5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Changes to the DEFAULT cipherlist:
8 - Prefer (EC)DHE handshakes over plain RSA.
9 - Prefer AEAD ciphers over legacy ciphers.
10 - Prefer ECDSA over RSA when both certificates are available.
11 - Prefer TLSv1.2 ciphers/PRF.
12 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
16 *) Change the ECC default curve list to be this, in order: x25519,
17 secp256r1, secp521r1, secp384r1.
20 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
21 disabled by default. They can be re-enabled using the
22 enable-weak-ssl-ciphers option to Configure.
25 *) If the server has ALPN configured, but supports no protocols that the
26 client advertises, send a fatal "no_application_protocol" alert.
27 This behaviour is SHALL in RFC 7301, though it isn't universally
28 implemented by other servers.
31 *) Add X25519 support.
32 Integrate support for X25519 into EC library. This includes support
33 for public and private key encoding using the format documented in
34 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
35 OID from that draft, encodes public keys using little endian
36 format in the ECPoint structure and private keys using
37 little endian form in the privateKey field of the ECPrivateKey
38 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
41 Note: the current version supports key generation, public and
42 private key encoding and ECDH key agreement using the EC API.
43 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
47 *) Deprecate SRP_VBASE_get_by_user.
48 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
49 In order to fix an unavoidable memory leak (CVE-2016-0798),
50 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
51 seed, even if the seed is configured.
53 Users should use SRP_VBASE_get1_by_user instead. Note that in
54 SRP_VBASE_get1_by_user, caller must free the returned value. Note
55 also that even though configuring the SRP seed attempts to hide
56 invalid usernames by continuing the handshake with fake
57 credentials, this behaviour is not constant time and no strong
58 guarantees are made that the handshake is indistinguishable from
62 *) Configuration change; it's now possible to build dynamic engines
63 without having to build shared libraries and vice versa. This
64 only applies to the engines in engines/, those in crypto/engine/
65 will always be built into libcrypto (i.e. "static").
67 Building dynamic engines is enabled by default; to disable, use
68 the configuration option "disable-dynamic-engine".
70 The only requirements for building dynamic engines are the
71 presence of the DSO module and building with position independent
72 code, so they will also automatically be disabled if configuring
73 with "disable-dso" or "disable-pic".
75 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
76 are also taken away from openssl/opensslconf.h, as they are
80 *) Configuration change; if there is a known flag to compile
81 position independent code, it will always be applied on the
82 libcrypto and libssl object files, and never on the application
83 object files. This means other libraries that use routines from
84 libcrypto / libssl can be made into shared libraries regardless
85 of how OpenSSL was configured.
87 If this isn't desirable, the configuration options "disable-pic"
88 or "no-pic" can be used to disable the use of PIC. This will
89 also disable building shared libraries and dynamic engines.
92 *) Removed JPAKE code. It was experimental and has no wide use.
95 *) The INSTALL_PREFIX Makefile variable has been renamed to
96 DESTDIR. That makes for less confusion on what this variable
97 is for. Also, the configuration option --install_prefix is
101 *) Heartbeat for TLS has been removed and is disabled by default
102 for DTLS; configure with enable-heartbeats. Code that uses the
103 old #define's might need to be updated.
104 [Emilia Käsper, Rich Salz]
106 *) Rename REF_CHECK to REF_DEBUG.
109 *) New "unified" build system
111 The "unified" build system is aimed to be a common system for all
112 platforms we support. With it comes new support for VMS.
114 This system builds supports building in a different directory tree
115 than the source tree. It produces one Makefile (for unix family
116 or lookalikes), or one descrip.mms (for VMS).
118 The source of information to make the Makefile / descrip.mms is
119 small files called 'build.info', holding the necessary
120 information for each directory with source to compile, and a
121 template in Configurations, like unix-Makefile.tmpl or
124 We rely heavily on the perl module Text::Template.
127 *) Added support for auto-initialisation and de-initialisation of the library.
128 OpenSSL no longer requires explicit init or deinit routines to be called,
129 except in certain circumstances. See the OPENSSL_init_crypto() and
130 OPENSSL_init_ssl() man pages for further information.
133 *) The arguments to the DTLSv1_listen function have changed. Specifically the
134 "peer" argument is now expected to be a BIO_ADDR object.
136 *) Rewrite of BIO networking library. The BIO library lacked consistent
137 support of IPv6, and adding it required some more extensive
138 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
139 which hold all types of addresses and chains of address information.
140 It also introduces a new API, with functions like BIO_socket,
141 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
142 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
143 have been adapted accordingly.
146 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
150 *) CRIME protection: disable compression by default, even if OpenSSL is
151 compiled with zlib enabled. Applications can still enable compression
152 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
153 using the SSL_CONF library to configure compression.
156 *) The signature of the session callback configured with
157 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
158 was explicitly marked as 'const unsigned char*' instead of
162 *) Always DPURIFY. Remove the use of uninitialized memory in the
163 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
166 *) Removed many obsolete configuration items, including
167 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
168 MD2_CHAR, MD2_INT, MD2_LONG
170 IDEA_SHORT, IDEA_LONG
171 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
172 [Rich Salz, with advice from Andy Polyakov]
174 *) Many BN internals have been moved to an internal header file.
175 [Rich Salz with help from Andy Polyakov]
177 *) Configuration and writing out the results from it has changed.
178 Files such as Makefile include/openssl/opensslconf.h and are now
179 produced through general templates, such as Makefile.in and
180 crypto/opensslconf.h.in and some help from the perl module
183 Also, the center of configuration information is no longer
184 Makefile. Instead, Configure produces a perl module in
185 configdata.pm which holds most of the config data (in the hash
186 table %config), the target data that comes from the target
187 configuration in one of the Configurations/*.conf files (in
191 *) To clarify their intended purposes, the Configure options
192 --prefix and --openssldir change their semantics, and become more
193 straightforward and less interdependent.
195 --prefix shall be used exclusively to give the location INSTALLTOP
196 where programs, scripts, libraries, include files and manuals are
197 going to be installed. The default is now /usr/local.
199 --openssldir shall be used exclusively to give the default
200 location OPENSSLDIR where certificates, private keys, CRLs are
201 managed. This is also where the default openssl.cnf gets
203 If the directory given with this option is a relative path, the
204 values of both the --prefix value and the --openssldir value will
205 be combined to become OPENSSLDIR.
206 The default for --openssldir is INSTALLTOP/ssl.
208 Anyone who uses --openssldir to specify where OpenSSL is to be
209 installed MUST change to use --prefix instead.
212 *) The GOST engine was out of date and therefore it has been removed. An up
213 to date GOST engine is now being maintained in an external repository.
214 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
215 support for GOST ciphersuites (these are only activated if a GOST engine
219 *) EGD is no longer supported by default; use enable-egd when
221 [Ben Kaduk and Rich Salz]
223 *) The distribution now has Makefile.in files, which are used to
224 create Makefile's when Configure is run. *Configure must be run
225 before trying to build now.*
228 *) The return value for SSL_CIPHER_description() for error conditions
232 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
234 Obtaining and performing DNSSEC validation of TLSA records is
235 the application's responsibility. The application provides
236 the TLSA records of its choice to OpenSSL, and these are then
237 used to authenticate the peer.
239 The TLSA records need not even come from DNS. They can, for
240 example, be used to implement local end-entity certificate or
241 trust-anchor "pinning", where the "pin" data takes the form
242 of TLSA records, which can augment or replace verification
243 based on the usual WebPKI public certification authorities.
246 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
247 continues to support deprecated interfaces in default builds.
248 However, applications are strongly advised to compile their
249 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
250 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
251 or the 1.1.0 releases.
253 In environments in which all applications have been ported to
254 not use any deprecated interfaces OpenSSL's Configure script
255 should be used with the --api=1.1.0 option to entirely remove
256 support for the deprecated features from the library and
257 unconditionally disable them in the installed headers.
258 Essentially the same effect can be achieved with the "no-deprecated"
259 argument to Configure, except that this will always restrict
260 the build to just the latest API, rather than a fixed API
263 As applications are ported to future revisions of the API,
264 they should update their compile-time OPENSSL_API_COMPAT define
265 accordingly, but in most cases should be able to continue to
266 compile with later releases.
268 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
269 0x10000000L and 0x00908000L, respectively. However those
270 versions did not support the OPENSSL_API_COMPAT feature, and
271 so applications are not typically tested for explicit support
272 of just the undeprecated features of either release.
275 *) Add support for setting the minimum and maximum supported protocol.
276 It can bet set via the SSL_set_min_proto_version() and
277 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
278 MaxProtcol. It's recommended to use the new APIs to disable
279 protocols instead of disabling individual protocols using
280 SSL_set_options() or SSL_CONF's Protocol. This change also
281 removes support for disabling TLS 1.2 in the OpenSSL TLS
282 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
285 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
288 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
289 and integrates ECDSA and ECDH functionality into EC. Implementations can
290 now redirect key generation and no longer need to convert to or from
293 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
294 include the ec.h header file instead.
297 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
298 ciphers who are no longer supported and drops support the ephemeral RSA key
299 exchange. The LOW ciphers currently doesn't have any ciphers in it.
302 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
303 opaque. For HMAC_CTX, the following constructors and destructors
306 HMAC_CTX *HMAC_CTX_new(void);
307 void HMAC_CTX_free(HMAC_CTX *ctx);
309 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
310 destroy such methods has been added. See EVP_MD_meth_new(3) and
311 EVP_CIPHER_meth_new(3) for documentation.
314 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
315 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
316 EVP_MD_CTX_reset() should be called instead to reinitialise
317 an already created structure.
318 2) For consistency with the majority of our object creators and
319 destructors, EVP_MD_CTX_(create|destroy) were renamed to
320 EVP_MD_CTX_(new|free). The old names are retained as macros
321 for deprecated builds.
324 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
325 cryptographic operations to be performed asynchronously as long as an
326 asynchronous capable engine is used. See the ASYNC_start_job() man page for
327 further details. Libssl has also had this capability integrated with the
328 introduction of the new mode SSL_MODE_ASYNC and associated error
329 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
330 pages. This work was developed in partnership with Intel Corp.
333 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
334 always enabled now. If you want to disable the support you should
335 exclude it using the list of supported ciphers. This also means that the
336 "-no_ecdhe" option has been removed from s_server.
339 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
340 SSL_{CTX_}set1_curves() which can set a list.
343 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
344 curve you want to support using SSL_{CTX_}set1_curves().
347 *) State machine rewrite. The state machine code has been significantly
348 refactored in order to remove much duplication of code and solve issues
349 with the old code (see ssl/statem/README for further details). This change
350 does have some associated API changes. Notably the SSL_state() function
351 has been removed and replaced by SSL_get_state which now returns an
352 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
353 altogether. The previous handshake states defined in ssl.h and ssl3.h have
357 *) All instances of the string "ssleay" in the public API were replaced
358 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
359 Some error codes related to internal RSA_eay API's were renamed.
362 *) The demo files in crypto/threads were moved to demo/threads.
365 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
367 [Matt Caswell, Rich Salz]
369 *) New ASN.1 embed macro.
371 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
372 structure is not allocated: it is part of the parent. That is instead of
380 This reduces memory fragmentation and make it impossible to accidentally
381 set a mandatory field to NULL.
383 This currently only works for some fields specifically a SEQUENCE, CHOICE,
384 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
385 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
389 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
392 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
393 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
394 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
395 DES and RC4 ciphersuites.
398 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
399 This changes the decoding behaviour for some invalid messages,
400 though the change is mostly in the more lenient direction, and
401 legacy behaviour is preserved as much as possible.
404 *) Fix no-stdio build.
405 [ David Woodhouse <David.Woodhouse@intel.com> and also
406 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
408 *) New testing framework
409 The testing framework has been largely rewritten and is now using
410 perl and the perl modules Test::Harness and an extended variant of
411 Test::More called OpenSSL::Test to do its work. All test scripts in
412 test/ have been rewritten into test recipes, and all direct calls to
413 executables in test/Makefile have become individual recipes using the
414 simplified testing OpenSSL::Test::Simple.
416 For documentation on our testing modules, do:
418 perldoc test/testlib/OpenSSL/Test/Simple.pm
419 perldoc test/testlib/OpenSSL/Test.pm
423 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
424 are used; the latter aborts on memory leaks (usually checked on exit).
425 Some undocumented "set malloc, etc., hooks" functions were removed
426 and others were changed. All are now documented.
429 *) In DSA_generate_parameters_ex, if the provided seed is too short,
431 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
433 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
434 from RFC4279, RFC4785, RFC5487, RFC5489.
436 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
437 original RSA_PSK patch.
440 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
441 era flag was never set throughout the codebase (only read). Also removed
442 SSL3_FLAGS_POP_BUFFER which was only used if
443 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
446 *) Changed the default name options in the "ca", "crl", "req" and "x509"
447 to be "oneline" instead of "compat".
450 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
451 not aware of clients that still exhibit this bug, and the workaround
452 hasn't been working properly for a while.
455 *) The return type of BIO_number_read() and BIO_number_written() as well as
456 the corresponding num_read and num_write members in the BIO structure has
457 changed from unsigned long to uint64_t. On platforms where an unsigned
458 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
462 *) Given the pervasive nature of TLS extensions it is inadvisable to run
463 OpenSSL without support for them. It also means that maintaining
464 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
465 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
468 *) Removed support for the two export grade static DH ciphersuites
469 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
470 were newly added (along with a number of other static DH ciphersuites) to
471 1.0.2. However the two export ones have *never* worked since they were
472 introduced. It seems strange in any case to be adding new export
473 ciphersuites, and given "logjam" it also does not seem correct to fix them.
476 *) Version negotiation has been rewritten. In particular SSLv23_method(),
477 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
478 and turned into macros which simply call the new preferred function names
479 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
480 should use the new names instead. Also as part of this change the ssl23.h
481 header file has been removed.
484 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
485 code and the associated standard is no longer considered fit-for-purpose.
488 *) RT2547 was closed. When generating a private key, try to make the
489 output file readable only by the owner. This behavior change might
490 be noticeable when interacting with other software.
492 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
496 *) Added HTTP GET support to the ocsp command.
499 *) Changed default digest for the dgst and enc commands from MD5 to
503 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
506 *) Added support for TLS extended master secret from
507 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
508 initial patch which was a great help during development.
511 *) All libssl internal structures have been removed from the public header
512 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
513 now redundant). Users should not attempt to access internal structures
514 directly. Instead they should use the provided API functions.
517 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
518 Access to deprecated functions can be re-enabled by running config with
519 "enable-deprecated". In addition applications wishing to use deprecated
520 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
521 will, by default, disable some transitive includes that previously existed
522 in the header files (e.g. ec.h will no longer, by default, include bn.h)
525 *) Added support for OCB mode. OpenSSL has been granted a patent license
526 compatible with the OpenSSL license for use of OCB. Details are available
527 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
528 for OCB can be removed by calling config with no-ocb.
531 *) SSLv2 support has been removed. It still supports receiving a SSLv2
532 compatible client hello.
535 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
536 done while fixing the error code for the key-too-small case.
537 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
539 *) CA.sh has been removmed; use CA.pl instead.
542 *) Removed old DES API.
545 *) Remove various unsupported platforms:
551 Sinix/ReliantUNIX RM400
556 16-bit platforms such as WIN16
559 *) Clean up OPENSSL_NO_xxx #define's
560 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
561 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
562 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
563 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
564 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
565 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
566 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
567 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
568 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
569 Remove MS_STATIC; it's a relic from platforms <32 bits.
572 *) Cleaned up dead code
573 Remove all but one '#ifdef undef' which is to be looked at.
576 *) Clean up calling of xxx_free routines.
577 Just like free(), fix most of the xxx_free routines to accept
578 NULL. Remove the non-null checks from callers. Save much code.
581 *) Add secure heap for storage of private keys (when possible).
582 Add BIO_s_secmem(), CBIGNUM, etc.
583 Contributed by Akamai Technologies under our Corporate CLA.
586 *) Experimental support for a new, fast, unbiased prime candidate generator,
587 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
588 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
590 *) New output format NSS in the sess_id command line tool. This allows
591 exporting the session id and the master key in NSS keylog format.
592 [Martin Kaiser <martin@kaiser.cx>]
594 *) Harmonize version and its documentation. -f flag is used to display
596 [mancha <mancha1@zoho.com>]
598 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
599 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
600 [mancha <mancha1@zoho.com>]
602 *) Fix some double frees. These are not thought to be exploitable.
603 [mancha <mancha1@zoho.com>]
605 *) A missing bounds check in the handling of the TLS heartbeat extension
606 can be used to reveal up to 64k of memory to a connected client or
609 Thanks for Neel Mehta of Google Security for discovering this bug and to
610 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
611 preparing the fix (CVE-2014-0160)
612 [Adam Langley, Bodo Moeller]
614 *) Fix for the attack described in the paper "Recovering OpenSSL
615 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
616 by Yuval Yarom and Naomi Benger. Details can be obtained from:
617 http://eprint.iacr.org/2014/140
619 Thanks to Yuval Yarom and Naomi Benger for discovering this
620 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
621 [Yuval Yarom and Naomi Benger]
623 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
624 this fixes a limitation in previous versions of OpenSSL.
627 *) Experimental encrypt-then-mac support.
629 Experimental support for encrypt then mac from
630 draft-gutmann-tls-encrypt-then-mac-02.txt
632 To enable it set the appropriate extension number (0x42 for the test
633 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
635 For non-compliant peers (i.e. just about everything) this should have no
638 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
642 *) Add EVP support for key wrapping algorithms, to avoid problems with
643 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
644 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
645 algorithms and include tests cases.
648 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
652 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
653 MGF1 digest and OAEP label.
656 *) Make openssl verify return errors.
657 [Chris Palmer <palmer@google.com> and Ben Laurie]
659 *) New function ASN1_TIME_diff to calculate the difference between two
660 ASN1_TIME structures or one structure and the current time.
663 *) Update fips_test_suite to support multiple command line options. New
664 test to induce all self test errors in sequence and check expected
668 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
669 sign or verify all in one operation.
672 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
673 test programs and fips_test_suite. Includes functionality to parse
674 the minimal script output of fipsalgest.pl directly.
677 *) Add authorisation parameter to FIPS_module_mode_set().
680 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
683 *) Use separate DRBG fields for internal and external flags. New function
684 FIPS_drbg_health_check() to perform on demand health checking. Add
685 generation tests to fips_test_suite with reduced health check interval to
686 demonstrate periodic health checking. Add "nodh" option to
687 fips_test_suite to skip very slow DH test.
690 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
694 *) More extensive health check for DRBG checking many more failure modes.
695 New function FIPS_selftest_drbg_all() to handle every possible DRBG
696 combination: call this in fips_test_suite.
699 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
700 and POST to handle Dual EC cases.
703 *) Add support for canonical generation of DSA parameter 'g'. See
706 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
707 POST to handle HMAC cases.
710 *) Add functions FIPS_module_version() and FIPS_module_version_text()
711 to return numerical and string versions of the FIPS module number.
714 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
715 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
716 outside the validated module in the FIPS capable OpenSSL.
719 *) Minor change to DRBG entropy callback semantics. In some cases
720 there is no multiple of the block length between min_len and
721 max_len. Allow the callback to return more than max_len bytes
722 of entropy but discard any extra: it is the callback's responsibility
723 to ensure that the extra data discarded does not impact the
724 requested amount of entropy.
727 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
728 information in FIPS186-3, SP800-57 and SP800-131A.
731 *) CCM support via EVP. Interface is very similar to GCM case except we
732 must supply all data in one chunk (i.e. no update, final) and the
733 message length must be supplied if AAD is used. Add algorithm test
737 *) Initial version of POST overhaul. Add POST callback to allow the status
738 of POST to be monitored and/or failures induced. Modify fips_test_suite
739 to use callback. Always run all selftests even if one fails.
742 *) XTS support including algorithm test driver in the fips_gcmtest program.
743 Note: this does increase the maximum key length from 32 to 64 bytes but
744 there should be no binary compatibility issues as existing applications
745 will never use XTS mode.
748 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
749 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
750 performs algorithm blocking for unapproved PRNG types. Also do not
751 set PRNG type in FIPS_mode_set(): leave this to the application.
752 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
753 the standard OpenSSL PRNG: set additional data to a date time vector.
756 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
757 This shouldn't present any incompatibility problems because applications
758 shouldn't be using these directly and any that are will need to rethink
759 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
762 *) Extensive self tests and health checking required by SP800-90 DRBG.
763 Remove strength parameter from FIPS_drbg_instantiate and always
764 instantiate at maximum supported strength.
767 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
770 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
773 *) New function DH_compute_key_padded() to compute a DH key and pad with
774 leading zeroes if needed: this complies with SP800-56A et al.
777 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
778 anything, incomplete, subject to change and largely untested at present.
781 *) Modify fipscanisteronly build option to only build the necessary object
782 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
785 *) Add experimental option FIPSSYMS to give all symbols in
786 fipscanister.o and FIPS or fips prefix. This will avoid
787 conflicts with future versions of OpenSSL. Add perl script
788 util/fipsas.pl to preprocess assembly language source files
789 and rename any affected symbols.
792 *) Add selftest checks and algorithm block of non-fips algorithms in
793 FIPS mode. Remove DES2 from selftests.
796 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
797 return internal method without any ENGINE dependencies. Add new
798 tiny fips sign and verify functions.
