asn1pars program which parses nested
11961 [Dr Stephen Henson <shenson@bigfoot.com>]
11963 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11966 *) DSA fix for "ca" program.
11969 *) Added "-genkey" option to "dsaparam" program.
11972 *) Added RIPE MD160 (rmd160) message digest.
11975 *) Added -a (all) option to "ssleay version" command.
11978 *) Added PLATFORM define which is the id given to Configure.
11981 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11984 *) Extended the ASN.1 parser routines.
11987 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11990 *) Added a BN_CTX to the BN library.
11993 *) Fixed the weak key values in DES library
11996 *) Changed API in EVP library for cipher aliases.
11999 *) Added support for RC2/64bit cipher.
12002 *) Converted the lhash library to the crypto/mem.c functions.
12005 *) Added more recognized ASN.1 object ids.
12008 *) Added more RSA padding checks for SSL/TLS.
12011 *) Added BIO proxy/filter functionality.
12014 *) Added extra_certs to SSL_CTX which can be used
12015 send extra CA certificates to the client in the CA cert chain sending
12016 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12019 *) Now Fortezza is denied in the authentication phase because
12020 this is key exchange mechanism is not supported by SSLeay at all.
12023 *) Additional PKCS1 checks.
12026 *) Support the string "TLSv1" for all TLS v1 ciphers.
12029 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12030 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12033 *) Fixed a few memory leaks.
12036 *) Fixed various code and comment typos.
12039 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12040 bytes sent in the client random.
12041 [Edward Bishop <ebishop@spyglass.com>]