801 *) New build option no-ec2m to disable characteristic 2 code.
804 *) New build option "fipscanisteronly". This only builds fipscanister.o
805 and (currently) associated fips utilities. Uses the file Makefile.fips
806 instead of Makefile.org as the prototype.
809 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
810 Update fips_gcmtest to use IV generator.
813 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
814 setting output buffer to NULL. The *Final function must be
815 called although it will not retrieve any additional data. The tag
816 can be set or retrieved with a ctrl. The IV length is by default 12
817 bytes (96 bits) but can be set to an alternative value. If the IV
818 length exceeds the maximum IV length (currently 16 bytes) it cannot be
822 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
823 underlying do_cipher function handles all cipher semantics itself
824 including padding and finalisation. This is useful if (for example)
825 an ENGINE cipher handles block padding itself. The behaviour of
826 do_cipher is subtly changed if this flag is set: the return value
827 is the number of characters written to the output buffer (zero is
828 no longer an error code) or a negative error code. Also if the
829 input buffer is NULL and length 0 finalisation should be performed.
832 *) If a candidate issuer certificate is already part of the constructed
833 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
836 *) Improve forward-security support: add functions
838 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
839 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
841 for use by SSL/TLS servers; the callback function will be called whenever a
842 new session is created, and gets to decide whether the session may be
843 cached to make it resumable (return 0) or not (return 1). (As by the
844 SSL/TLS protocol specifications, the session_id sent by the server will be
845 empty to indicate that the session is not resumable; also, the server will
846 not generate RFC 4507 (RFC 5077) session tickets.)
848 A simple reasonable callback implementation is to return is_forward_secure.
849 This parameter will be set to 1 or 0 depending on the ciphersuite selected
850 by the SSL/TLS server library, indicating whether it can provide forward
852 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
854 *) New -verify_name option in command line utilities to set verification
858 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
859 Add CMAC pkey methods.
862 *) Experimental renegotiation in s_server -www mode. If the client
863 browses /reneg connection is renegotiated. If /renegcert it is
864 renegotiated requesting a certificate.
867 *) Add an "external" session cache for debugging purposes to s_server. This
868 should help trace issues which normally are only apparent in deployed
869 multi-process servers.
872 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
873 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
874 BIO_set_cipher() and some obscure PEM functions were changed so they
875 can now return an error. The RAND changes required a change to the
876 RAND_METHOD structure.
879 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
880 a gcc attribute to warn if the result of a function is ignored. This
881 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
882 whose return value is often ignored.
885 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
886 These allow SCTs (signed certificate timestamps) to be requested and
887 validated when establishing a connection.
888 [Rob Percival <robpercival@google.com>]
890 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
892 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
893 Builds that are not configured with "enable-weak-ssl-ciphers" will not
894 provide any "EXPORT" or "LOW" strength ciphers.
897 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
898 is by default disabled at build-time. Builds that are not configured with
899 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
900 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
901 will need to explicitly call either of:
903 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
905 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
907 as appropriate. Even if either of those is used, or the application
908 explicitly uses the version-specific SSLv2_method() or its client and
909 server variants, SSLv2 ciphers vulnerable to exhaustive search key
910 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
911 ciphers, and SSLv2 56-bit DES are no longer available.
915 *) Fix a double-free in DSA code
917 A double free bug was discovered when OpenSSL parses malformed DSA private
918 keys and could lead to a DoS attack or memory corruption for applications
919 that receive DSA private keys from untrusted sources. This scenario is
922 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
927 *) Disable SRP fake user seed to address a server memory leak.
929 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
931 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
932 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
933 was changed to ignore the "fake user" SRP seed, even if the seed
936 Users should use SRP_VBASE_get1_by_user instead. Note that in
937 SRP_VBASE_get1_by_user, caller must free the returned value. Note
938 also that even though configuring the SRP seed attempts to hide
939 invalid usernames by continuing the handshake with fake
940 credentials, this behaviour is not constant time and no strong
941 guarantees are made that the handshake is indistinguishable from
942 that of a valid user.
946 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
948 In the BN_hex2bn function the number of hex digits is calculated using an
949 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
950 large values of |i| this can result in |bn_expand| not allocating any
951 memory because |i * 4| is negative. This can leave the internal BIGNUM data
952 field as NULL leading to a subsequent NULL ptr deref. For very large values
953 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
954 In this case memory is allocated to the internal BIGNUM data field, but it
955 is insufficiently sized leading to heap corruption. A similar issue exists
956 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
957 is ever called by user applications with very large untrusted hex/dec data.
958 This is anticipated to be a rare occurrence.
960 All OpenSSL internal usage of these functions use data that is not expected
961 to be untrusted, e.g. config file data or application command line
962 arguments. If user developed applications generate config file data based
963 on untrusted data then it is possible that this could also lead to security
964 consequences. This is also anticipated to be rare.
966 This issue was reported to OpenSSL by Guido Vranken.
970 *) Fix memory issues in BIO_*printf functions
972 The internal |fmtstr| function used in processing a "%s" format string in
973 the BIO_*printf functions could overflow while calculating the length of a
974 string and cause an OOB read when printing very long strings.
976 Additionally the internal |doapr_outch| function can attempt to write to an
977 OOB memory location (at an offset from the NULL pointer) in the event of a
978 memory allocation failure. In 1.0.2 and below this could be caused where
979 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
980 could be in processing a very long "%s" format string. Memory leaks can
983 The first issue may mask the second issue dependent on compiler behaviour.
984 These problems could enable attacks where large amounts of untrusted data
985 is passed to the BIO_*printf functions. If applications use these functions
986 in this way then they could be vulnerable. OpenSSL itself uses these
987 functions when printing out human-readable dumps of ASN.1 data. Therefore
988 applications that print this data could be vulnerable if the data is from
989 untrusted sources. OpenSSL command line applications could also be
990 vulnerable where they print out ASN.1 data, or if untrusted data is passed
991 as command line arguments.
993 Libssl is not considered directly vulnerable. Additionally certificates etc
994 received via remote connections via libssl are also unlikely to be able to
995 trigger these issues because of message size limits enforced within libssl.
997 This issue was reported to OpenSSL Guido Vranken.
1001 *) Side channel attack on modular exponentiation
1003 A side-channel attack was found which makes use of cache-bank conflicts on
1004 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1005 of RSA keys. The ability to exploit this issue is limited as it relies on
1006 an attacker who has control of code in a thread running on the same
1007 hyper-threaded core as the victim thread which is performing decryptions.
1009 This issue was reported to OpenSSL by Yuval Yarom, The University of
1010 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1011 Nadia Heninger, University of Pennsylvania with more information at
1012 http://cachebleed.info.
1016 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1017 if no keysize is specified with default_bits. This fixes an
1018 omission in an earlier change that changed all RSA/DSA key generation
1019 apps to use 2048 bits by default.
1022 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1023 *) DH small subgroups
1025 Historically OpenSSL only ever generated DH parameters based on "safe"
1026 primes. More recently (in version 1.0.2) support was provided for
1027 generating X9.42 style parameter files such as those required for RFC 5114
1028 support. The primes used in such files may not be "safe". Where an
1029 application is using DH configured with parameters based on primes that are
1030 not "safe" then an attacker could use this fact to find a peer's private
1031 DH exponent. This attack requires that the attacker complete multiple
1032 handshakes in which the peer uses the same private DH exponent. For example
1033 this could be used to discover a TLS server's private DH exponent if it's
1034 reusing the private DH exponent or it's using a static DH ciphersuite.
1036 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1037 TLS. It is not on by default. If the option is not set then the server
1038 reuses the same private DH exponent for the life of the server process and
1039 would be vulnerable to this attack. It is believed that many popular
1040 applications do set this option and would therefore not be at risk.
1042 The fix for this issue adds an additional check where a "q" parameter is
1043 available (as is the case in X9.42 based parameters). This detects the
1044 only known attack, and is the only possible defense for static DH
1045 ciphersuites. This could have some performance impact.
1047 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1048 default and cannot be disabled. This could have some performance impact.
1050 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1054 *) SSLv2 doesn't block disabled ciphers
1056 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1057 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1058 been disabled, provided that the SSLv2 protocol was not also disabled via
1061 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1062 and Sebastian Schinzel.
1066 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1068 *) BN_mod_exp may produce incorrect results on x86_64
1070 There is a carry propagating bug in the x86_64 Montgomery squaring
1071 procedure. No EC algorithms are affected. Analysis suggests that attacks
1072 against RSA and DSA as a result of this defect would be very difficult to
1073 perform and are not believed likely. Attacks against DH are considered just
1074 feasible (although very difficult) because most of the work necessary to
1075 deduce information about a private key may be performed offline. The amount
1076 of resources required for such an attack would be very significant and
1077 likely only accessible to a limited number of attackers. An attacker would
1078 additionally need online access to an unpatched system using the target
1079 private key in a scenario with persistent DH parameters and a private
1080 key that is shared between multiple clients. For example this can occur by
1081 default in OpenSSL DHE based SSL/TLS ciphersuites.
1083 This issue was reported to OpenSSL by Hanno Böck.
1087 *) Certificate verify crash with missing PSS parameter
1089 The signature verification routines will crash with a NULL pointer
1090 dereference if presented with an ASN.1 signature using the RSA PSS
1091 algorithm and absent mask generation function parameter. Since these
1092 routines are used to verify certificate signature algorithms this can be
1093 used to crash any certificate verification operation and exploited in a
1094 DoS attack. Any application which performs certificate verification is
1095 vulnerable including OpenSSL clients and servers which enable client
1098 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1102 *) X509_ATTRIBUTE memory leak
1104 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1105 memory. This structure is used by the PKCS#7 and CMS routines so any
1106 application which reads PKCS#7 or CMS data from untrusted sources is
1107 affected. SSL/TLS is not affected.
1109 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1114 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1115 This changes the decoding behaviour for some invalid messages,
1116 though the change is mostly in the more lenient direction, and
1117 legacy behaviour is preserved as much as possible.
1120 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1122 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1124 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1126 *) Alternate chains certificate forgery
1128 During certificate verfification, OpenSSL will attempt to find an
1129 alternative certificate chain if the first attempt to build such a chain
1130 fails. An error in the implementation of this logic can mean that an
1131 attacker could cause certain checks on untrusted certificates to be
1132 bypassed, such as the CA flag, enabling them to use a valid leaf
1133 certificate to act as a CA and "issue" an invalid certificate.
1135 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1139 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1141 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1142 incompatibility in the handling of HMAC. The previous ABI has now been
1146 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1148 *) Malformed ECParameters causes infinite loop
1150 When processing an ECParameters structure OpenSSL enters an infinite loop
1151 if the curve specified is over a specially malformed binary polynomial
1154 This can be used to perform denial of service against any
1155 system which processes public keys, certificate requests or
1156 certificates. This includes TLS clients and TLS servers with
1157 client authentication enabled.
1159 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1163 *) Exploitable out-of-bounds read in X509_cmp_time
1165 X509_cmp_time does not properly check the length of the ASN1_TIME
1166 string and can read a few bytes out of bounds. In addition,
1167 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1170 An attacker can use this to craft malformed certificates and CRLs of
1171 various sizes and potentially cause a segmentation fault, resulting in
1172 a DoS on applications that verify certificates or CRLs. TLS clients
1173 that verify CRLs are affected. TLS clients and servers with client
1174 authentication enabled may be affected if they use custom verification
1177 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1178 independently by Hanno Böck.
1182 *) PKCS7 crash with missing EnvelopedContent
1184 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1185 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1186 with missing content and trigger a NULL pointer dereference on parsing.
1188 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1189 structures from untrusted sources are affected. OpenSSL clients and
1190 servers are not affected.
1192 This issue was reported to OpenSSL by Michal Zalewski (Google).
1196 *) CMS verify infinite loop with unknown hash function
1198 When verifying a signedData message the CMS code can enter an infinite loop
1199 if presented with an unknown hash function OID. This can be used to perform
1200 denial of service against any system which verifies signedData messages using
1202 This issue was reported to OpenSSL by Johannes Bauer.
1206 *) Race condition handling NewSessionTicket
1208 If a NewSessionTicket is received by a multi-threaded client when attempting to
1209 reuse a previous ticket then a race condition can occur potentially leading to
1210 a double free of the ticket data.
1214 *) Only support 256-bit or stronger elliptic curves with the
1215 'ecdh_auto' setting (server) or by default (client). Of supported
1216 curves, prefer P-256 (both).
1219 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1221 *) ClientHello sigalgs DoS fix
1223 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1224 invalid signature algorithms extension a NULL pointer dereference will
1225 occur. This can be exploited in a DoS attack against the server.
1227 This issue was was reported to OpenSSL by David Ramos of Stanford
1230 [Stephen Henson and Matt Caswell]
1232 *) Multiblock corrupted pointer fix
1234 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1235 feature only applies on 64 bit x86 architecture platforms that support AES
1236 NI instructions. A defect in the implementation of "multiblock" can cause
1237 OpenSSL's internal write buffer to become incorrectly set to NULL when
1238 using non-blocking IO. Typically, when the user application is using a
1239 socket BIO for writing, this will only result in a failed connection.
1240 However if some other BIO is used then it is likely that a segmentation
1241 fault will be triggered, thus enabling a potential DoS attack.
1243 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1247 *) Segmentation fault in DTLSv1_listen fix
1249 The DTLSv1_listen function is intended to be stateless and processes the
1250 initial ClientHello from many peers. It is common for user code to loop
1251 over the call to DTLSv1_listen until a valid ClientHello is received with
1252 an associated cookie. A defect in the implementation of DTLSv1_listen means
1253 that state is preserved in the SSL object from one invocation to the next
1254 that can lead to a segmentation fault. Errors processing the initial
1255 ClientHello can trigger this scenario. An example of such an error could be
1256 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1259 This issue was reported to OpenSSL by Per Allansson.
1263 *) Segmentation fault in ASN1_TYPE_cmp fix
1265 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1266 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1267 certificate signature algorithm consistency this can be used to crash any
1268 certificate verification operation and exploited in a DoS attack. Any
1269 application which performs certificate verification is vulnerable including
1270 OpenSSL clients and servers which enable client authentication.
1274 *) Segmentation fault for invalid PSS parameters fix
1276 The signature verification routines will crash with a NULL pointer
1277 dereference if presented with an ASN.1 signature using the RSA PSS
1278 algorithm and invalid parameters. Since these routines are used to verify
1279 certificate signature algorithms this can be used to crash any
1280 certificate verification operation and exploited in a DoS attack. Any
1281 application which performs certificate verification is vulnerable including
1282 OpenSSL clients and servers which enable client authentication.
1284 This issue was was reported to OpenSSL by Brian Carpenter.
1288 *) ASN.1 structure reuse memory corruption fix
1290 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1291 memory corruption via an invalid write. Such reuse is and has been
1292 strongly discouraged and is believed to be rare.
1294 Applications that parse structures containing CHOICE or ANY DEFINED BY
1295 components may be affected. Certificate parsing (d2i_X509 and related
1296 functions) are however not affected. OpenSSL clients and servers are
1301 *) PKCS7 NULL pointer dereferences fix
1303 The PKCS#7 parsing code does not handle missing outer ContentInfo
1304 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1305 missing content and trigger a NULL pointer dereference on parsing.
1307 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1308 otherwise parse PKCS#7 structures from untrusted sources are
1309 affected. OpenSSL clients and servers are not affected.
1311 This issue was reported to OpenSSL by Michal Zalewski (Google).
1315 *) DoS via reachable assert in SSLv2 servers fix
1317 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1318 servers that both support SSLv2 and enable export cipher suites by sending
1319 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1321 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1322 (OpenSSL development team).
1326 *) Empty CKE with client auth and DHE fix
1328 If client auth is used then a server can seg fault in the event of a DHE
1329 ciphersuite being selected and a zero length ClientKeyExchange message
1330 being sent by the client. This could be exploited in a DoS attack.
1334 *) Handshake with unseeded PRNG fix
1336 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1337 with an unseeded PRNG. The conditions are:
1338 - The client is on a platform where the PRNG has not been seeded
1339 automatically, and the user has not seeded manually
1340 - A protocol specific client method version has been used (i.e. not
1341 SSL_client_methodv23)
1342 - A ciphersuite is used that does not require additional random data from
1343 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1345 If the handshake succeeds then the client random that has been used will
1346 have been generated from a PRNG with insufficient entropy and therefore the
1347 output may be predictable.
1349 For example using the following command with an unseeded openssl will
1350 succeed on an unpatched platform:
1352 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1356 *) Use After Free following d2i_ECPrivatekey error fix
1358 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1359 could cause a use after free condition. This, in turn, could cause a double
1360 free in several private key parsing functions (such as d2i_PrivateKey
1361 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1362 for applications that receive EC private keys from untrusted
1363 sources. This scenario is considered rare.
1365 This issue was discovered by the BoringSSL project and fixed in their
1370 *) X509_to_X509_REQ NULL pointer deref fix
1372 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1373 the certificate key is invalid. This function is rarely used in practice.
1375 This issue was discovered by Brian Carpenter.
1379 *) Removed the export ciphers from the DEFAULT ciphers
1382 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1384 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1385 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1386 So far those who have to target multiple plaforms would compromise
1387 and argue that binary targeting say ARMv5 would still execute on
1388 ARMv8. "Universal" build resolves this compromise by providing
1389 near-optimal performance even on newer platforms.
1392 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1393 (other platforms pending).
1394 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1396 *) Add support for the SignedCertificateTimestampList certificate and
1397 OCSP response extensions from RFC6962.
1400 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1401 for corner cases. (Certain input points at infinity could lead to
1402 bogus results, with non-infinity inputs mapped to infinity too.)
1405 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1406 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1407 common cases are optimized and there still is room for further
1408 improvements. Vector Permutation AES for Altivec is also added.
1411 *) Add support for little-endian ppc64 Linux target.
1412 [Marcelo Cerri (IBM)]
1414 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1415 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1416 are optimized and there still is room for further improvements.
1417 Both 32- and 64-bit modes are supported.
1418 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1420 *) Improved ARMv7 NEON support.
1423 *) Support for SPARC Architecture 2011 crypto extensions, first
1424 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1425 SHA256/512, MD5, GHASH and modular exponentiation.
1426 [Andy Polyakov, David Miller]
1428 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1430 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1432 *) Support for new and upcoming Intel processors, including AVX2,
1433 BMI and SHA ISA extensions. This includes additional "stitched"
1434 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1437 This work was sponsored by Intel Corp.
1440 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1441 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1442 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1445 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1446 this fixes a limiation in previous versions of OpenSSL.
1449 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1450 MGF1 digest and OAEP label.
1453 *) Add EVP support for key wrapping algorithms, to avoid problems with
1454 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1455 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1456 algorithms and include tests cases.
1459 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1461 [Douglas E. Engert, Steve Henson]
1463 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1464 difference in days and seconds between two tm or ASN1_TIME structures.
1467 *) Add -rev test option to s_server to just reverse order of characters
1468 received by client and send back to server. Also prints an abbreviated
1469 summary of the connection parameters.
1472 *) New option -brief for s_client and s_server to print out a brief summary
1473 of connection parameters.
1476 *) Add callbacks for arbitrary TLS extensions.
1477 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1479 *) New option -crl_download in several openssl utilities to download CRLs
1480 from CRLDP extension in certificates.
1483 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1486 *) New function X509_CRL_diff to generate a delta CRL from the difference
1487 of two full CRLs. Add support to "crl" utility.
1490 *) New functions to set lookup_crls function and to retrieve
1491 X509_STORE from X509_STORE_CTX.
1494 *) Print out deprecated issuer and subject unique ID fields in
1498 *) Extend OCSP I/O functions so they can be used for simple general purpose
1499 HTTP as well as OCSP. New wrapper function which can be used to download
1500 CRLs using the OCSP API.
1503 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1506 *) SSL_CONF* functions. These provide a common framework for application
1507 configuration using configuration files or command lines.
1510 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1511 message callback and prints the results. Needs compile time option
1512 "enable-ssl-trace". New options to s_client and s_server to enable
1516 *) New ctrl and macro to retrieve supported points extensions.
1517 Print out extension in s_server and s_client.
1520 *) New functions to retrieve certificate signature and signature
1524 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1528 *) New Suite B modes for TLS code. These use and enforce the requirements
1529 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1530 only use Suite B curves. The Suite B modes can be set by using the
1531 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1534 *) New chain verification flags for Suite B levels of security. Check
1535 algorithms are acceptable when flags are set in X509_verify_cert.
1538 *) Make tls1_check_chain return a set of flags indicating checks passed
1539 by a certificate chain. Add additional tests to handle client
1540 certificates: checks for matching certificate type and issuer name
1544 *) If an attempt is made to use a signature algorithm not in the peer
1545 preference list abort the handshake. If client has no suitable
1546 signature algorithms in response to a certificate request do not
1547 use the certificate.
1550 *) If server EC tmp key is not in client preference list abort handshake.
1553 *) Add support for certificate stores in CERT structure. This makes it
1554 possible to have different stores per SSL structure or one store in
1555 the parent SSL_CTX. Include distint stores for certificate chain
1556 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1557 to build and store a certificate chain in CERT structure: returing
1558 an error if the chain cannot be built: this will allow applications
1559 to test if a chain is correctly configured.
1561 Note: if the CERT based stores are not set then the parent SSL_CTX
1562 store is used to retain compatibility with existing behaviour.
1566 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1567 mask based on the current session, check mask when sending client
1568 hello and checking the requested ciphersuite.
1571 *) New ctrls to retrieve and set certificate types in a certificate
1572 request message. Print out received values in s_client. If certificate
1573 types is not set with custom values set sensible values based on
1574 supported signature algorithms.
1577 *) Support for distinct client and server supported signature algorithms.
1580 *) Add certificate callback. If set this is called whenever a certificate
1581 is required by client or server. An application can decide which
1582 certificate chain to present based on arbitrary criteria: for example
1583 supported signature algorithms. Add very simple example to s_server.
1584 This fixes many of the problems and restrictions of the existing client
1585 certificate callback: for example you can now clear an existing
1586 certificate and specify the whole chain.
1589 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1590 the certificate can be used for (if anything). Set valid_flags field
1591 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1592 to have similar checks in it.
1594 Add new "cert_flags" field to CERT structure and include a "strict mode".
1595 This enforces some TLS certificate requirements (such as only permitting
1596 certificate signature algorithms contained in the supported algorithms
1597 extension) which some implementations ignore: this option should be used
1598 with caution as it could cause interoperability issues.
1601 *) Update and tidy signature algorithm extension processing. Work out
1602 shared signature algorithms based on preferences and peer algorithms
1603 and print them out in s_client and s_server. Abort handshake if no
1604 shared signature algorithms.
1607 *) Add new functions to allow customised supported signature algorithms
1608 for SSL and SSL_CTX structures. Add options to s_client and s_server
1612 *) New function SSL_certs_clear() to delete all references to certificates
1613 from an SSL structure. Before this once a certificate had been added
1614 it couldn't be removed.
1617 *) Integrate hostname, email address and IP address checking with certificate
1618 verification. New verify options supporting checking in opensl utility.
1621 *) Fixes and wildcard matching support to hostname and email checking
1622 functions. Add manual page.
1623 [Florian Weimer (Red Hat Product Security Team)]
1625 *) New functions to check a hostname email or IP address against a
1626 certificate. Add options x509 utility to print results of checks against
1630 *) Fix OCSP checking.
1631 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1633 *) Initial experimental support for explicitly trusted non-root CAs.
1634 OpenSSL still tries to build a complete chain to a root but if an
1635 intermediate CA has a trust setting included that is used. The first
1636 setting is used: whether to trust (e.g., -addtrust option to the x509
1640 *) Add -trusted_first option which attempts to find certificates in the
1641 trusted store even if an untrusted chain is also supplied.
1644 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1645 platform support for Linux and Android.
1648 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1651 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1652 When in FIPS mode the approved implementations are used as normal,
1653 when not in FIPS mode the internal unapproved versions are used instead.
1654 This means that the FIPS capable OpenSSL isn't forced to use the
1655 (often lower performance) FIPS implementations outside FIPS mode.
1658 *) Transparently support X9.42 DH parameters when calling
1659 PEM_read_bio_DHparameters. This means existing applications can handle
1660 the new parameter format automatically.
1663 *) Initial experimental support for X9.42 DH parameter format: mainly
1664 to support use of 'q' parameter for RFC5114 parameters.
1667 *) Add DH parameters from RFC5114 including test data to dhtest.
1670 *) Support for automatic EC temporary key parameter selection. If enabled
1671 the most preferred EC parameters are automatically used instead of
1672 hardcoded fixed parameters. Now a server just has to call:
1673 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1674 support ECDH and use the most appropriate parameters.
1677 *) Enhance and tidy EC curve and point format TLS extension code. Use
1678 static structures instead of allocation if default values are used.
1679 New ctrls to set curves we wish to support and to retrieve shared curves.
1680 Print out shared curves in s_server. New options to s_server and s_client
1681 to set list of supported curves.
1684 *) New ctrls to retrieve supported signature algorithms and
1685 supported curve values as an array of NIDs. Extend openssl utility
1686 to print out received values.
1689 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1690 between NIDs and the more common NIST names such as "P-256". Enhance
1691 ecparam utility and ECC method to recognise the NIST names for curves.
1694 *) Enhance SSL/TLS certificate chain handling to support different
1695 chains for each certificate instead of one chain in the parent SSL_CTX.
1698 *) Support for fixed DH ciphersuite client authentication: where both
1699 server and client use DH certificates with common parameters.
1702 *) Support for fixed DH ciphersuites: those requiring DH server
1706 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1708 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1709 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1710 X509_CINF_get_signature were reverted post internal team review.
1712 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1714 *) Build fixes for the Windows and OpenVMS platforms
1715 [Matt Caswell and Richard Levitte]
1717 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1719 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1720 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1721 dereference. This could lead to a Denial Of Service attack. Thanks to
1722 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1726 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1727 dtls1_buffer_record function under certain conditions. In particular this
1728 could occur if an attacker sent repeated DTLS records with the same
1729 sequence number but for the next epoch. The memory leak could be exploited
1730 by an attacker in a Denial of Service attack through memory exhaustion.
1731 Thanks to Chris Mueller for reporting this issue.
1735 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1736 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1737 method would be set to NULL which could later result in a NULL pointer
1738 dereference. Thanks to Frank Schmirler for reporting this issue.
1742 *) Abort handshake if server key exchange message is omitted for ephemeral
1745 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1746 reporting this issue.
1750 *) Remove non-export ephemeral RSA code on client and server. This code
1751 violated the TLS standard by allowing the use of temporary RSA keys in
1752 non-export ciphersuites and could be used by a server to effectively
1753 downgrade the RSA key length used to a value smaller than the server
1754 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1755 INRIA or reporting this issue.
1759 *) Fixed issue where DH client certificates are accepted without verification.
1760 An OpenSSL server will accept a DH certificate for client authentication
1761 without the certificate verify message. This effectively allows a client to
1762 authenticate without the use of a private key. This only affects servers
1763 which trust a client certificate authority which issues certificates
1764 containing DH keys: these are extremely rare and hardly ever encountered.
1765 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1770 *) Ensure that the session ID context of an SSL is updated when its
1771 SSL_CTX is updated via SSL_set_SSL_CTX.
1773 The session ID context is typically set from the parent SSL_CTX,
1774 and can vary with the CTX.
1777 *) Fix various certificate fingerprint issues.
1779 By using non-DER or invalid encodings outside the signed portion of a
1780 certificate the fingerprint can be changed without breaking the signature.
1781 Although no details of the signed portion of the certificate can be changed
1782 this can cause problems with some applications: e.g. those using the
1783 certificate fingerprint for blacklists.
1785 1. Reject signatures with non zero unused bits.
1787 If the BIT STRING containing the signature has non zero unused bits reject
1788 the signature. All current signature algorithms require zero unused bits.
1790 2. Check certificate algorithm consistency.
1792 Check the AlgorithmIdentifier inside TBS matches the one in the
1793 certificate signature. NB: this will result in signature failure
1794 errors for some broken certificates.
1796 Thanks to Konrad Kraszewski from Google for reporting this issue.
1798 3. Check DSA/ECDSA signatures use DER.
1800 Reencode DSA/ECDSA signatures and compare with the original received
1801 signature. Return an error if there is a mismatch.
1803 This will reject various cases including garbage after signature
1804 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1805 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1806 (negative or with leading zeroes).
1808 Further analysis was conducted and fixes were developed by Stephen Henson
1809 of the OpenSSL core team.
1814 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1815 results on some platforms, including x86_64. This bug occurs at random
1816 with a very low probability, and is not known to be exploitable in any
1817 way, though its exact impact is difficult to determine. Thanks to Pieter
1818 Wuille (Blockstream) who reported this issue and also suggested an initial
1819 fix. Further analysis was conducted by the OpenSSL development team and
1820 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1821 the OpenSSL core team.
1825 *) Do not resume sessions on the server if the negotiated protocol
1826 version does not match the session's version. Resuming with a different
1827 version, while not strictly forbidden by the RFC, is of questionable
1828 sanity and breaks all known clients.
1829 [David Benjamin, Emilia Käsper]
1831 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1832 early CCS messages during renegotiation. (Note that because
1833 renegotiation is encrypted, this early CCS was not exploitable.)
1836 *) Tighten client-side session ticket handling during renegotiation:
1837 ensure that the client only accepts a session ticket if the server sends
1838 the extension anew in the ServerHello. Previously, a TLS client would
1839 reuse the old extension state and thus accept a session ticket if one was
1840 announced in the initial ServerHello.
1842 Similarly, ensure that the client requires a session ticket if one
1843 was advertised in the ServerHello. Previously, a TLS client would
1844 ignore a missing NewSessionTicket message.
1847 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1849 *) SRTP Memory Leak.
1851 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1852 sends a carefully crafted handshake message, to cause OpenSSL to fail
1853 to free up to 64k of memory causing a memory leak. This could be
1854 exploited in a Denial Of Service attack. This issue affects OpenSSL
1855 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1856 whether SRTP is used or configured. Implementations of OpenSSL that
1857 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1859 The fix was developed by the OpenSSL team.
1863 *) Session Ticket Memory Leak.
1865 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1866 integrity of that ticket is first verified. In the event of a session
1867 ticket integrity check failing, OpenSSL will fail to free memory
1868 causing a memory leak. By sending a large number of invalid session
1869 tickets an attacker could exploit this issue in a Denial Of Service
1874 *) Build option no-ssl3 is incomplete.
1876 When OpenSSL is configured with "no-ssl3" as a build option, servers
1877 could accept and complete a SSL 3.0 handshake, and clients could be
1878 configured to send them.
1880 [Akamai and the OpenSSL team]
1882 *) Add support for TLS_FALLBACK_SCSV.
1883 Client applications doing fallback retries should call
1884 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1886 [Adam Langley, Bodo Moeller]
1888 *) Add additional DigestInfo checks.
1890 Reencode DigestInto in DER and check against the original when
1891 verifying RSA signature: this will reject any improperly encoded
1892 DigestInfo structures.
1894 Note: this is a precautionary measure and no attacks are currently known.
1898 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1900 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1901 SRP code can be overrun an internal buffer. Add sanity check that
1902 g, A, B < N to SRP code.
1904 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1905 Group for discovering this issue.
1909 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1910 TLS 1.0 instead of higher protocol versions when the ClientHello message
1911 is badly fragmented. This allows a man-in-the-middle attacker to force a
1912 downgrade to TLS 1.0 even if both the server and the client support a
1913 higher protocol version, by modifying the client's TLS records.
1915 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1916 researching this issue.
1920 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1921 to a denial of service attack. A malicious server can crash the client
1922 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1923 ciphersuite and sending carefully crafted handshake messages.
1925 Thanks to Felix Gröbert (Google) for discovering and researching this
1930 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1931 to leak memory. This can be exploited through a Denial of Service attack.
1932 Thanks to Adam Langley for discovering and researching this issue.
1936 *) An attacker can force openssl to consume large amounts of memory whilst
1937 processing DTLS handshake messages. This can be exploited through a
1938 Denial of Service attack.
1939 Thanks to Adam Langley for discovering and researching this issue.
1943 *) An attacker can force an error condition which causes openssl to crash
1944 whilst processing DTLS packets due to memory being freed twice. This
1945 can be exploited through a Denial of Service attack.
1946 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1951 *) If a multithreaded client connects to a malicious server using a resumed
1952 session and the server sends an ec point format extension it could write
1953 up to 255 bytes to freed memory.
1955 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1960 *) A malicious server can crash an OpenSSL client with a null pointer
1961 dereference (read) by specifying an SRP ciphersuite even though it was not
1962 properly negotiated with the client. This can be exploited through a
1963 Denial of Service attack.
1965 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1966 discovering and researching this issue.
1970 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1971 X509_name_oneline, X509_name_print_ex et al. to leak some information
1972 from the stack. Applications may be affected if they echo pretty printing
1973 output to the attacker.
1975 Thanks to Ivan Fratric (Google) for discovering this issue.
1977 [Emilia Käsper, and Steve Henson]
1979 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1980 for corner cases. (Certain input points at infinity could lead to
1981 bogus results, with non-infinity inputs mapped to infinity too.)
1984 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1986 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1987 handshake can force the use of weak keying material in OpenSSL
1988 SSL/TLS clients and servers.
1990 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1991 researching this issue. (CVE-2014-0224)
1992 [KIKUCHI Masashi, Steve Henson]
1994 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1995 OpenSSL DTLS client the code can be made to recurse eventually crashing
1998 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2000 [Imre Rad, Steve Henson]
2002 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2003 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2004 client or server. This is potentially exploitable to run arbitrary
2005 code on a vulnerable client or server.
2007 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2008 [Jüri Aedla, Steve Henson]
2010 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2011 are subject to a denial of service attack.
2013 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2014 this issue. (CVE-2014-3470)
2015 [Felix Gröbert, Ivan Fratric, Steve Henson]
2017 *) Harmonize version and its documentation. -f flag is used to display
2019 [mancha <mancha1@zoho.com>]
2021 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2022 in i2d_ECPrivateKey.
2023 [mancha <mancha1@zoho.com>]
2025 *) Fix some double frees. These are not thought to be exploitable.
2026 [mancha <mancha1@zoho.com>]
2028 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2030 *) A missing bounds check in the handling of the TLS heartbeat extension
2031 can be used to reveal up to 64k of memory to a connected client or
2034 Thanks for Neel Mehta of Google Security for discovering this bug and to
2035 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2036 preparing the fix (CVE-2014-0160)
2037 [Adam Langley, Bodo Moeller]
2039 *) Fix for the attack described in the paper "Recovering OpenSSL
2040 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2041 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2042 http://eprint.iacr.org/2014/140
2044 Thanks to Yuval Yarom and Naomi Benger for discovering this
2045 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2046 [Yuval Yarom and Naomi Benger]
2048 *) TLS pad extension: draft-agl-tls-padding-03
2050 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2051 TLS client Hello record length value would otherwise be > 255 and
2052 less that 512 pad with a dummy extension containing zeroes so it
2053 is at least 512 bytes long.
2055 [Adam Langley, Steve Henson]
2057 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2059 *) Fix for TLS record tampering bug. A carefully crafted invalid
2060 handshake could crash OpenSSL with a NULL pointer exception.
2061 Thanks to Anton Johansson for reporting this issues.
2064 *) Keep original DTLS digest and encryption contexts in retransmission
2065 structures so we can use the previous session parameters if they need
2066 to be resent. (CVE-2013-6450)
2069 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2070 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2071 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2072 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2073 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2074 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2075 [Rob Stradling, Adam Langley]
2077 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2079 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2080 supporting platforms or when small records were transferred.
2081 [Andy Polyakov, Steve Henson]
2083 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2085 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2087 This addresses the flaw in CBC record processing discovered by
2088 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2089 at: http://www.isg.rhul.ac.uk/tls/
2091 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2092 Security Group at Royal Holloway, University of London
2093 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2094 Emilia Käsper for the initial patch.
2096 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2098 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2099 ciphersuites which can be exploited in a denial of service attack.
2100 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2101 and detecting this bug and to Wolfgang Ettlinger
2102 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2106 *) Return an error when checking OCSP signatures when key is NULL.
2107 This fixes a DoS attack. (CVE-2013-0166)
2110 *) Make openssl verify return errors.
2111 [Chris Palmer <palmer@google.com> and Ben Laurie]
2113 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2114 the right response is stapled. Also change SSL_get_certificate()
2115 so it returns the certificate actually sent.
2116 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2117 [Rob Stradling <rob.stradling@comodo.com>]
2119 *) Fix possible deadlock when decoding public keys.
2122 *) Don't use TLS 1.0 record version number in initial client hello
2126 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2128 *) Sanity check record length before skipping explicit IV in TLS
2129 1.2, 1.1 and DTLS to fix DoS attack.
2131 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2132 fuzzing as a service testing platform.
2136 *) Initialise tkeylen properly when encrypting CMS messages.
2137 Thanks to Solar Designer of Openwall for reporting this issue.
2140 *) In FIPS mode don't try to use composite ciphers as they are not
2144 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2146 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2147 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2148 mean any application compiled against OpenSSL 1.0.0 headers setting
2149 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2150 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2151 0x10000000L Any application which was previously compiled against
2152 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2153 will need to be recompiled as a result. Letting be results in
2154 inability to disable specifically TLS 1.1 and in client context,
2155 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2158 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2159 disable just protocol X, but all protocols above X *if* there are
2160 protocols *below* X still enabled. In more practical terms it means
2161 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2162 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2163 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2167 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2169 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2170 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2171 in CRYPTO_realloc_clean.
2173 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2174 issue and to Adam Langley <agl@chromium.org> for fixing it.
2176 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2178 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2181 *) Workarounds for some broken servers that "hang" if a client hello
2182 record length exceeds 255 bytes.
2184 1. Do not use record version number > TLS 1.0 in initial client
2185 hello: some (but not all) hanging servers will now work.
2186 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2187 the number of ciphers sent in the client hello. This should be
2188 set to an even number, such as 50, for example by passing:
2189 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2190 Most broken servers should now work.
2191 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2192 TLS 1.2 client support entirely.
2195 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2198 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2200 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2201 STRING form instead of a DigestInfo.
2204 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2205 and the RSA_sign/RSA_verify functions. This was made more apparent when
2206 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2207 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2208 the correct format in RSA_verify so both forms transparently work.
2211 *) Some servers which support TLS 1.0 can choke if we initially indicate
2212 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2213 encrypted premaster secret. As a workaround use the maximum permitted
2214 client version in client hello, this should keep such servers happy
2215 and still work with previous versions of OpenSSL.
2218 *) Add support for TLS/DTLS heartbeats.
2219 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2221 *) Add support for SCTP.
2222 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2224 *) Improved PRNG seeding for VOS.
2225 [Paul Green <Paul.Green@stratus.com>]
2227 *) Extensive assembler packs updates, most notably:
2229 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2230 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2231 - x86_64: bit-sliced AES implementation;
2232 - ARM: NEON support, contemporary platforms optimizations;
2233 - s390x: z196 support;
2234 - *: GHASH and GF(2^m) multiplication implementations;
2238 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2239 (removal of unnecessary code)
2240 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2242 *) Add TLS key material exporter from RFC 5705.
2245 *) Add DTLS-SRTP negotiation from RFC 5764.
2248 *) Add Next Protocol Negotiation,
2249 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2250 disabled with a no-npn flag to config or Configure. Code donated
2252 [Adam Langley <agl@google.com> and Ben Laurie]
2254 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2255 NIST-P256, NIST-P521, with constant-time single point multiplication on
2256 typical inputs. Compiler support for the nonstandard type __uint128_t is
2257 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2258 Code made available under Apache License version 2.0.
2260 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2261 line to include this in your build of OpenSSL, and run "make depend" (or
2262 "make update"). This enables the following EC_METHODs:
2264 EC_GFp_nistp224_method()
2265 EC_GFp_nistp256_method()
2266 EC_GFp_nistp521_method()
2268 EC_GROUP_new_by_curve_name() will automatically use these (while
2269 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2271 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2273 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2274 all platforms. Move ssize_t definition from e_os.h to the public
2275 header file e_os2.h as it now appears in public header file cms.h
2278 *) New -sigopt option to the ca, req and x509 utilities. Additional
2279 signature parameters can be passed using this option and in
2283 *) Add RSA PSS signing function. This will generate and set the
2284 appropriate AlgorithmIdentifiers for PSS based on those in the
2285 corresponding EVP_MD_CTX structure. No application support yet.
2288 *) Support for companion algorithm specific ASN1 signing routines.
2289 New function ASN1_item_sign_ctx() signs a pre-initialised
2290 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2291 the appropriate parameters.
2294 *) Add new algorithm specific ASN1 verification initialisation function
2295 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2296 handling will be the same no matter what EVP_PKEY_METHOD is used.
2297 Add a PSS handler to support verification of PSS signatures: checked
2298 against a number of sample certificates.
2301 *) Add signature printing for PSS. Add PSS OIDs.
2302 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2304 *) Add algorithm specific signature printing. An individual ASN1 method
2305 can now print out signatures instead of the standard hex dump.
2307 More complex signatures (e.g. PSS) can print out more meaningful
2308 information. Include DSA version that prints out the signature
2312 *) Password based recipient info support for CMS library: implementing
2316 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2317 neatly separates the code into cipher and PBE sections and is required
2318 for some algorithms that split PBES2 into separate pieces (such as
2319 password based CMS).
2322 *) Session-handling fixes:
2323 - Fix handling of connections that are resuming with a session ID,
2324 but also support Session Tickets.
2325 - Fix a bug that suppressed issuing of a new ticket if the client
2326 presented a ticket with an expired session.
2327 - Try to set the ticket lifetime hint to something reasonable.
2328 - Make tickets shorter by excluding irrelevant information.
2329 - On the client side, don't ignore renewed tickets.
2330 [Adam Langley, Bodo Moeller (Google)]
2332 *) Fix PSK session representation.
2335 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2337 This work was sponsored by Intel.
2340 *) Add GCM support to TLS library. Some custom code is needed to split
2341 the IV between the fixed (from PRF) and explicit (from TLS record)
2342 portions. This adds all GCM ciphersuites supported by RFC5288 and
2343 RFC5289. Generalise some AES* cipherstrings to include GCM and
2344 add a special AESGCM string for GCM only.
2347 *) Expand range of ctrls for AES GCM. Permit setting invocation
2348 field on decrypt and retrieval of invocation field only on encrypt.
2351 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2352 As required by RFC5289 these ciphersuites cannot be used if for
2353 versions of TLS earlier than 1.2.
2356 *) For FIPS capable OpenSSL interpret a NULL default public key method
2357 as unset and return the appropriate default but do *not* set the default.
2358 This means we can return the appropriate method in applications that
2359 switch between FIPS and non-FIPS modes.
2362 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2363 ENGINE is used then we cannot handle that in the FIPS module so we
2364 keep original code iff non-FIPS operations are allowed.
2367 *) Add -attime option to openssl utilities.
2368 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2370 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2373 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2374 FIPS EC methods unconditionally for now.
2377 *) New build option no-ec2m to disable characteristic 2 code.
2380 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2381 all cases can be covered as some introduce binary incompatibilities.
2384 *) Redirect RSA operations to FIPS module including keygen,
2385 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2388 *) Add similar low level API blocking to ciphers.
2391 *) Low level digest APIs are not approved in FIPS mode: any attempt
2392 to use these will cause a fatal error. Applications that *really* want
2393 to use them can use the private_* version instead.
2396 *) Redirect cipher operations to FIPS module for FIPS builds.
2399 *) Redirect digest operations to FIPS module for FIPS builds.
2402 *) Update build system to add "fips" flag which will link in fipscanister.o
2403 for static and shared library builds embedding a signature if needed.
2406 *) Output TLS supported curves in preference order instead of numerical
2407 order. This is currently hardcoded for the highest order curves first.
2408 This should be configurable so applications can judge speed vs strength.
2411 *) Add TLS v1.2 server support for client authentication.
2414 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2418 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2419 FIPS modules versions.
2422 *) Add TLS v1.2 client side support for client authentication. Keep cache
2423 of handshake records longer as we don't know the hash algorithm to use
2424 until after the certificate request message is received.
2427 *) Initial TLS v1.2 client support. Add a default signature algorithms
2428 extension including all the algorithms we support. Parse new signature
2429 format in client key exchange. Relax some ECC signing restrictions for
2430 TLS v1.2 as indicated in RFC5246.
2433 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2434 to new signature format when needed using client digest preference.
2435 All server ciphersuites should now work correctly in TLS v1.2. No client
2436 support yet and no support for client certificates.
2439 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2440 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2441 ciphersuites. At present only RSA key exchange ciphersuites work with
2442 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2443 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2444 and version checking.
2447 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2448 with this defined it will not be affected by any changes to ssl internal
2449 structures. Add several utility functions to allow openssl application
2450 to work with OPENSSL_NO_SSL_INTERN defined.
2454 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2456 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2459 *) Permit abbreviated handshakes when renegotiating using the function
2460 SSL_renegotiate_abbreviated().
2461 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2463 *) Add call to ENGINE_register_all_complete() to
2464 ENGINE_load_builtin_engines(), so some implementations get used
2465 automatically instead of needing explicit application support.
2468 *) Add support for TLS key exporter as described in RFC5705.
2469 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2471 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2472 a few changes are required:
2474 Add SSL_OP_NO_TLSv1_1 flag.
2475 Add TLSv1_1 methods.
2476 Update version checking logic to handle version 1.1.
2477 Add explicit IV handling (ported from DTLS code).
2478 Add command line options to s_client/s_server.
2481 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2483 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2484 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2485 content decryption and always return the same error. Note: this attack
2486 needs on average 2^20 messages so it only affects automated senders. The
2487 old behaviour can be reenabled in the CMS code by setting the
2488 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2489 an MMA defence is not necessary.
2490 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2491 this issue. (CVE-2012-0884)
2494 *) Fix CVE-2011-4619: make sure we really are receiving a
2495 client hello before rejecting multiple SGC restarts. Thanks to
2496 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2499 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2501 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2502 Thanks to Antonio Martin, Enterprise Secure Access Research and
2503 Development, Cisco Systems, Inc. for discovering this bug and
2504 preparing a fix. (CVE-2012-0050)
2507 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2509 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2510 of the Vaudenay padding oracle attack on CBC mode encryption
2511 which enables an efficient plaintext recovery attack against
2512 the OpenSSL implementation of DTLS. Their attack exploits timing
2513 differences arising during decryption processing. A research
2514 paper describing this attack can be found at:
2515 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2516 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2517 Security Group at Royal Holloway, University of London
2518 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2519 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2520 for preparing the fix. (CVE-2011-4108)
2521 [Robin Seggelmann, Michael Tuexen]
2523 *) Clear bytes used for block padding of SSL 3.0 records.
2525 [Adam Langley (Google)]
2527 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2528 Kadianakis <desnacked@gmail.com> for discovering this issue and
2529 Adam Langley for preparing the fix. (CVE-2011-4619)
2530 [Adam Langley (Google)]
2532 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2533 [Andrey Kulikov <amdeich@gmail.com>]
2535 *) Prevent malformed RFC3779 data triggering an assertion failure.
2536 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2537 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2538 [Rob Austein <sra@hactrn.net>]
2540 *) Improved PRNG seeding for VOS.
2541 [Paul Green <Paul.Green@stratus.com>]
2543 *) Fix ssl_ciph.c set-up race.
2544 [Adam Langley (Google)]
2546 *) Fix spurious failures in ecdsatest.c.
2547 [Emilia Käsper (Google)]
2549 *) Fix the BIO_f_buffer() implementation (which was mixing different
2550 interpretations of the '..._len' fields).
2551 [Adam Langley (Google)]
2553 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2554 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2555 threads won't reuse the same blinding coefficients.
2557 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2558 lock to call BN_BLINDING_invert_ex, and avoids one use of
2559 BN_BLINDING_update for each BN_BLINDING structure (previously,
2560 the last update always remained unused).
2561 [Emilia Käsper (Google)]
2563 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2564 [Bob Buckholz (Google)]
2566 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2568 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2569 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2570 [Kaspar Brand <ossl@velox.ch>]
2572 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2573 for multi-threaded use of ECDH. (CVE-2011-3210)
2574 [Adam Langley (Google)]
2576 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2579 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2580 signature public key algorithm by using OID xref utilities instead.
2581 Before this you could only use some ECC ciphersuites with SHA1 only.
2584 *) Add protection against ECDSA timing attacks as mentioned in the paper
2585 by Billy Bob Brumley and Nicola Tuveri, see:
2587 http://eprint.iacr.org/2011/232.pdf
2589 [Billy Bob Brumley and Nicola Tuveri]
2591 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2593 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2594 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2596 *) Fix bug in string printing code: if *any* escaping is enabled we must
2597 escape the escape character (backslash) or the resulting string is
2601 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2603 *) Disable code workaround for ancient and obsolete Netscape browsers
2604 and servers: an attacker can use it in a ciphersuite downgrade attack.
2605 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2608 *) Fixed J-PAKE implementation error, originally discovered by
2609 Sebastien Martini, further info and confirmation from Stefan
2610 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2613 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2615 *) Fix extension code to avoid race conditions which can result in a buffer
2616 overrun vulnerability: resumed sessions must not be modified as they can
2617 be shared by multiple threads. CVE-2010-3864
2620 *) Fix WIN32 build system to correctly link an ENGINE directory into
2624 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2626 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2628 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2630 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2632 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2633 context. The operation can be customised via the ctrl mechanism in
2634 case ENGINEs want to include additional functionality.
2637 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2640 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2641 output hashes compatible with older versions of OpenSSL.
2642 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2644 *) Fix compression algorithm handling: if resuming a session use the
2645 compression algorithm of the resumed session instead of determining
2646 it from client hello again. Don't allow server to change algorithm.
2649 *) Add load_crls() function to apps tidying load_certs() too. Add option
2650 to verify utility to allow additional CRLs to be included.
2653 *) Update OCSP request code to permit adding custom headers to the request:
2654 some responders need this.
2657 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2659 [Julia Lawall <julia@diku.dk>]
2661 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2662 needlessly dereferenced structures, used obsolete functions and
2663 didn't handle all updated verify codes correctly.
2666 *) Disable MD2 in the default configuration.
2669 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2670 indicate the initial BIO being pushed or popped. This makes it possible
2671 to determine whether the BIO is the one explicitly called or as a result
2672 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2673 it handles reference counts correctly and doesn't zero out the I/O bio
2674 when it is not being explicitly popped. WARNING: applications which
2675 included workarounds for the old buggy behaviour will need to be modified
2676 or they could free up already freed BIOs.
2679 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2680 renaming to all platforms (within the 0.9.8 branch, this was
2681 done conditionally on Netware platforms to avoid a name clash).
2682 [Guenter <lists@gknw.net>]
2684 *) Add ECDHE and PSK support to DTLS.
2685 [Michael Tuexen <tuexen@fh-muenster.de>]
2687 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2691 *) Add "missing" function EVP_MD_flags() (without this the only way to
2692 retrieve a digest flags is by accessing the structure directly. Update
2693 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2694 or cipher is registered as in the "from" argument. Print out all
2695 registered digests in the dgst usage message instead of manually
2696 attempting to work them out.
2699 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2700 this allows the use of compression and extensions. Change default cipher
2701 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2702 by default unless an application cipher string requests it.
2705 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2706 key ids to find matching certificates and keys but some PKCS#12 files
2707 don't follow the (somewhat unwritten) rules and this strategy fails.
2708 Now just gather all certificates together and the first private key
2709 then look for the first certificate that matches the key.
2712 *) Support use of registered digest and cipher names for dgst and cipher
2713 commands instead of having to add each one as a special case. So now
2720 openssl dgst -sha256 foo
2722 and this works for ENGINE based algorithms too.
2726 *) Update Gost ENGINE to support parameter files.
2727 [Victor B. Wagner <vitus@cryptocom.ru>]
2729 *) Support GeneralizedTime in ca utility.
2730 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2732 *) Enhance the hash format used for certificate directory links. The new
2733 form uses the canonical encoding (meaning equivalent names will work
2734 even if they aren't identical) and uses SHA1 instead of MD5. This form
2735 is incompatible with the older format and as a result c_rehash should
2736 be used to rebuild symbolic links.
2739 *) Make PKCS#8 the default write format for private keys, replacing the
2740 traditional format. This form is standardised, more secure and doesn't
2741 include an implicit MD5 dependency.
2744 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2745 committed to OpenSSL should pass this lot as a minimum.
2748 *) Add session ticket override functionality for use by EAP-FAST.
2749 [Jouni Malinen <j@w1.fi>]
2751 *) Modify HMAC functions to return a value. Since these can be implemented
2752 in an ENGINE errors can occur.
2755 *) Type-checked OBJ_bsearch_ex.
2758 *) Type-checked OBJ_bsearch. Also some constification necessitated
2759 by type-checking. Still to come: TXT_DB, bsearch(?),
2760 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2764 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2765 seconds to a tm structure directly, instead of going through OS
2766 specific date routines. This avoids any issues with OS routines such
2767 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2768 and X509_time_adj_ex() to cover the extended range. The existing
2769 X509_time_adj() is still usable and will no longer have any date issues.
2772 *) Delta CRL support. New use deltas option which will attempt to locate
2773 and search any appropriate delta CRLs available.
2775 This work was sponsored by Google.
2778 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2779 code and add additional score elements. Validate alternate CRL paths
2780 as part of the CRL checking and indicate a new error "CRL path validation
2781 error" in this case. Applications wanting additional details can use
2782 the verify callback and check the new "parent" field. If this is not
2783 NULL CRL path validation is taking place. Existing applications wont
2784 see this because it requires extended CRL support which is off by
2787 This work was sponsored by Google.
2790 *) Support for freshest CRL extension.
2792 This work was sponsored by Google.
2795 *) Initial indirect CRL support. Currently only supported in the CRLs
2796 passed directly and not via lookup. Process certificate issuer
2797 CRL entry extension and lookup CRL entries by bother issuer name
2798 and serial number. Check and process CRL issuer entry in IDP extension.
2800 This work was sponsored by Google.
2803 *) Add support for distinct certificate and CRL paths. The CRL issuer
2804 certificate is validated separately in this case. Only enabled if
2805 an extended CRL support flag is set: this flag will enable additional
2806 CRL functionality in future.
2808 This work was sponsored by Google.
2811 *) Add support for policy mappings extension.
2813 This work was sponsored by Google.
2816 *) Fixes to pathlength constraint, self issued certificate handling,
2817 policy processing to align with RFC3280 and PKITS tests.
2819 This work was sponsored by Google.
2822 *) Support for name constraints certificate extension. DN, email, DNS
2823 and URI types are currently supported.
2825 This work was sponsored by Google.
2828 *) To cater for systems that provide a pointer-based thread ID rather
2829 than numeric, deprecate the current numeric thread ID mechanism and
2830 replace it with a structure and associated callback type. This
2831 mechanism allows a numeric "hash" to be extracted from a thread ID in
2832 either case, and on platforms where pointers are larger than 'long',
2833 mixing is done to help ensure the numeric 'hash' is usable even if it
2834 can't be guaranteed unique. The default mechanism is to use "&errno"
2835 as a pointer-based thread ID to distinguish between threads.
2837 Applications that want to provide their own thread IDs should now use
2838 CRYPTO_THREADID_set_callback() to register a callback that will call
2839 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2841 Note that ERR_remove_state() is now deprecated, because it is tied
2842 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2843 to free the current thread's error state should be replaced by
2844 ERR_remove_thread_state(NULL).
2846 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2847 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2848 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2849 application was previously providing a numeric thread callback that
2850 was inappropriate for distinguishing threads, then uniqueness might
2851 have been obtained with &errno that happened immediately in the
2852 intermediate development versions of OpenSSL; this is no longer the
2853 case, the numeric thread callback will now override the automatic use
2855 [Geoff Thorpe, with help from Bodo Moeller]
2857 *) Initial support for different CRL issuing certificates. This covers a
2858 simple case where the self issued certificates in the chain exist and
2859 the real CRL issuer is higher in the existing chain.
2861 This work was sponsored by Google.
2864 *) Removed effectively defunct crypto/store from the build.
2867 *) Revamp of STACK to provide stronger type-checking. Still to come:
2868 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2869 ASN1_STRING, CONF_VALUE.
2872 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2873 RAM on SSL connections. This option can save about 34k per idle SSL.
2876 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2877 STACK, TXT_DB, bsearch, qsort.
2880 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2881 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2882 support for data, signedData, compressedData, digestedData and
2883 encryptedData, envelopedData types included. Scripts to check against
2884 RFC4134 examples draft and interop and consistency checks of many
2885 content types and variants.
2888 *) Add options to enc utility to support use of zlib compression BIO.
2891 *) Extend mk1mf to support importing of options and assembly language
2892 files from Configure script, currently only included in VC-WIN32.
2893 The assembly language rules can now optionally generate the source
2894 files from the associated perl scripts.
2897 *) Implement remaining functionality needed to support GOST ciphersuites.
2898 Interop testing has been performed using CryptoPro implementations.
2899 [Victor B. Wagner <vitus@cryptocom.ru>]
2901 *) s390x assembler pack.
2904 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2908 *) Implement Opaque PRF Input TLS extension as specified in
2909 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2910 official specification yet and no extension type assignment by
2911 IANA exists, this extension (for now) will have to be explicitly
2912 enabled when building OpenSSL by providing the extension number
2913 to use. For example, specify an option
2915 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2917 to the "config" or "Configure" script to enable the extension,
2918 assuming extension number 0x9527 (which is a completely arbitrary
2919 and unofficial assignment based on the MD5 hash of the Internet
2920 Draft). Note that by doing so, you potentially lose
2921 interoperability with other TLS implementations since these might
2922 be using the same extension number for other purposes.
2924 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2925 opaque PRF input value to use in the handshake. This will create
2926 an interal copy of the length-'len' string at 'src', and will
2927 return non-zero for success.
2929 To get more control and flexibility, provide a callback function
2932 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2933 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2937 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2940 Callback function 'cb' will be called in handshakes, and is
2941 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2942 Argument 'arg' is for application purposes (the value as given to
2943 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2944 be provided to the callback function). The callback function
2945 has to return non-zero to report success: usually 1 to use opaque
2946 PRF input just if possible, or 2 to enforce use of the opaque PRF
2947 input. In the latter case, the library will abort the handshake
2948 if opaque PRF input is not successfully negotiated.
2950 Arguments 'peerinput' and 'len' given to the callback function
2951 will always be NULL and 0 in the case of a client. A server will
2952 see the client's opaque PRF input through these variables if
2953 available (NULL and 0 otherwise). Note that if the server
2954 provides an opaque PRF input, the length must be the same as the
2955 length of the client's opaque PRF input.
2957 Note that the callback function will only be called when creating
2958 a new session (session resumption can resume whatever was
2959 previously negotiated), and will not be called in SSL 2.0
2960 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2961 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2962 for applications that need to enforce opaque PRF input.
2966 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2969 [Victor B. Wagner <vitus@cryptocom.ru>]
2971 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2972 RFC4507bis. The encrypted ticket format is an encrypted encoded
2973 SSL_SESSION structure, that way new session features are automatically
2976 If a client application caches session in an SSL_SESSION structure
2977 support is transparent because tickets are now stored in the encoded
2980 The SSL_CTX structure automatically generates keys for ticket
2981 protection in servers so again support should be possible
2982 with no application modification.
2984 If a client or server wishes to disable RFC4507 support then the option
2985 SSL_OP_NO_TICKET can be set.
2987 Add a TLS extension debugging callback to allow the contents of any client
2988 or server extensions to be examined.
2990 This work was sponsored by Google.
2993 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2994 OpenSSL should now compile cleanly on gcc 4.2
2995 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2997 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2998 support including streaming MAC support: this is required for GOST
2999 ciphersuite support.
3000 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3002 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3003 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3004 to output in BER and PEM format.
3007 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3008 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3009 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3010 ENGINE support for HMAC keys which are unextractable. New -mac and
3011 -macopt options to dgst utility.
3014 *) New option -sigopt to dgst utility. Update dgst to use
3015 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3016 alternative signing parameters such as X9.31 or PSS in the dgst
3020 *) Change ssl_cipher_apply_rule(), the internal function that does
3021 the work each time a ciphersuite string requests enabling
3022 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3023 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3024 the order of disabled ciphersuites such that those ciphersuites
3025 that most recently went from enabled to disabled not only stay
3026 in order with respect to each other, but also have higher priority
3027 than other disabled ciphersuites the next time ciphersuites are
3030 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3031 the same ciphersuites as with "HIGH" alone, but in a specific
3032 order where the PSK ciphersuites come first (since they are the
3033 most recently disabled ciphersuites when "HIGH" is parsed).
3035 Also, change ssl_create_cipher_list() (using this new
3036 funcionality) such that between otherwise identical
3037 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3041 *) Change ssl_create_cipher_list() so that it automatically
3042 arranges the ciphersuites in reasonable order before starting
3043 to process the rule string. Thus, the definition for "DEFAULT"
3044 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3045 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3046 This makes it much easier to arrive at a reasonable default order
3047 in applications for which anonymous ciphers are OK (meaning
3048 that you can't actually use DEFAULT).
3049 [Bodo Moeller; suggested by Victor Duchovni]
3051 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3052 processing) into multiple integers instead of setting
3053 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3054 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3055 (These masks as well as the individual bit definitions are hidden
3056 away into the non-exported interface ssl/ssl_locl.h, so this
3057 change to the definition of the SSL_CIPHER structure shouldn't
3058 affect applications.) This give us more bits for each of these
3059 categories, so there is no longer a need to coagulate AES128 and
3060 AES256 into a single algorithm bit, and to coagulate Camellia128
3061 and Camellia256 into a single algorithm bit, which has led to all
3064 Thus, among other things, the kludge introduced in 0.9.7m and
3065 0.9.8e for masking out AES256 independently of AES128 or masking
3066 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3068 With the change, we also introduce new ciphersuite aliases that
3069 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3073 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3074 Use the leftmost N bytes of the signature input if the input is
3075 larger than the prime q (with N being the size in bytes of q).
3078 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3079 it yet and it is largely untested.
3082 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3085 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3086 some compilers (gcc 4.2 and later) reject their use. Safestack is
3087 reimplemented. Update ASN1 to avoid use of legacy functions.
3090 *) Win32/64 targets are linked with Winsock2.
3093 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3094 to external functions. This can be used to increase CRL handling
3095 efficiency especially when CRLs are very large by (for example) storing
3096 the CRL revoked certificates in a database.
3099 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3100 new CRLs added to a directory can be used. New command line option
3101 -verify_return_error to s_client and s_server. This causes real errors
3102 to be returned by the verify callback instead of carrying on no matter
3103 what. This reflects the way a "real world" verify callback would behave.
3106 *) GOST engine, supporting several GOST algorithms and public key formats.
3107 Kindly donated by Cryptocom.
3110 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3111 partitioned by DP are handled but no indirect CRL or reason partitioning
3112 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3113 selected via a scoring technique which handles IDP and AKID in CRLs.
3116 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3117 will ultimately be used for all verify operations: this will remove the
3118 X509_STORE dependency on certificate verification and allow alternative
3119 lookup methods. X509_STORE based implementations of these two callbacks.
3122 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3123 Modify get_crl() to find a valid (unexpired) CRL if possible.
3126 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3127 this would be called X509_CRL_cmp() but that name is already used by
3128 a function that just compares CRL issuer names. Cache several CRL
3129 extensions in X509_CRL structure and cache CRLDP in X509.
3132 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3133 this maps equivalent X509_NAME structures into a consistent structure.
3134 Name comparison can then be performed rapidly using memcmp().
3137 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3141 *) Allow digests to supply their own micalg string for S/MIME type using
3142 the ctrl EVP_MD_CTRL_MICALG.
3145 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3146 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3147 ctrl. It can then customise the structure before and/or after signing
3151 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3152 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3153 to free up any added signature OIDs.
3156 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3157 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3158 digest and cipher tables. New options added to openssl utility:
3159 list-message-digest-algorithms and list-cipher-algorithms.
3162 *) Change the array representation of binary polynomials: the list
3163 of degrees of non-zero coefficients is now terminated with -1.
3164 Previously it was terminated with 0, which was also part of the
3165 value; thus, the array representation was not applicable to
3166 polynomials where t^0 has coefficient zero. This change makes
3167 the array representation useful in a more general context.
3170 *) Various modifications and fixes to SSL/TLS cipher string
3171 handling. For ECC, the code now distinguishes between fixed ECDH
3172 with RSA certificates on the one hand and with ECDSA certificates
3173 on the other hand, since these are separate ciphersuites. The
3174 unused code for Fortezza ciphersuites has been removed.
3176 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3177 (not "ECDHE"). For consistency with the code for DH
3178 certificates, use of ECDH certificates is now considered ECDH
3179 authentication, not RSA or ECDSA authentication (the latter is
3180 merely the CA's signing algorithm and not actively used in the
3183 The temporary ciphersuite alias "ECCdraft" is no longer
3184 available, and ECC ciphersuites are no longer excluded from "ALL"
3185 and "DEFAULT". The following aliases now exist for RFC 4492
3186 ciphersuites, most of these by analogy with the DH case:
3188 kECDHr - ECDH cert, signed with RSA
3189 kECDHe - ECDH cert, signed with ECDSA
3190 kECDH - ECDH cert (signed with either RSA or ECDSA)
3191 kEECDH - ephemeral ECDH
3192 ECDH - ECDH cert or ephemeral ECDH
3198 AECDH - anonymous ECDH
3199 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3203 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3204 Use correct micalg parameters depending on digest(s) in signed message.
3207 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3208 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3211 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3212 an engine to register a method. Add ENGINE lookups for methods and
3213 functional reference processing.
3216 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3217 EVP_{Sign,Verify}* which allow an application to customise the signature
3221 *) New -resign option to smime utility. This adds one or more signers
3222 to an existing PKCS#7 signedData structure. Also -md option to use an
3223 alternative message digest algorithm for signing.
3226 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3227 create PKCS7 structures containing multiple signers. Update smime
3228 application to support multiple signers.
3231 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3235 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3236 Reorganize PBE internals to lookup from a static table using NIDs,
3237 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3238 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3239 PRF which will be automatically used with PBES2.
3242 *) Replace the algorithm specific calls to generate keys in "req" with the
3246 *) Update PKCS#7 enveloped data routines to use new API. This is now
3247 supported by any public key method supporting the encrypt operation. A
3248 ctrl is added to allow the public key algorithm to examine or modify
3249 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3253 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3254 a default digest type to use. In most cases this will be SHA1 but some
3255 algorithms (such as GOST) need to specify an alternative digest. The
3256 return value indicates how strong the preference is 1 means optional and
3257 2 is mandatory (that is it is the only supported type). Modify
3258 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3259 use the default md. Update openssl utilities to use the default digest
3260 type for signing if it is not explicitly indicated.
3263 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3264 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3265 signing method from the key type. This effectively removes the link
3266 between digests and public key types.
3269 *) Add an OID cross reference table and utility functions. Its purpose is to
3270 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3271 rsaEncryption. This will allow some of the algorithm specific hackery
3272 needed to use the correct OID to be removed.
3275 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3276 structures for PKCS7_sign(). They are now set up by the relevant public
3280 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3283 *) Add support for key derivation (agreement) in the API, DH method and
3287 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3288 public and private key formats. As a side effect these add additional
3289 command line functionality not previously available: DSA signatures can be
3290 generated and verified using pkeyutl and DH key support and generation in
3295 [Oliver Tappe <zooey@hirschkaefer.de>]
3297 *) New make target "install_html_docs" installs HTML renditions of the
3299 [Oliver Tappe <zooey@hirschkaefer.de>]
3301 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3302 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3303 support key and parameter generation and add initial key generation
3304 functionality for RSA.
3307 *) Add functions for main EVP_PKEY_method operations. The undocumented
3308 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3309 EVP_PKEY_{encrypt,decrypt}_old.
3312 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3313 key API, doesn't do much yet.
3316 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3317 public key algorithms. New option to openssl utility:
3318 "list-public-key-algorithms" to print out info.
3321 *) Implement the Supported Elliptic Curves Extension for
3322 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3325 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3326 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3329 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3330 utilities such as rsa, dsa, dsaparam etc except they process any key
3334 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3335 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3336 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3340 *) Initial support for pluggable public key ASN1.
3341 De-spaghettify the public key ASN1 handling. Move public and private
3342 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3343 algorithm specific handling to a single module within the relevant
3344 algorithm directory. Add functions to allow (near) opaque processing
3345 of public and private key structures.
3348 *) Implement the Supported Point Formats Extension for
3349 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3352 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3353 for the psk identity [hint] and the psk callback functions to the
3354 SSL_SESSION, SSL and SSL_CTX structure.
3357 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3361 SSL_CTX_use_psk_identity_hint
3362 SSL_get_psk_identity_hint
3363 SSL_get_psk_identity
3364 SSL_use_psk_identity_hint
3366 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3368 *) Add RFC 3161 compliant time stamp request creation, response generation
3369 and response verification functionality.
3370 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3372 *) Add initial support for TLS extensions, specifically for the server_name
3373 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3374 have new members for a host name. The SSL data structure has an
3375 additional member SSL_CTX *initial_ctx so that new sessions can be
3376 stored in that context to allow for session resumption, even after the
3377 SSL has been switched to a new SSL_CTX in reaction to a client's
3378 server_name extension.
3380 New functions (subject to change):
3382 SSL_get_servername()
3383 SSL_get_servername_type()
3386 New CTRL codes and macros (subject to change):
3388 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3389 - SSL_CTX_set_tlsext_servername_callback()
3390 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3391 - SSL_CTX_set_tlsext_servername_arg()
3392 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3394 openssl s_client has a new '-servername ...' option.
3396 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3397 '-key2 ...', '-servername_fatal' (subject to change). This allows
3398 testing the HostName extension for a specific single host name ('-cert'
3399 and '-key' remain fallbacks for handshakes without HostName
3400 negotiation). If the unrecognized_name alert has to be sent, this by
3401 default is a warning; it becomes fatal with the '-servername_fatal'
3404 [Peter Sylvester, Remy Allais, Christophe Renou]
3406 *) Whirlpool hash implementation is added.
3409 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3410 bn(64,32). Because of instruction set limitations it doesn't have
3411 any negative impact on performance. This was done mostly in order
3412 to make it possible to share assembler modules, such as bn_mul_mont
3413 implementations, between 32- and 64-bit builds without hassle.
3416 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3417 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3421 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3422 dedicated Montgomery multiplication procedure, is introduced.
3423 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3424 "64-bit" performance on certain 32-bit targets.
3427 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3428 in SSL structures. New SSL ctrl to set maximum send fragment size.
3429 Save memory by seeting the I/O buffer sizes dynamically instead of
3430 using the maximum available value.
3433 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3434 in addition to the text details.
3437 *) Very, very preliminary EXPERIMENTAL support for printing of general
3438 ASN1 structures. This currently produces rather ugly output and doesn't
3439 handle several customised structures at all.
3442 *) Integrated support for PVK file format and some related formats such
3443 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3444 these in the 'rsa' and 'dsa' utilities.
3447 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3450 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3451 place for the (very old) "NETSCAPE" format certificates which are now
3452 handled using new ASN1 code equivalents.
3455 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3456 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3457 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3460 *) Modify CRL distribution points extension code to print out previously
3461 unsupported fields. Enhance extension setting code to allow setting of
3465 *) Add print and set support for Issuing Distribution Point CRL extension.
3468 *) Change 'Configure' script to enable Camellia by default.
3471 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3473 *) When rejecting SSL/TLS records due to an incorrect version number, never
3474 update s->server with a new major version number. As of
3475 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3476 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3477 the previous behavior could result in a read attempt at NULL when
3478 receiving specific incorrect SSL/TLS records once record payload
3479 protection is active. (CVE-2010-0740)
3480 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3482 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3483 could be crashed if the relevant tables were not present (e.g. chrooted).
3484 [Tomas Hoger <thoger@redhat.com>]
3486 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3488 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3489 [Martin Olsson, Neel Mehta]
3491 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3492 accommodate for stack sorting, always a write lock!).
3495 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3496 excessive delays in the RAND_poll(): over a minute. As a workaround
3497 include a time check in the inner Heap32Next loop too.
3500 *) The code that handled flushing of data in SSL/TLS originally used the
3501 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3502 the problem outlined in PR#1949. The fix suggested there however can
3503 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3504 of Apache). So instead simplify the code to flush unconditionally.
3505 This should be fine since flushing with no data to flush is a no op.
3508 *) Handle TLS versions 2.0 and later properly and correctly use the
3509 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3510 off ancient servers have a habit of sticking around for a while...
3513 *) Modify compression code so it frees up structures without using the
3514 ex_data callbacks. This works around a problem where some applications
3515 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3516 restarting) then use compression (e.g. SSL with compression) later.
3517 This results in significant per-connection memory leaks and
3518 has caused some security issues including CVE-2008-1678 and
3522 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3523 change when encrypting or decrypting.
3526 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3527 connect and renegotiate with servers which do not support RI.
3528 Until RI is more widely deployed this option is enabled by default.
3531 *) Add "missing" ssl ctrls to clear options and mode.
3534 *) If client attempts to renegotiate and doesn't support RI respond with
3535 a no_renegotiation alert as required by RFC5746. Some renegotiating
3536 TLS clients will continue a connection gracefully when they receive
3537 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3538 waiting for a server hello which it will never receive. Now we treat a
3539 received no_renegotiation alert as a fatal error. This is because
3540 applications requesting a renegotiation might well expect it to succeed
3541 and would have no code in place to handle the server denying it so the
3542 only safe thing to do is to terminate the connection.
3545 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3546 peer supports secure renegotiation and 0 otherwise. Print out peer
3547 renegotiation support in s_client/s_server.
3550 *) Replace the highly broken and deprecated SPKAC certification method with
3551 the updated NID creation version. This should correctly handle UTF8.
3554 *) Implement RFC5746. Re-enable renegotiation but require the extension
3555 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3556 turns out to be a bad idea. It has been replaced by
3557 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3558 SSL_CTX_set_options(). This is really not recommended unless you
3559 know what you are doing.
3560 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3562 *) Fixes to stateless session resumption handling. Use initial_ctx when
3563 issuing and attempting to decrypt tickets in case it has changed during
3564 servername handling. Use a non-zero length session ID when attempting
3565 stateless session resumption: this makes it possible to determine if
3566 a resumption has occurred immediately after receiving server hello
3567 (several places in OpenSSL subtly assume this) instead of later in
3571 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3572 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3573 fixes for a few places where the return code is not checked
3575 [Julia Lawall <julia@diku.dk>]
3577 *) Add --strict-warnings option to Configure script to include devteam
3578 warnings in other configurations.
3581 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3582 makes it possible to install openssl libraries in locations which
3583 have names other than "lib", for example "/usr/lib64" which some
3585 [Steve Henson, based on patch from Jeremy Utley]
3587 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3588 X690 8.9.12 and can produce some misleading textual output of OIDs.
3589 [Steve Henson, reported by Dan Kaminsky]
3591 *) Delete MD2 from algorithm tables. This follows the recommendation in
3592 several standards that it is not used in new applications due to
3593 several cryptographic weaknesses. For binary compatibility reasons
3594 the MD2 API is still compiled in by default.
3597 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3601 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3602 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3604 [Guenter <lists@gknw.net>]
3606 *) Fix the server certificate chain building code to use X509_verify_cert(),
3607 it used to have an ad-hoc builder which was unable to cope with anything
3608 other than a simple chain.
3609 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3611 *) Don't check self signed certificate signatures in X509_verify_cert()
3612 by default (a flag can override this): it just wastes time without
3613 adding any security. As a useful side effect self signed root CAs
3614 with non-FIPS digests are now usable in FIPS mode.
3617 *) In dtls1_process_out_of_seq_message() the check if the current message
3618 is already buffered was missing. For every new message was memory
3619 allocated, allowing an attacker to perform an denial of service attack
3620 with sending out of seq handshake messages until there is no memory
3621 left. Additionally every future messege was buffered, even if the
3622 sequence number made no sense and would be part of another handshake.
3623 So only messages with sequence numbers less than 10 in advance will be
3624 buffered. (CVE-2009-1378)
3625 [Robin Seggelmann, discovered by Daniel Mentz]
3627 *) Records are buffered if they arrive with a future epoch to be
3628 processed after finishing the corresponding handshake. There is
3629 currently no limitation to this buffer allowing an attacker to perform
3630 a DOS attack with sending records with future epochs until there is no
3631 memory left. This patch adds the pqueue_size() function to determine
3632 the size of a buffer and limits the record buffer to 100 entries.
3634 [Robin Seggelmann, discovered by Daniel Mentz]
3636 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3637 parent structure is freed. (CVE-2009-1379)
3640 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3641 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3644 [Ilya O. <vrghost@gmail.com>]
3646 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3648 *) Disable renegotiation completely - this fixes a severe security
3649 problem (CVE-2009-3555) at the cost of breaking all
3650 renegotiation. Renegotiation can be re-enabled by setting
3651 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3652 run-time. This is really not recommended unless you know what
3656 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3658 *) Don't set val to NULL when freeing up structures, it is freed up by
3659 underlying code. If sizeof(void *) > sizeof(long) this can result in
3660 zeroing past the valid field. (CVE-2009-0789)
3661 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3663 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3664 checked correctly. This would allow some invalid signed attributes to
3665 appear to verify correctly. (CVE-2009-0591)
3666 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3668 *) Reject UniversalString and BMPString types with invalid lengths. This
3669 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3670 a legal length. (CVE-2009-0590)
3673 *) Set S/MIME signing as the default purpose rather than setting it
3674 unconditionally. This allows applications to override it at the store
3678 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3679 to handle some structures.
3682 *) Improve efficiency of mem_gets: don't search whole buffer each time
3684 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3686 *) New -hex option for openssl rand.
3689 *) Print out UTF8String and NumericString when parsing ASN1.
3692 *) Support NumericString type for name components.
3695 *) Allow CC in the environment to override the automatically chosen
3696 compiler. Note that nothing is done to ensure flags work with the
3700 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3702 *) Properly check EVP_VerifyFinal() and similar return values
3704 [Ben Laurie, Bodo Moeller, Google Security Team]
3706 *) Enable TLS extensions by default.
3709 *) Allow the CHIL engine to be loaded, whether the application is
3710 multithreaded or not. (This does not release the developer from the
3711 obligation to set up the dynamic locking callbacks.)
3712 [Sander Temme <sander@temme.net>]
3714 *) Use correct exit code if there is an error in dgst command.
3715 [Steve Henson; problem pointed out by Roland Dirlewanger]
3717 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3718 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3721 *) Add experimental JPAKE support, including demo authentication in
3722 s_client and s_server.
3725 *) Set the comparison function in v3_addr_canonize().
3726 [Rob Austein <sra@hactrn.net>]
3728 *) Add support for XMPP STARTTLS in s_client.
3729 [Philip Paeps <philip@freebsd.org>]
3731 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3732 to ensure that even with this option, only ciphersuites in the
3733 server's preference list will be accepted. (Note that the option
3734 applies only when resuming a session, so the earlier behavior was
3735 just about the algorithm choice for symmetric cryptography.)
3738 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3740 *) Fix NULL pointer dereference if a DTLS server received
3741 ChangeCipherSpec as first record (CVE-2009-1386).
3744 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3745 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3748 *) The fix in 0.9.8c that supposedly got rid of unsafe
3749 double-checked locking was incomplete for RSA blinding,
3750 addressing just one layer of what turns out to have been
3751 doubly unsafe triple-checked locking.
3753 So now fix this for real by retiring the MONT_HELPER macro
3754 in crypto/rsa/rsa_eay.c.
3756 [Bodo Moeller; problem pointed out by Marius Schilder]
3758 *) Various precautionary measures:
3760 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3762 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3763 (NB: This would require knowledge of the secret session ticket key
3764 to exploit, in which case you'd be SOL either way.)
3766 - Change bn_nist.c so that it will properly handle input BIGNUMs
3767 outside the expected range.
3769 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3772 [Neel Mehta, Bodo Moeller]
3774 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3775 the load fails. Useful for distros.
3776 [Ben Laurie and the FreeBSD team]
3778 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3781 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3784 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3786 This work was sponsored by Logica.
3789 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3790 keystores. Support for SSL/TLS client authentication too.
3791 Not compiled unless enable-capieng specified to Configure.
3793 This work was sponsored by Logica.
3796 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3797 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3798 attribute creation routines such as certifcate requests and PKCS#12
3802 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3804 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3805 handshake which could lead to a cilent crash as found using the
3806 Codenomicon TLS test suite (CVE-2008-1672)
3807 [Steve Henson, Mark Cox]
3809 *) Fix double free in TLS server name extensions which could lead to
3810 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3813 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3815 Clear the error queue to ensure that error entries left from
3816 older function calls do not interfere with the correct operation.
3817 [Lutz Jaenicke, Erik de Castro Lopo]
3819 *) Remove root CA certificates of commercial CAs:
3821 The OpenSSL project does not recommend any specific CA and does not
3822 have any policy with respect to including or excluding any CA.
3823 Therefore it does not make any sense to ship an arbitrary selection
3824 of root CA certificates with the OpenSSL software.
3827 *) RSA OAEP patches to fix two separate invalid memory reads.
3828 The first one involves inputs when 'lzero' is greater than
3829 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3830 before the beginning of from). The second one involves inputs where
3831 the 'db' section contains nothing but zeroes (there is a one-byte
3832 invalid read after the end of 'db').
3833 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3835 *) Partial backport from 0.9.9-dev:
3837 Introduce bn_mul_mont (dedicated Montgomery multiplication
3838 procedure) as a candidate for BIGNUM assembler implementation.
3839 While 0.9.9-dev uses assembler for various architectures, only
3840 x86_64 is available by default here in the 0.9.8 branch, and
3841 32-bit x86 is available through a compile-time setting.
3843 To try the 32-bit x86 assembler implementation, use Configure
3844 option "enable-montasm" (which exists only for this backport).
3846 As "enable-montasm" for 32-bit x86 disclaims code stability
3847 anyway, in this constellation we activate additional code
3848 backported from 0.9.9-dev for further performance improvements,
3849 namely BN_from_montgomery_word. (To enable this otherwise,
3850 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3852 [Andy Polyakov (backport partially by Bodo Moeller)]
3854 *) Add TLS session ticket callback. This allows an application to set
3855 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3856 values. This is useful for key rollover for example where several key
3857 sets may exist with different names.
3860 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3861 This was broken until now in 0.9.8 releases, such that the only way
3862 a registered ENGINE could be used (assuming it initialises
3863 successfully on the host) was to explicitly set it as the default
3864 for the relevant algorithms. This is in contradiction with 0.9.7
3865 behaviour and the documentation. With this fix, when an ENGINE is
3866 registered into a given algorithm's table of implementations, the
3867 'uptodate' flag is reset so that auto-discovery will be used next
3868 time a new context for that algorithm attempts to select an
3870 [Ian Lister (tweaked by Geoff Thorpe)]
3872 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3873 implemention in the following ways:
3875 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3878 Lack of BER streaming support means one pass streaming processing is
3879 only supported if data is detached: setting the streaming flag is
3880 ignored for embedded content.
3882 CMS support is disabled by default and must be explicitly enabled
3883 with the enable-cms configuration option.
3886 *) Update the GMP engine glue to do direct copies between BIGNUM and
3887 mpz_t when openssl and GMP use the same limb size. Otherwise the
3888 existing "conversion via a text string export" trick is still used.
3889 [Paul Sheer <paulsheer@gmail.com>]
3891 *) Zlib compression BIO. This is a filter BIO which compressed and
3892 uncompresses any data passed through it.
3895 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3896 RFC3394 compatible AES key wrapping.
3899 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3900 sets string data without copying. X509_ALGOR_set0() and
3901 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3902 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3903 from an X509_ATTRIBUTE structure optionally checking it occurs only
3904 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3908 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3909 to get the expected BN_FLG_CONSTTIME behavior.
3910 [Bodo Moeller (Google)]
3914 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3915 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3916 - added some more tests to do_tests.pl
3917 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3918 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3919 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3920 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3921 - various changes to netware.pl to enable gcc-cross builds on Win32
3923 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3924 - various changes to fix missing prototype warnings
3925 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3926 - added AES, WHIRLPOOL and CPUID assembler code to build files
3927 - added missing AES assembler make rules to mk1mf.pl
3928 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3929 [Guenter Knauf <eflash@gmx.net>]
3931 *) Implement certificate status request TLS extension defined in RFC3546.
3932 A client can set the appropriate parameters and receive the encoded
3933 OCSP response via a callback. A server can query the supplied parameters
3934 and set the encoded OCSP response in the callback. Add simplified examples
3935 to s_client and s_server.
3938 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3940 *) Fix various bugs:
3941 + Binary incompatibility of ssl_ctx_st structure
3942 + DTLS interoperation with non-compliant servers
3943 + Don't call get_session_cb() without proposed session
3944 + Fix ia64 assembler code
3945 [Andy Polyakov, Steve Henson]
3947 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3949 *) DTLS Handshake overhaul. There were longstanding issues with
3950 OpenSSL DTLS implementation, which were making it impossible for
3951 RFC 4347 compliant client to communicate with OpenSSL server.
3952 Unfortunately just fixing these incompatibilities would "cut off"
3953 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3954 server keeps tolerating non RFC compliant syntax. The opposite is
3955 not true, 0.9.8f client can not communicate with earlier server.
3956 This update even addresses CVE-2007-4995.
3959 *) Changes to avoid need for function casts in OpenSSL: some compilers
3960 (gcc 4.2 and later) reject their use.
3961 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3964 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3965 RFC4507bis. The encrypted ticket format is an encrypted encoded
3966 SSL_SESSION structure, that way new session features are automatically
3969 If a client application caches session in an SSL_SESSION structure
3970 support is transparent because tickets are now stored in the encoded
3973 The SSL_CTX structure automatically generates keys for ticket
3974 protection in servers so again support should be possible
3975 with no application modification.
3977 If a client or server wishes to disable RFC4507 support then the option
3978 SSL_OP_NO_TICKET can be set.
3980 Add a TLS extension debugging callback to allow the contents of any client
3981 or server extensions to be examined.
3983 This work was sponsored by Google.
3986 *) Add initial support for TLS extensions, specifically for the server_name
3987 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3988 have new members for a host name. The SSL data structure has an
3989 additional member SSL_CTX *initial_ctx so that new sessions can be
3990 stored in that context to allow for session resumption, even after the
3991 SSL has been switched to a new SSL_CTX in reaction to a client's
3992 server_name extension.
3994 New functions (subject to change):
3996 SSL_get_servername()
3997 SSL_get_servername_type()
4000 New CTRL codes and macros (subject to change):
4002 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4003 - SSL_CTX_set_tlsext_servername_callback()
4004 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4005 - SSL_CTX_set_tlsext_servername_arg()
4006 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4008 openssl s_client has a new '-servername ...' option.
4010 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4011 '-key2 ...', '-servername_fatal' (subject to change). This allows
4012 testing the HostName extension for a specific single host name ('-cert'
4013 and '-key' remain fallbacks for handshakes without HostName
4014 negotiation). If the unrecognized_name alert has to be sent, this by
4015 default is a warning; it becomes fatal with the '-servername_fatal'
4018 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4020 *) Add AES and SSE2 assembly language support to VC++ build.
4023 *) Mitigate attack on final subtraction in Montgomery reduction.
4026 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4027 (which previously caused an internal error).
4030 *) Squeeze another 10% out of IGE mode when in != out.
4033 *) AES IGE mode speedup.
4034 [Dean Gaudet (Google)]
4036 *) Add the Korean symmetric 128-bit cipher SEED (see
4037 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4038 add SEED ciphersuites from RFC 4162:
4040 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4041 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4042 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4043 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4045 To minimize changes between patchlevels in the OpenSSL 0.9.8
4046 series, SEED remains excluded from compilation unless OpenSSL
4047 is configured with 'enable-seed'.
4048 [KISA, Bodo Moeller]
4050 *) Mitigate branch prediction attacks, which can be practical if a
4051 single processor is shared, allowing a spy process to extract
4052 information. For detailed background information, see
4053 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4054 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4055 and Necessary Software Countermeasures"). The core of the change
4056 are new versions BN_div_no_branch() and
4057 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4058 respectively, which are slower, but avoid the security-relevant
4059 conditional branches. These are automatically called by BN_div()
4060 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4061 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4062 remove a conditional branch.
4064 BN_FLG_CONSTTIME is the new name for the previous
4065 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4066 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4067 in the exponent causes BN_mod_exp_mont() to use the alternative
4068 implementation in BN_mod_exp_mont_consttime().) The old name
4069 remains as a deprecated alias.
4071 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4072 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4073 constant-time implementations for more than just exponentiation.
4074 Here too the old name is kept as a deprecated alias.
4076 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4077 the BN_BLINDING structure gets an independent copy of the
4078 modulus. This means that the previous "BIGNUM *m" argument to
4079 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4080 essentially becomes "const BIGNUM *m", although we can't actually
4081 change this in the header file before 0.9.9. It allows
4082 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4083 enable BN_FLG_CONSTTIME.
4085 [Matthew D Wood (Intel Corp)]
4087 *) In the SSL/TLS server implementation, be strict about session ID
4088 context matching (which matters if an application uses a single
4089 external cache for different purposes). Previously,
4090 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4091 set. This did ensure strict client verification, but meant that,
4092 with applications using a single external cache for quite
4093 different requirements, clients could circumvent ciphersuite
4094 restrictions for a given session ID context by starting a session
4095 in a different context.
4098 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4099 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4100 authentication-only ciphersuites.
4103 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4104 not complete and could lead to a possible single byte overflow
4105 (CVE-2007-5135) [Ben Laurie]
4107 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4109 *) Since AES128 and AES256 (and similarly Camellia128 and
4110 Camellia256) share a single mask bit in the logic of
4111 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4112 kludge to work properly if AES128 is available and AES256 isn't
4113 (or if Camellia128 is available and Camellia256 isn't).
4116 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4117 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4118 When a point or a seed is encoded in a BIT STRING, we need to
4119 prevent the removal of trailing zero bits to get the proper DER
4120 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4121 of a NamedBitList, for which trailing 0 bits need to be removed.)
4124 *) Have SSL/TLS server implementation tolerate "mismatched" record
4125 protocol version while receiving ClientHello even if the
4126 ClientHello is fragmented. (The server can't insist on the
4127 particular protocol version it has chosen before the ServerHello
4128 message has informed the client about his choice.)
4131 *) Add RFC 3779 support.
4132 [Rob Austein for ARIN, Ben Laurie]
4134 *) Load error codes if they are not already present instead of using a
4135 static variable. This allows them to be cleanly unloaded and reloaded.
4136 Improve header file function name parsing.
4139 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4140 or CAPABILITY handshake as required by RFCs.
4143 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4145 *) Introduce limits to prevent malicious keys being able to
4146 cause a denial of service. (CVE-2006-2940)
4147 [Steve Henson, Bodo Moeller]
4149 *) Fix ASN.1 parsing of certain invalid structures that can result
4150 in a denial of service. (CVE-2006-2937) [Steve Henson]
4152 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4153 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4155 *) Fix SSL client code which could crash if connecting to a
4156 malicious SSLv2 server. (CVE-2006-4343)
4157 [Tavis Ormandy and Will Drewry, Google Security Team]
4159 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4160 match only those. Before that, "AES256-SHA" would be interpreted
4161 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4162 the same strength classification in 0.9.7h) as we currently only
4163 have a single AES bit in the ciphersuite description bitmap.
4164 That change, however, also applied to ciphersuite strings such as
4165 "RC4-MD5" that intentionally matched multiple ciphersuites --
4166 namely, SSL 2.0 ciphersuites in addition to the more common ones
4167 from SSL 3.0/TLS 1.0.
4169 So we change the selection algorithm again: Naming an explicit
4170 ciphersuite selects this one ciphersuite, and any other similar
4171 ciphersuite (same bitmap) from *other* protocol versions.
4172 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4173 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4175 Since SSL 2.0 does not have any ciphersuites for which the
4176 128/256 bit distinction would be relevant, this works for now.
4177 The proper fix will be to use different bits for AES128 and
4178 AES256, which would have avoided the problems from the beginning;
4179 however, bits are scarce, so we can only do this in a new release
4180 (not just a patchlevel) when we can change the SSL_CIPHER
4181 definition to split the single 'unsigned long mask' bitmap into
4182 multiple values to extend the available space.
4186 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4188 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4189 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4191 *) Add AES IGE and biIGE modes.
4194 *) Change the Unix randomness entropy gathering to use poll() when
4195 possible instead of select(), since the latter has some
4196 undesirable limitations.
4197 [Darryl Miles via Richard Levitte and Bodo Moeller]
4199 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4200 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4201 cannot be implicitly activated as part of, e.g., the "AES" alias.
4202 However, please upgrade to OpenSSL 0.9.9[-dev] for
4203 non-experimental use of the ECC ciphersuites to get TLS extension
4204 support, which is required for curve and point format negotiation
4205 to avoid potential handshake problems.
4208 *) Disable rogue ciphersuites:
4210 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4211 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4212 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4214 The latter two were purportedly from
4215 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4218 Also deactivate the remaining ciphersuites from
4219 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4220 unofficial, and the ID has long expired.
4223 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4224 dual-core machines) and other potential thread-safety issues.
4227 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4228 versions), which is now available for royalty-free use
4229 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4230 Also, add Camellia TLS ciphersuites from RFC 4132.
4232 To minimize changes between patchlevels in the OpenSSL 0.9.8
4233 series, Camellia remains excluded from compilation unless OpenSSL
4234 is configured with 'enable-camellia'.
4237 *) Disable the padding bug check when compression is in use. The padding
4238 bug check assumes the first packet is of even length, this is not
4239 necessarily true if compresssion is enabled and can result in false
4240 positives causing handshake failure. The actual bug test is ancient
4241 code so it is hoped that implementations will either have fixed it by
4242 now or any which still have the bug do not support compression.
4245 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4247 *) When applying a cipher rule check to see if string match is an explicit
4248 cipher suite and only match that one cipher suite if it is.
4251 *) Link in manifests for VC++ if needed.
4252 [Austin Ziegler <halostatue@gmail.com>]
4254 *) Update support for ECC-based TLS ciphersuites according to
4255 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4256 TLS extensions, which are supported starting with the 0.9.9
4257 branch, not in the OpenSSL 0.9.8 branch).
4260 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4261 opaque EVP_CIPHER_CTX handling.
4264 *) Fixes and enhancements to zlib compression code. We now only use
4265 "zlib1.dll" and use the default __cdecl calling convention on Win32
4266 to conform with the standards mentioned here:
4267 http://www.zlib.net/DLL_FAQ.txt
4268 Static zlib linking now works on Windows and the new --with-zlib-include
4269 --with-zlib-lib options to Configure can be used to supply the location
4270 of the headers and library. Gracefully handle case where zlib library
4274 *) Several fixes and enhancements to the OID generation code. The old code
4275 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4276 handle numbers larger than ULONG_MAX, truncated printing and had a
4277 non standard OBJ_obj2txt() behaviour.
4280 *) Add support for building of engines under engine/ as shared libraries
4281 under VC++ build system.
4284 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4285 Hopefully, we will not see any false combination of paths any more.
4288 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4290 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4291 (part of SSL_OP_ALL). This option used to disable the
4292 countermeasure against man-in-the-middle protocol-version
4293 rollback in the SSL 2.0 server implementation, which is a bad
4294 idea. (CVE-2005-2969)
4296 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4297 for Information Security, National Institute of Advanced Industrial
4298 Science and Technology [AIST], Japan)]
4300 *) Add two function to clear and return the verify parameter flags.
4303 *) Keep cipherlists sorted in the source instead of sorting them at
4304 runtime, thus removing the need for a lock.
4307 *) Avoid some small subgroup attacks in Diffie-Hellman.
4308 [Nick Mathewson and Ben Laurie]
4310 *) Add functions for well-known primes.
4313 *) Extended Windows CE support.
4314 [Satoshi Nakamura and Andy Polyakov]
4316 *) Initialize SSL_METHOD structures at compile time instead of during
4317 runtime, thus removing the need for a lock.
4320 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4321 attempting to decrypt each encrypted key in turn. Add support to
4325 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4327 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4330 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4333 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4334 key into the same file any more.
4337 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4340 *) Add -utf8 command line and config file option to 'ca'.
4341 [Stefan <stf@udoma.org]
4343 *) Removed the macro des_crypt(), as it seems to conflict with some
4344 libraries. Use DES_crypt().
4347 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4348 involves renaming the source and generated shared-libs for
4349 both. The engines will accept the corrected or legacy ids
4350 ('ncipher' and '4758_cca' respectively) when binding. NB,
4351 this only applies when building 'shared'.
4352 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4354 *) Add attribute functions to EVP_PKEY structure. Modify
4355 PKCS12_create() to recognize a CSP name attribute and
4356 use it. Make -CSP option work again in pkcs12 utility.
4359 *) Add new functionality to the bn blinding code:
4360 - automatic re-creation of the BN_BLINDING parameters after
4361 a fixed number of uses (currently 32)
4362 - add new function for parameter creation
4363 - introduce flags to control the update behaviour of the
4364 BN_BLINDING parameters
4365 - hide BN_BLINDING structure
4366 Add a second BN_BLINDING slot to the RSA structure to improve
4367 performance when a single RSA object is shared among several
4371 *) Add support for DTLS.
4372 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4374 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4375 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4378 *) Remove buggy and incomplete DH cert support from
4379 ssl/ssl_rsa.c and ssl/s3_both.c
4382 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4383 the apps/openssl applications.
4386 *) Compile clean with "-Wall -Wmissing-prototypes
4387 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4388 DEBUG_SAFESTACK must also be set.
4391 *) Change ./Configure so that certain algorithms can be disabled by default.
4392 The new counterpiece to "no-xxx" is "enable-xxx".
4394 The patented RC5 and MDC2 algorithms will now be disabled unless
4395 "enable-rc5" and "enable-mdc2", respectively, are specified.
4397 (IDEA remains enabled despite being patented. This is because IDEA
4398 is frequently required for interoperability, and there is no license
4399 fee for non-commercial use. As before, "no-idea" can be used to
4400 avoid this algorithm.)
4404 *) Add processing of proxy certificates (see RFC 3820). This work was
4405 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4406 EGEE (Enabling Grids for E-science in Europe).
4409 *) RC4 performance overhaul on modern architectures/implementations, such
4410 as Intel P4, IA-64 and AMD64.
4413 *) New utility extract-section.pl. This can be used specify an alternative
4414 section number in a pod file instead of having to treat each file as
4415 a separate case in Makefile. This can be done by adding two lines to the
4418 =for comment openssl_section:XXX
4420 The blank line is mandatory.
4424 *) New arguments -certform, -keyform and -pass for s_client and s_server
4425 to allow alternative format key and certificate files and passphrase
4429 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4430 update associated structures and add various utility functions.
4432 Add new policy related verify parameters, include policy checking in
4433 standard verify code. Enhance 'smime' application with extra parameters
4434 to support policy checking and print out.
4437 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4438 Nehemiah processors. These extensions support AES encryption in hardware
4439 as well as RNG (though RNG support is currently disabled).
4440 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4442 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4445 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4446 [Andy Polyakov and a number of other people]
4448 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4449 implementation contributed by IBM.
4450 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4452 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4453 exponent rather than 'unsigned long'. There is a corresponding change to
4454 the new 'rsa_keygen' element of the RSA_METHOD structure.
4455 [Jelte Jansen, Geoff Thorpe]
4457 *) Functionality for creating the initial serial number file is now
4458 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4460 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4461 number file to 1, which is bound to cause problems. To avoid
4462 the problems while respecting compatibility between different 0.9.7
4463 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4464 CA.pl for serial number initialization. With the new release 0.9.8,
4465 we can fix the problem directly in the 'ca' utility.)
4468 *) Reduced header interdepencies by declaring more opaque objects in
4469 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4470 give fewer recursive includes, which could break lazy source code - so
4471 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4472 developers should define this symbol when building and using openssl to
4473 ensure they track the recommended behaviour, interfaces, [etc], but
4474 backwards-compatible behaviour prevails when this isn't defined.
4477 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4480 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4481 This will generate a random key of the appropriate length based on the
4482 cipher context. The EVP_CIPHER can provide its own random key generation
4483 routine to support keys of a specific form. This is used in the des and
4484 3des routines to generate a key of the correct parity. Update S/MIME
4485 code to use new functions and hence generate correct parity DES keys.
4486 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4487 valid (weak or incorrect parity).
4490 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4491 as looking them up. This is useful when the verified structure may contain
4492 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4493 present unless the new PKCS7_NO_CRL flag is asserted.
4496 *) Extend ASN1 oid configuration module. It now additionally accepts the
4499 shortName = some long name, 1.2.3.4
4502 *) Reimplemented the BN_CTX implementation. There is now no more static
4503 limitation on the number of variables it can handle nor the depth of the
4504 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4505 information can now expand as required, and rather than having a single
4506 static array of bignums, BN_CTX now uses a linked-list of such arrays
4507 allowing it to expand on demand whilst maintaining the usefulness of
4508 BN_CTX's "bundling".
4511 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4512 to allow all RSA operations to function using a single BN_CTX.
4515 *) Preliminary support for certificate policy evaluation and checking. This
4516 is initially intended to pass the tests outlined in "Conformance Testing
4517 of Relying Party Client Certificate Path Processing Logic" v1.07.
4520 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4521 remained unused and not that useful. A variety of other little bignum
4522 tweaks and fixes have also been made continuing on from the audit (see
4526 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4527 associated ASN1, EVP and SSL functions and old ASN1 macros.
4530 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4531 and this should never fail. So the return value from the use of
4532 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4533 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4536 *) BN_CTX_get() should return zero-valued bignums, providing the same
4537 initialised value as BN_new().
4538 [Geoff Thorpe, suggested by Ulf Möller]
4540 *) Support for inhibitAnyPolicy certificate extension.
4543 *) An audit of the BIGNUM code is underway, for which debugging code is
4544 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4545 is considered valid when processing BIGNUMs, and causes execution to
4546 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4547 further steps are taken to deliberately pollute unused data in BIGNUM
4548 structures to try and expose faulty code further on. For now, openssl will
4549 (in its default mode of operation) continue to tolerate the inconsistent
4550 forms that it has tolerated in the past, but authors and packagers should
4551 consider trying openssl and their own applications when compiled with
4552 these debugging symbols defined. It will help highlight potential bugs in
4553 their own code, and will improve the test coverage for OpenSSL itself. At
4554 some point, these tighter rules will become openssl's default to improve
4555 maintainability, though the assert()s and other overheads will remain only
4556 in debugging configurations. See bn.h for more details.
4557 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4559 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4560 that can only be obtained through BN_CTX_new() (which implicitly
4561 initialises it). The presence of this function only made it possible
4562 to overwrite an existing structure (and cause memory leaks).
4565 *) Because of the callback-based approach for implementing LHASH as a
4566 template type, lh_insert() adds opaque objects to hash-tables and
4567 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4568 to clean up those corresponding objects before destroying the hash table
4569 (and losing the object pointers). So some over-zealous constifications in
4570 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4571 objects as "const" and the lh_doall[_arg] callback wrappers are not
4572 prototyped to have "const" restrictions on the object pointers they are
4573 given (and so aren't required to cast them away any more).
4576 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4577 (speed) prefers to use its own implementation. The two implementations
4578 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4579 its object type properly exposed (MS_TM) instead of casting to/from "char
4580 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4581 aren't necessarily the greatest nomenclatures - but this is what was used
4582 internally to the implementation so I've used that for now.
4585 *) Ensure that deprecated functions do not get compiled when
4586 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4587 the self-tests were still using deprecated key-generation functions so
4588 these have been updated also.
4591 *) Reorganise PKCS#7 code to separate the digest location functionality
4592 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4593 New function PKCS7_set_digest() to set the digest type for PKCS#7
4594 digestedData type. Add additional code to correctly generate the
4595 digestedData type and add support for this type in PKCS7 initialization
4599 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4600 structure of type "other".
4603 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4604 sure the loop does correctly stop and breaking ("division by zero")
4605 modulus operations are not performed. The (pre-generated) prime
4606 table crypto/bn/bn_prime.h was already correct, but it could not be
4607 re-generated on some platforms because of the "division by zero"
4608 situation in the script.
4609 [Ralf S. Engelschall]
4611 *) Update support for ECC-based TLS ciphersuites according to
4612 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4613 SHA-1 now is only used for "small" curves (where the
4614 representation of a field element takes up to 24 bytes); for
4615 larger curves, the field element resulting from ECDH is directly
4616 used as premaster secret.
4617 [Douglas Stebila (Sun Microsystems Laboratories)]
4619 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4620 curve secp160r1 to the tests.
4621 [Douglas Stebila (Sun Microsystems Laboratories)]
4623 *) Add the possibility to load symbols globally with DSO.
4624 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4626 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4627 control of the error stack.
4630 *) Add support for STORE in ENGINE.
4633 *) Add the STORE type. The intention is to provide a common interface
4634 to certificate and key stores, be they simple file-based stores, or
4635 HSM-type store, or LDAP stores, or...
4636 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4639 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4640 pass a list of arguments to any function as well as provide a way
4641 for a function to pass data back to the caller.
4644 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4645 works like BUF_strdup() but can be used to duplicate a portion of
4646 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4650 *) Add the function sk_find_ex() which works like sk_find(), but will
4651 return an index to an element even if an exact match couldn't be
4652 found. The index is guaranteed to point at the element where the
4653 searched-for key would be inserted to preserve sorting order.
4656 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4657 takes an extra flags argument for optional functionality. Currently,
4658 the following flags are defined:
4660 OBJ_BSEARCH_VALUE_ON_NOMATCH
4661 This one gets OBJ_bsearch_ex() to return a pointer to the first
4662 element where the comparing function returns a negative or zero
4665 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4666 This one gets OBJ_bsearch_ex() to return a pointer to the first
4667 element where the comparing function returns zero. This is useful
4668 if there are more than one element where the comparing function
4672 *) Make it possible to create self-signed certificates with 'openssl ca'
4673 in such a way that the self-signed certificate becomes part of the
4674 CA database and uses the same mechanisms for serial number generation
4675 as all other certificate signing. The new flag '-selfsign' enables
4676 this functionality. Adapt CA.sh and CA.pl.in.
4679 *) Add functionality to check the public key of a certificate request
4680 against a given private. This is useful to check that a certificate
4681 request can be signed by that key (self-signing).
4684 *) Make it possible to have multiple active certificates with the same
4685 subject in the CA index file. This is done only if the keyword
4686 'unique_subject' is set to 'no' in the main CA section (default
4687 if 'CA_default') of the configuration file. The value is saved
4688 with the database itself in a separate index attribute file,
4689 named like the index file with '.attr' appended to the name.
4692 *) Generate muti valued AVAs using '+' notation in config files for
4696 *) Support for nameConstraints certificate extension.
4699 *) Support for policyConstraints certificate extension.
4702 *) Support for policyMappings certificate extension.
4705 *) Make sure the default DSA_METHOD implementation only uses its
4706 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4707 and change its own handlers to be NULL so as to remove unnecessary
4708 indirection. This lets alternative implementations fallback to the
4709 default implementation more easily.
4712 *) Support for directoryName in GeneralName related extensions
4716 *) Make it possible to link applications using Makefile.shared.
4717 Make that possible even when linking against static libraries!
4720 *) Support for single pass processing for S/MIME signing. This now
4721 means that S/MIME signing can be done from a pipe, in addition
4722 cleartext signing (multipart/signed type) is effectively streaming
4723 and the signed data does not need to be all held in memory.
4725 This is done with a new flag PKCS7_STREAM. When this flag is set
4726 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4727 is done after the data is output (and digests calculated) in
4728 SMIME_write_PKCS7().
4731 *) Add full support for -rpath/-R, both in shared libraries and
4732 applications, at least on the platforms where it's known how
4736 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4737 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4738 will now compute a table of multiples of the generator that
4739 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4740 faster (notably in the case of a single point multiplication,
4741 scalar * generator).
4742 [Nils Larsch, Bodo Moeller]
4744 *) IPv6 support for certificate extensions. The various extensions
4745 which use the IP:a.b.c.d can now take IPv6 addresses using the
4746 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4750 *) Added an ENGINE that implements RSA by performing private key
4751 exponentiations with the GMP library. The conversions to and from
4752 GMP's mpz_t format aren't optimised nor are any montgomery forms
4753 cached, and on x86 it appears OpenSSL's own performance has caught up.
4754 However there are likely to be other architectures where GMP could
4755 provide a boost. This ENGINE is not built in by default, but it can be
4756 specified at Configure time and should be accompanied by the necessary
4757 linker additions, eg;
4758 ./config -DOPENSSL_USE_GMP -lgmp
4761 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4762 testing availability of engines with "-t" - the old behaviour is
4763 produced by increasing the feature's verbosity with "-tt".
4766 *) ECDSA routines: under certain error conditions uninitialized BN objects
4767 could be freed. Solution: make sure initialization is performed early
4768 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4772 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4773 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4774 software implementations. For DSA and DH, parameter generation can
4775 also be overriden by providing the appropriate method callbacks.
4778 *) Change the "progress" mechanism used in key-generation and
4779 primality testing to functions that take a new BN_GENCB pointer in
4780 place of callback/argument pairs. The new API functions have "_ex"
4781 postfixes and the older functions are reimplemented as wrappers for
4782 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4783 declarations of the old functions to help (graceful) attempts to
4784 migrate to the new functions. Also, the new key-generation API
4785 functions operate on a caller-supplied key-structure and return
4786 success/failure rather than returning a key or NULL - this is to
4787 help make "keygen" another member function of RSA_METHOD etc.
4789 Example for using the new callback interface:
4791 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4795 BN_GENCB_set(&my_cb, my_callback, my_arg);
4797 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4798 /* For the meaning of a, b in calls to my_callback(), see the
4799 * documentation of the function that calls the callback.
4800 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4801 * my_callback should return 1 if it wants BN_is_prime_ex()
4802 * to continue, or 0 to stop.
4807 *) Change the ZLIB compression method to be stateful, and make it
4808 available to TLS with the number defined in
4809 draft-ietf-tls-compression-04.txt.
4812 *) Add the ASN.1 structures and functions for CertificatePair, which
4813 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4815 CertificatePair ::= SEQUENCE {
4816 forward [0] Certificate OPTIONAL,
4817 reverse [1] Certificate OPTIONAL,
4818 -- at least one of the pair shall be present -- }
4820 Also implement the PEM functions to read and write certificate
4821 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4823 This needed to be defined, mostly for the sake of the LDAP
4824 attribute crossCertificatePair, but may prove useful elsewhere as
4828 *) Make it possible to inhibit symlinking of shared libraries in
4829 Makefile.shared, for Cygwin's sake.
4832 *) Extend the BIGNUM API by creating a function
4833 void BN_set_negative(BIGNUM *a, int neg);
4834 and a macro that behave like
4835 int BN_is_negative(const BIGNUM *a);
4837 to avoid the need to access 'a->neg' directly in applications.
4840 *) Implement fast modular reduction for pseudo-Mersenne primes
4841 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4842 EC_GROUP_new_curve_GFp() will now automatically use this
4844 [Nils Larsch <nla@trustcenter.de>]
4846 *) Add new lock type (CRYPTO_LOCK_BN).
4849 *) Change the ENGINE framework to automatically load engines
4850 dynamically from specific directories unless they could be
4851 found to already be built in or loaded. Move all the
4852 current engines except for the cryptodev one to a new
4854 The engines in engines/ are built as shared libraries if
4855 the "shared" options was given to ./Configure or ./config.
4856 Otherwise, they are inserted in libcrypto.a.
4857 /usr/local/ssl/engines is the default directory for dynamic
4858 engines, but that can be overriden at configure time through
4859 the usual use of --prefix and/or --openssldir, and at run
4860 time with the environment variable OPENSSL_ENGINES.
4861 [Geoff Thorpe and Richard Levitte]
4863 *) Add Makefile.shared, a helper makefile to build shared
4864 libraries. Adapt Makefile.org.
4867 *) Add version info to Win32 DLLs.
4868 [Peter 'Luna' Runestig" <peter@runestig.com>]
4870 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4871 can be added using this API to created arbitrary PKCS#12
4872 files while avoiding the low level API.
4874 New options to PKCS12_create(), key or cert can be NULL and
4875 will then be omitted from the output file. The encryption
4876 algorithm NIDs can be set to -1 for no encryption, the mac
4877 iteration count can be set to 0 to omit the mac.
4879 Enhance pkcs12 utility by making the -nokeys and -nocerts
4880 options work when creating a PKCS#12 file. New option -nomac
4881 to omit the mac, NONE can be set for an encryption algorithm.
4882 New code is modified to use the enhanced PKCS12_create()
4883 instead of the low level API.
4886 *) Extend ASN1 encoder to support indefinite length constructed
4887 encoding. This can output sequences tags and octet strings in
4888 this form. Modify pk7_asn1.c to support indefinite length
4889 encoding. This is experimental and needs additional code to
4890 be useful, such as an ASN1 bio and some enhanced streaming
4893 Extend template encode functionality so that tagging is passed
4894 down to the template encoder.
4897 *) Let 'openssl req' fail if an argument to '-newkey' is not
4898 recognized instead of using RSA as a default.
4901 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4902 As these are not official, they are not included in "ALL";
4903 the "ECCdraft" ciphersuite group alias can be used to select them.
4904 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4906 *) Add ECDH engine support.
4907 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4909 *) Add ECDH in new directory crypto/ecdh/.
4910 [Douglas Stebila (Sun Microsystems Laboratories)]
4912 *) Let BN_rand_range() abort with an error after 100 iterations
4913 without success (which indicates a broken PRNG).
4916 *) Change BN_mod_sqrt() so that it verifies that the input value
4917 is really the square of the return value. (Previously,
4918 BN_mod_sqrt would show GIGO behaviour.)
4921 *) Add named elliptic curves over binary fields from X9.62, SECG,
4922 and WAP/WTLS; add OIDs that were still missing.
4924 [Sheueling Chang Shantz and Douglas Stebila
4925 (Sun Microsystems Laboratories)]
4927 *) Extend the EC library for elliptic curves over binary fields
4928 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4931 EC_GF2m_simple_method
4935 EC_GROUP_new_curve_GF2m
4936 EC_GROUP_set_curve_GF2m
4937 EC_GROUP_get_curve_GF2m
4938 EC_POINT_set_affine_coordinates_GF2m
4939 EC_POINT_get_affine_coordinates_GF2m
4940 EC_POINT_set_compressed_coordinates_GF2m
4942 Point compression for binary fields is disabled by default for
4943 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4946 As binary polynomials are represented as BIGNUMs, various members
4947 of the EC_GROUP and EC_POINT data structures can be shared
4948 between the implementations for prime fields and binary fields;
4949 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4950 are essentially identical to their ..._GFp counterparts.
4951 (For simplicity, the '..._GFp' prefix has been dropped from
4952 various internal method names.)
4954 An internal 'field_div' method (similar to 'field_mul' and
4955 'field_sqr') has been added; this is used only for binary fields.
4957 [Sheueling Chang Shantz and Douglas Stebila
4958 (Sun Microsystems Laboratories)]
4960 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4961 through methods ('mul', 'precompute_mult').
4963 The generic implementations (now internally called 'ec_wNAF_mul'
4964 and 'ec_wNAF_precomputed_mult') remain the default if these
4965 methods are undefined.
4967 [Sheueling Chang Shantz and Douglas Stebila
4968 (Sun Microsystems Laboratories)]
4970 *) New function EC_GROUP_get_degree, which is defined through
4971 EC_METHOD. For curves over prime fields, this returns the bit
4972 length of the modulus.
4974 [Sheueling Chang Shantz and Douglas Stebila
4975 (Sun Microsystems Laboratories)]
4977 *) New functions EC_GROUP_dup, EC_POINT_dup.
4978 (These simply call ..._new and ..._copy).
4980 [Sheueling Chang Shantz and Douglas Stebila
4981 (Sun Microsystems Laboratories)]
4983 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4984 Polynomials are represented as BIGNUMs (where the sign bit is not
4985 used) in the following functions [macros]:
4988 BN_GF2m_sub [= BN_GF2m_add]
4989 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4990 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4991 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4993 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4994 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4995 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4996 BN_GF2m_cmp [= BN_ucmp]
4998 (Note that only the 'mod' functions are actually for fields GF(2^m).
4999 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5001 For some functions, an the irreducible polynomial defining a
5002 field can be given as an 'unsigned int[]' with strictly
5003 decreasing elements giving the indices of those bits that are set;
5004 i.e., p[] represents the polynomial
5005 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5007 p[0] > p[1] > ... > p[k] = 0.
5008 This applies to the following functions:
5013 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5014 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5016 BN_GF2m_mod_sqrt_arr
5017 BN_GF2m_mod_solve_quad_arr
5021 Conversion can be performed by the following functions:
5026 bntest.c has additional tests for binary polynomial arithmetic.
5028 Two implementations for BN_GF2m_mod_div() are available.
5029 The default algorithm simply uses BN_GF2m_mod_inv() and
5030 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5031 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5032 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5034 [Sheueling Chang Shantz and Douglas Stebila
5035 (Sun Microsystems Laboratories)]
5037 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5038 functionality is disabled at compile-time.
5039 [Douglas Stebila <douglas.stebila@sun.com>]
5041 *) Change default behaviour of 'openssl asn1parse' so that more
5042 information is visible when viewing, e.g., a certificate:
5044 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5045 mode the content of non-printable OCTET STRINGs is output in a
5046 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5047 avoid the appearance of a printable string.
5048 [Nils Larsch <nla@trustcenter.de>]
5050 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5052 EC_GROUP_set_asn1_flag()
5053 EC_GROUP_get_asn1_flag()
5054 EC_GROUP_set_point_conversion_form()
5055 EC_GROUP_get_point_conversion_form()
5056 These control ASN1 encoding details:
5057 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5058 has been set to OPENSSL_EC_NAMED_CURVE.
5059 - Points are encoded in uncompressed form by default; options for
5060 asn1_for are as for point2oct, namely
5061 POINT_CONVERSION_COMPRESSED
5062 POINT_CONVERSION_UNCOMPRESSED
5063 POINT_CONVERSION_HYBRID
5065 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5068 EC_GROUP_get0_seed()
5069 EC_GROUP_get_seed_len()
5070 This is used only for ASN1 purposes (so far).
5071 [Nils Larsch <nla@trustcenter.de>]
5073 *) Add 'field_type' member to EC_METHOD, which holds the NID
5074 of the appropriate field type OID. The new function
5075 EC_METHOD_get_field_type() returns this value.
5076 [Nils Larsch <nla@trustcenter.de>]
5081 EC_POINT_point2hex()
5082 EC_POINT_hex2point()
5083 providing useful interfaces to EC_POINT_point2oct() and
5084 EC_POINT_oct2point().
5085 [Nils Larsch <nla@trustcenter.de>]
5087 *) Change internals of the EC library so that the functions
5088 EC_GROUP_set_generator()
5089 EC_GROUP_get_generator()
5090 EC_GROUP_get_order()
5091 EC_GROUP_get_cofactor()
5092 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5093 to methods, which would lead to unnecessary code duplication when
5094 adding different types of curves.
5095 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5097 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5098 arithmetic, and such that modified wNAFs are generated
5099 (which avoid length expansion in many cases).
5102 *) Add a function EC_GROUP_check_discriminant() (defined via
5103 EC_METHOD) that verifies that the curve discriminant is non-zero.
5105 Add a function EC_GROUP_check() that makes some sanity tests
5106 on a EC_GROUP, its generator and order. This includes
5107 EC_GROUP_check_discriminant().
5108 [Nils Larsch <nla@trustcenter.de>]
5110 *) Add ECDSA in new directory crypto/ecdsa/.
5112 Add applications 'openssl ecparam' and 'openssl ecdsa'
5113 (these are based on 'openssl dsaparam' and 'openssl dsa').
5115 ECDSA support is also included in various other files across the
5116 library. Most notably,
5117 - 'openssl req' now has a '-newkey ecdsa:file' option;
5118 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5119 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5120 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5121 them suitable for ECDSA where domain parameters must be
5122 extracted before the specific public key;
5123 - ECDSA engine support has been added.
5124 [Nils Larsch <nla@trustcenter.de>]
5126 *) Include some named elliptic curves, and add OIDs from X9.62,
5127 SECG, and WAP/WTLS. Each curve can be obtained from the new
5129 EC_GROUP_new_by_curve_name(),
5130 and the list of available named curves can be obtained with
5131 EC_get_builtin_curves().
5132 Also add a 'curve_name' member to EC_GROUP objects, which can be
5134 EC_GROUP_set_curve_name()
5135 EC_GROUP_get_curve_name()
5136 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5138 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5139 was actually never needed) and in BN_mul(). The removal in BN_mul()
5140 required a small change in bn_mul_part_recursive() and the addition
5141 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5142 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5143 bn_sub_words() and bn_add_words() except they take arrays with
5147 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5149 *) Cleanse PEM buffers before freeing them since they may contain
5151 [Benjamin Bennett <ben@psc.edu>]
5153 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5154 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5155 authentication-only ciphersuites.
5158 *) Since AES128 and AES256 share a single mask bit in the logic of
5159 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5160 kludge to work properly if AES128 is available and AES256 isn't.
5163 *) Expand security boundary to match 1.1.1 module.
5166 *) Remove redundant features: hash file source, editing of test vectors
5167 modify fipsld to use external fips_premain.c signature.
5170 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5171 run algorithm test programs.
5174 *) Make algorithm test programs more tolerant of whitespace.
5177 *) Have SSL/TLS server implementation tolerate "mismatched" record
5178 protocol version while receiving ClientHello even if the
5179 ClientHello is fragmented. (The server can't insist on the
5180 particular protocol version it has chosen before the ServerHello
5181 message has informed the client about his choice.)
5184 *) Load error codes if they are not already present instead of using a
5185 static variable. This allows them to be cleanly unloaded and reloaded.
5188 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5190 *) Introduce limits to prevent malicious keys being able to
5191 cause a denial of service. (CVE-2006-2940)
5192 [Steve Henson, Bodo Moeller]
5194 *) Fix ASN.1 parsing of certain invalid structures that can result
5195 in a denial of service. (CVE-2006-2937) [Steve Henson]
5197 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5198 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5200 *) Fix SSL client code which could crash if connecting to a
5201 malicious SSLv2 server. (CVE-2006-4343)
5202 [Tavis Ormandy and Will Drewry, Google Security Team]
5204 *) Change ciphersuite string processing so that an explicit
5205 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5206 will no longer include "AES128-SHA"), and any other similar
5207 ciphersuite (same bitmap) from *other* protocol versions (so that
5208 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5209 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5210 changes from 0.9.8b and 0.9.8d.
5213 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5215 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5216 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5218 *) Change the Unix randomness entropy gathering to use poll() when
5219 possible instead of select(), since the latter has some
5220 undesirable limitations.
5221 [Darryl Miles via Richard Levitte and Bodo Moeller]
5223 *) Disable rogue ciphersuites:
5225 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5226 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5227 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5229 The latter two were purportedly from
5230 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5233 Also deactive the remaining ciphersuites from
5234 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5235 unofficial, and the ID has long expired.
5238 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5239 dual-core machines) and other potential thread-safety issues.
5242 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5244 *) Adapt fipsld and the build system to link against the validated FIPS
5245 module in FIPS mode.
5248 *) Fixes for VC++ 2005 build under Windows.
5251 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5252 from a Windows bash shell such as MSYS. It is autodetected from the
5253 "config" script when run from a VC++ environment. Modify standard VC++
5254 build to use fipscanister.o from the GNU make build.
5257 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5259 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5260 The value now differs depending on if you build for FIPS or not.
5261 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5262 safely run with a non-FIPSed libcrypto, as it may crash because of
5263 the difference induced by this change.
5266 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5268 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5269 (part of SSL_OP_ALL). This option used to disable the
5270 countermeasure against man-in-the-middle protocol-version
5271 rollback in the SSL 2.0 server implementation, which is a bad
5272 idea. (CVE-2005-2969)
5274 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5275 for Information Security, National Institute of Advanced Industrial
5276 Science and Technology [AIST], Japan)]
5278 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5279 mainly for FIPS compliance and not fully integrated at this stage.
5282 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5283 the exponentiation using a fixed-length exponent. (Otherwise,
5284 the information leaked through timing could expose the secret key
5285 after many signatures; cf. Bleichenbacher's attack on DSA with
5289 *) Make a new fixed-window mod_exp implementation the default for
5290 RSA, DSA, and DH private-key operations so that the sequence of
5291 squares and multiplies and the memory access pattern are
5292 independent of the particular secret key. This will mitigate
5293 cache-timing and potential related attacks.
5295 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5296 and this is automatically used by BN_mod_exp_mont() if the new flag
5297 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5298 will use this BN flag for private exponents unless the flag
5299 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5300 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5302 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5304 *) Change the client implementation for SSLv23_method() and
5305 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5306 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5307 (Previously, the SSL 2.0 backwards compatible Client Hello
5308 message format would be used even with SSL_OP_NO_SSLv2.)
5311 *) Add support for smime-type MIME parameter in S/MIME messages which some
5315 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5316 a threadsafe manner. Modify rsa code to use new function and add calls
5317 to dsa and dh code (which had race conditions before).
5320 *) Include the fixed error library code in the C error file definitions
5321 instead of fixing them up at runtime. This keeps the error code
5322 structures constant.
5325 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5327 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5330 *) Fixes for newer kerberos headers. NB: the casts are needed because
5331 the 'length' field is signed on one version and unsigned on another
5332 with no (?) obvious way to tell the difference, without these VC++
5333 complains. Also the "definition" of FAR (blank) is no longer included
5334 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5335 some needed definitions.
5338 *) Undo Cygwin change.
5341 *) Added support for proxy certificates according to RFC 3820.
5342 Because they may be a security thread to unaware applications,
5343 they must be explicitly allowed in run-time. See
5344 docs/HOWTO/proxy_certificates.txt for further information.
5347 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5349 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5350 server and client random values. Previously
5351 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5352 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5354 This change has negligible security impact because:
5356 1. Server and client random values still have 24 bytes of pseudo random
5359 2. Server and client random values are sent in the clear in the initial
5362 3. The master secret is derived using the premaster secret (48 bytes in
5363 size for static RSA ciphersuites) as well as client server and random
5366 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5369 [Stephen Henson, reported by UK NISCC]
5371 *) Use Windows randomness collection on Cygwin.
5374 *) Fix hang in EGD/PRNGD query when communication socket is closed
5375 prematurely by EGD/PRNGD.
5376 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5378 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5381 *) Back-port of selected performance improvements from development
5382 branch, as well as improved support for PowerPC platforms.
5385 *) Add lots of checks for memory allocation failure, error codes to indicate
5386 failure and freeing up memory if a failure occurs.
5387 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5389 *) Add new -passin argument to dgst.
5392 *) Perform some character comparisons of different types in X509_NAME_cmp:
5393 this is needed for some certificates that reencode DNs into UTF8Strings
5394 (in violation of RFC3280) and can't or wont issue name rollover
5398 *) Make an explicit check during certificate validation to see that
5399 the CA setting in each certificate on the chain is correct. As a
5400 side effect always do the following basic checks on extensions,
5401 not just when there's an associated purpose to the check:
5403 - if there is an unhandled critical extension (unless the user
5404 has chosen to ignore this fault)
5405 - if the path length has been exceeded (if one is set at all)
5406 - that certain extensions fit the associated purpose (if one has
5410 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5412 *) Avoid a race condition when CRLs are checked in a multi threaded
5413 environment. This would happen due to the reordering of the revoked
5414 entries during signature checking and serial number lookup. Now the
5415 encoding is cached and the serial number sort performed under a lock.
5416 Add new STACK function sk_is_sorted().
5419 *) Add Delta CRL to the extension code.
5422 *) Various fixes to s3_pkt.c so alerts are sent properly.
5423 [David Holmes <d.holmes@f5.com>]
5425 *) Reduce the chances of duplicate issuer name and serial numbers (in
5426 violation of RFC3280) using the OpenSSL certificate creation utilities.
5427 This is done by creating a random 64 bit value for the initial serial
5428 number when a serial number file is created or when a self signed
5429 certificate is created using 'openssl req -x509'. The initial serial
5430 number file is created using 'openssl x509 -next_serial' in CA.pl
5431 rather than being initialized to 1.
5434 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5436 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5437 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5438 [Joe Orton, Steve Henson]
5440 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5442 [Joe Orton, Steve Henson]
5444 *) Make it possible to have multiple active certificates with the same
5445 subject in the CA index file. This is done only if the keyword
5446 'unique_subject' is set to 'no' in the main CA section (default
5447 if 'CA_default') of the configuration file. The value is saved
5448 with the database itself in a separate index attribute file,
5449 named like the index file with '.attr' appended to the name.
5452 *) X509 verify fixes. Disable broken certificate workarounds when
5453 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5454 keyUsage extension present. Don't accept CRLs with unhandled critical
5455 extensions: since verify currently doesn't process CRL extensions this
5456 rejects a CRL with *any* critical extensions. Add new verify error codes
5460 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5461 A clarification of RFC2560 will require the use of OCTET STRINGs and
5462 some implementations cannot handle the current raw format. Since OpenSSL
5463 copies and compares OCSP nonces as opaque blobs without any attempt at
5464 parsing them this should not create any compatibility issues.
5467 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5468 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5469 this HMAC (and other) operations are several times slower than OpenSSL
5473 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5474 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5476 *) Use the correct content when signing type "other".
5479 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5481 *) Fix various bugs revealed by running the NISCC test suite:
5483 Stop out of bounds reads in the ASN1 code when presented with
5484 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5486 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5488 If verify callback ignores invalid public key errors don't try to check
5489 certificate signature with the NULL public key.
5493 *) New -ignore_err option in ocsp application to stop the server
5494 exiting on the first error in a request.
5497 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5498 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5502 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5503 extra data after the compression methods not only for TLS 1.0
5504 but also for SSL 3.0 (as required by the specification).
5505 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5507 *) Change X509_certificate_type() to mark the key as exported/exportable
5508 when it's 512 *bits* long, not 512 bytes.
5511 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5512 blocks during encryption.
5515 *) Various fixes to base64 BIO and non blocking I/O. On write
5516 flushes were not handled properly if the BIO retried. On read
5517 data was not being buffered properly and had various logic bugs.
5518 This also affects blocking I/O when the data being decoded is a
5522 *) Various S/MIME bugfixes and compatibility changes:
5523 output correct application/pkcs7 MIME type if
5524 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5525 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5526 of files as .eml work). Correctly handle very long lines in MIME
5530 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5532 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5533 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5534 a protocol version number mismatch like a decryption error
5535 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5538 *) Turn on RSA blinding by default in the default implementation
5539 to avoid a timing attack. Applications that don't want it can call
5540 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5541 They would be ill-advised to do so in most cases.
5542 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5544 *) Change RSA blinding code so that it works when the PRNG is not
5545 seeded (in this case, the secret RSA exponent is abused as
5546 an unpredictable seed -- if it is not unpredictable, there
5547 is no point in blinding anyway). Make RSA blinding thread-safe
5548 by remembering the creator's thread ID in rsa->blinding and
5549 having all other threads use local one-time blinding factors
5550 (this requires more computation than sharing rsa->blinding, but
5551 avoids excessive locking; and if an RSA object is not shared
5552 between threads, blinding will still be very fast).
5555 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5556 ENGINE as defaults for all supported algorithms irrespective of
5557 the 'flags' parameter. 'flags' is now honoured, so applications
5558 should make sure they are passing it correctly.
5561 *) Target "mingw" now allows native Windows code to be generated in
5562 the Cygwin environment as well as with the MinGW compiler.
5565 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5567 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5568 via timing by performing a MAC computation even if incorrect
5569 block cipher padding has been found. This is a countermeasure
5570 against active attacks where the attacker has to distinguish
5571 between bad padding and a MAC verification error. (CVE-2003-0078)
5573 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5574 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5575 Martin Vuagnoux (EPFL, Ilion)]
5577 *) Make the no-err option work as intended. The intention with no-err
5578 is not to have the whole error stack handling routines removed from
5579 libcrypto, it's only intended to remove all the function name and
5580 reason texts, thereby removing some of the footprint that may not
5581 be interesting if those errors aren't displayed anyway.
5583 NOTE: it's still possible for any application or module to have it's
5584 own set of error texts inserted. The routines are there, just not
5585 used by default when no-err is given.
5588 *) Add support for FreeBSD on IA64.
5589 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5591 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5592 Kerberos function mit_des_cbc_cksum(). Before this change,
5593 the value returned by DES_cbc_cksum() was like the one from
5594 mit_des_cbc_cksum(), except the bytes were swapped.
5595 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5597 *) Allow an application to disable the automatic SSL chain building.
5598 Before this a rather primitive chain build was always performed in
5599 ssl3_output_cert_chain(): an application had no way to send the
5600 correct chain if the automatic operation produced an incorrect result.
5602 Now the chain builder is disabled if either:
5604 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5606 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5608 The reasoning behind this is that an application would not want the
5609 auto chain building to take place if extra chain certificates are
5610 present and it might also want a means of sending no additional
5611 certificates (for example the chain has two certificates and the
5615 *) Add the possibility to build without the ENGINE framework.
5616 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5618 *) Under Win32 gmtime() can return NULL: check return value in
5619 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5622 *) DSA routines: under certain error conditions uninitialized BN objects
5623 could be freed. Solution: make sure initialization is performed early
5624 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5625 Nils Larsch <nla@trustcenter.de> via PR#459)
5628 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5629 checked on reconnect on the client side, therefore session resumption
5630 could still fail with a "ssl session id is different" error. This
5631 behaviour is masked when SSL_OP_ALL is used due to
5632 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5633 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5634 followup to PR #377.
5637 *) IA-32 assembler support enhancements: unified ELF targets, support
5638 for SCO/Caldera platforms, fix for Cygwin shared build.
5641 *) Add support for FreeBSD on sparc64. As a consequence, support for
5642 FreeBSD on non-x86 processors is separate from x86 processors on
5643 the config script, much like the NetBSD support.
5644 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5646 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5648 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5651 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5652 code (06) was taken as the first octet of the session ID and the last
5653 octet was ignored consequently. As a result SSLv2 client side session
5654 caching could not have worked due to the session ID mismatch between
5656 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5660 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5661 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5665 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5666 seems that in spite of existing for more than a year, many application
5667 author have done nothing to provide the necessary callbacks, which
5668 means that this particular engine will not work properly anywhere.
5669 This is a very unfortunate situation which forces us, in the name
5670 of usability, to give the hw_ncipher.c a static lock, which is part
5672 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5673 appear in 0.9.8 or later. We EXPECT application authors to have
5674 dealt properly with this when 0.9.8 is released (unless we actually
5675 make such changes in the libcrypto locking code that changes will
5676 have to be made anyway).
5679 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5680 octets have been read, EOF or an error occurs. Without this change
5681 some truncated ASN1 structures will not produce an error.
5684 *) Disable Heimdal support, since it hasn't been fully implemented.
5685 Still give the possibility to force the use of Heimdal, but with
5686 warnings and a request that patches get sent to openssl-dev.
5689 *) Add the VC-CE target, introduce the WINCE sysname, and add
5690 INSTALL.WCE and appropriate conditionals to make it build.
5691 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5693 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5694 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5695 edit numbers of the version.
5696 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5698 *) Introduce safe string copy and catenation functions
5699 (BUF_strlcpy() and BUF_strlcat()).
5700 [Ben Laurie (CHATS) and Richard Levitte]
5702 *) Avoid using fixed-size buffers for one-line DNs.
5703 [Ben Laurie (CHATS)]
5705 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5706 resizing buffers containing secrets, and use where appropriate.
5707 [Ben Laurie (CHATS)]
5709 *) Avoid using fixed size buffers for configuration file location.
5710 [Ben Laurie (CHATS)]
5712 *) Avoid filename truncation for various CA files.
5713 [Ben Laurie (CHATS)]
5715 *) Use sizeof in preference to magic numbers.
5716 [Ben Laurie (CHATS)]
5718 *) Avoid filename truncation in cert requests.
5719 [Ben Laurie (CHATS)]
5721 *) Add assertions to check for (supposedly impossible) buffer
5723 [Ben Laurie (CHATS)]
5725 *) Don't cache truncated DNS entries in the local cache (this could
5726 potentially lead to a spoofing attack).
5727 [Ben Laurie (CHATS)]
5729 *) Fix various buffers to be large enough for hex/decimal
5730 representations in a platform independent manner.
5731 [Ben Laurie (CHATS)]
5733 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5734 resizing buffers containing secrets, and use where appropriate.
5735 [Ben Laurie (CHATS)]
5737 *) Add BIO_indent() to avoid much slightly worrying code to do
5739 [Ben Laurie (CHATS)]
5741 *) Convert sprintf()/BIO_puts() to BIO_printf().
5742 [Ben Laurie (CHATS)]
5744 *) buffer_gets() could terminate with the buffer only half
5746 [Ben Laurie (CHATS)]
5748 *) Add assertions to prevent user-supplied crypto functions from
5749 overflowing internal buffers by having large block sizes, etc.
5750 [Ben Laurie (CHATS)]
5752 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5754 [Ben Laurie (CHATS)]
5756 *) Eliminate unused copy of key in RC4.
5757 [Ben Laurie (CHATS)]
5759 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5760 [Ben Laurie (CHATS)]
5762 *) Fix off-by-one error in EGD path.
5763 [Ben Laurie (CHATS)]
5765 *) If RANDFILE path is too long, ignore instead of truncating.
5766 [Ben Laurie (CHATS)]
5768 *) Eliminate unused and incorrectly sized X.509 structure
5770 [Ben Laurie (CHATS)]
5772 *) Eliminate unused and dangerous function knumber().
5773 [Ben Laurie (CHATS)]
5775 *) Eliminate unused and dangerous structure, KSSL_ERR.
5776 [Ben Laurie (CHATS)]
5778 *) Protect against overlong session ID context length in an encoded
5779 session object. Since these are local, this does not appear to be
5781 [Ben Laurie (CHATS)]
5783 *) Change from security patch (see 0.9.6e below) that did not affect
5784 the 0.9.6 release series:
5786 Remote buffer overflow in SSL3 protocol - an attacker could
5787 supply an oversized master key in Kerberos-enabled versions.
5789 [Ben Laurie (CHATS)]
5791 *) Change the SSL kerb5 codes to match RFC 2712.
5794 *) Make -nameopt work fully for req and add -reqopt switch.
5795 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5797 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5798 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5800 *) Make sure tests can be performed even if the corresponding algorithms
5801 have been removed entirely. This was also the last step to make
5802 OpenSSL compilable with DJGPP under all reasonable conditions.
5803 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5805 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5806 to allow version independent disabling of normally unselected ciphers,
5807 which may be activated as a side-effect of selecting a single cipher.
5809 (E.g., cipher list string "RSA" enables ciphersuites that are left
5810 out of "ALL" because they do not provide symmetric encryption.
5811 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5812 [Lutz Jaenicke, Bodo Moeller]
5814 *) Add appropriate support for separate platform-dependent build
5815 directories. The recommended way to make a platform-dependent
5816 build directory is the following (tested on Linux), maybe with
5819 # Place yourself outside of the OpenSSL source tree. In
5820 # this example, the environment variable OPENSSL_SOURCE
5821 # is assumed to contain the absolute OpenSSL source directory.
5822 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5823 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5824 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5825 mkdir -p `dirname $F`
5826 ln -s $OPENSSL_SOURCE/$F $F
5829 To be absolutely sure not to disturb the source tree, a "make clean"
5830 is a good thing. If it isn't successful, don't worry about it,
5831 it probably means the source directory is very clean.
5834 *) Make sure any ENGINE control commands make local copies of string
5835 pointers passed to them whenever necessary. Otherwise it is possible
5836 the caller may have overwritten (or deallocated) the original string
5837 data when a later ENGINE operation tries to use the stored values.
5838 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5840 *) Improve diagnostics in file reading and command-line digests.
5841 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5843 *) Add AES modes CFB and OFB to the object database. Correct an
5844 error in AES-CFB decryption.
5847 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5848 allows existing EVP_CIPHER_CTX structures to be reused after
5849 calling EVP_*Final(). This behaviour is used by encryption
5850 BIOs and some applications. This has the side effect that
5851 applications must explicitly clean up cipher contexts with
5852 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5855 *) Check the values of dna and dnb in bn_mul_recursive before calling
5856 bn_mul_comba (a non zero value means the a or b arrays do not contain
5857 n2 elements) and fallback to bn_mul_normal if either is not zero.
5860 *) Fix escaping of non-ASCII characters when using the -subj option
5861 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5864 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5865 form for "surname", serialNumber has no short form.
5866 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5867 therefore remove "mail" short name for "internet 7".
5868 The OID for unique identifiers in X509 certificates is
5869 x500UniqueIdentifier, not uniqueIdentifier.
5870 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5873 *) Add an "init" command to the ENGINE config module and auto initialize
5874 ENGINEs. Without any "init" command the ENGINE will be initialized
5875 after all ctrl commands have been executed on it. If init=1 the
5876 ENGINE is initailized at that point (ctrls before that point are run
5877 on the uninitialized ENGINE and after on the initialized one). If
5878 init=0 then the ENGINE will not be iniatialized at all.
5881 *) Fix the 'app_verify_callback' interface so that the user-defined
5882 argument is actually passed to the callback: In the
5883 SSL_CTX_set_cert_verify_callback() prototype, the callback
5884 declaration has been changed from
5887 int (*cb)(X509_STORE_CTX *,void *);
5888 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5889 i=s->ctx->app_verify_callback(&ctx)
5890 has been changed into
5891 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5893 To update applications using SSL_CTX_set_cert_verify_callback(),
5894 a dummy argument can be added to their callback functions.
5895 [D. K. Smetters <smetters@parc.xerox.com>]
5897 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5898 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5900 *) Add and OPENSSL_LOAD_CONF define which will cause
5901 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5902 This allows older applications to transparently support certain
5903 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5904 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5905 load the config file and OPENSSL_add_all_algorithms_conf() which will
5906 always load it have also been added.
5909 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5910 Adjust NIDs and EVP layer.
5911 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5913 *) Config modules support in openssl utility.
5915 Most commands now load modules from the config file,
5916 though in a few (such as version) this isn't done
5917 because it couldn't be used for anything.
5919 In the case of ca and req the config file used is
5920 the same as the utility itself: that is the -config
5921 command line option can be used to specify an
5925 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5926 use "openssl_conf" if filename is NULL use default openssl config file.
5929 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5930 config section name. Add a new flag to tolerate a missing config file
5931 and move code to CONF_modules_load_file().
5934 *) Support for crypto accelerator cards from Accelerated Encryption
5935 Processing, www.aep.ie. (Use engine 'aep')
5936 The support was copied from 0.9.6c [engine] and adapted/corrected
5937 to work with the new engine framework.
5938 [AEP Inc. and Richard Levitte]
5940 *) Support for SureWare crypto accelerator cards from Baltimore
5941 Technologies. (Use engine 'sureware')
5942 The support was copied from 0.9.6c [engine] and adapted
5943 to work with the new engine framework.
5946 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5947 make the newer ENGINE framework commands for the CHIL engine work.
5948 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5950 *) Make it possible to produce shared libraries on ReliantUNIX.
5951 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5953 *) Add the configuration target debug-linux-ppro.
5954 Make 'openssl rsa' use the general key loading routines
5955 implemented in apps.c, and make those routines able to
5956 handle the key format FORMAT_NETSCAPE and the variant
5958 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5960 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5961 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5963 *) Add -keyform to rsautl, and document -engine.
5964 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5966 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5967 BIO_R_NO_SUCH_FILE error code rather than the generic
5968 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5971 *) Add new functions
5973 ERR_peek_last_error_line
5974 ERR_peek_last_error_line_data.
5975 These are similar to
5978 ERR_peek_error_line_data,
5979 but report on the latest error recorded rather than the first one
5980 still in the error queue.
5981 [Ben Laurie, Bodo Moeller]
5983 *) default_algorithms option in ENGINE config module. This allows things
5985 default_algorithms = ALL
5986 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5989 *) Preliminary ENGINE config module.
5992 *) New experimental application configuration code.
5995 *) Change the AES code to follow the same name structure as all other
5996 symmetric ciphers, and behave the same way. Move everything to
5997 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5998 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6000 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6001 [Ben Laurie and Theo de Raadt]
6003 *) Add option to output public keys in req command.
6004 [Massimiliano Pala madwolf@openca.org]
6006 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6007 (up to about 10% better than before for P-192 and P-224).
6010 *) New functions/macros
6012 SSL_CTX_set_msg_callback(ctx, cb)
6013 SSL_CTX_set_msg_callback_arg(ctx, arg)
6014 SSL_set_msg_callback(ssl, cb)
6015 SSL_set_msg_callback_arg(ssl, arg)
6017 to request calling a callback function
6019 void cb(int write_p, int version, int content_type,
6020 const void *buf, size_t len, SSL *ssl, void *arg)
6022 whenever a protocol message has been completely received
6023 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6024 protocol version according to which the SSL library interprets
6025 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6026 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6027 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6028 specification (change_cipher_spec(20), alert(21), handshake(22)).
6029 'buf' and 'len' point to the actual message, 'ssl' to the
6030 SSL object, and 'arg' is the application-defined value set by
6031 SSL[_CTX]_set_msg_callback_arg().
6033 'openssl s_client' and 'openssl s_server' have new '-msg' options
6034 to enable a callback that displays all protocol messages.
6037 *) Change the shared library support so shared libraries are built as
6038 soon as the corresponding static library is finished, and thereby get
6039 openssl and the test programs linked against the shared library.
6040 This still only happens when the keyword "shard" has been given to
6041 the configuration scripts.
6043 NOTE: shared library support is still an experimental thing, and
6044 backward binary compatibility is still not guaranteed.
6045 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6047 *) Add support for Subject Information Access extension.
6048 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6050 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6051 additional bytes when new memory had to be allocated, not just
6052 when reusing an existing buffer.
6055 *) New command line and configuration option 'utf8' for the req command.
6056 This allows field values to be specified as UTF8 strings.
6059 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6060 runs for the former and machine-readable output for the latter.
6063 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6064 of the e-mail address in the DN (i.e., it will go into a certificate
6065 extension only). The new configuration file option 'email_in_dn = no'
6066 has the same effect.
6067 [Massimiliano Pala madwolf@openca.org]
6069 *) Change all functions with names starting with des_ to be starting
6070 with DES_ instead. Add wrappers that are compatible with libdes,
6071 but are named _ossl_old_des_*. Finally, add macros that map the
6072 des_* symbols to the corresponding _ossl_old_des_* if libdes
6073 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6074 desired, the des_* symbols will be mapped to DES_*, with one
6077 Since we provide two compatibility mappings, the user needs to
6078 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6079 compatibility is desired. The default (i.e., when that macro
6080 isn't defined) is OpenSSL 0.9.6c compatibility.
6082 There are also macros that enable and disable the support of old
6083 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6084 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6085 are defined, the default will apply: to support the old des routines.
6087 In either case, one must include openssl/des.h to get the correct
6088 definitions. Do not try to just include openssl/des_old.h, that
6091 NOTE: This is a major break of an old API into a new one. Software
6092 authors are encouraged to switch to the DES_ style functions. Some
6093 time in the future, des_old.h and the libdes compatibility functions
6094 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6095 default), and then completely removed.
6098 *) Test for certificates which contain unsupported critical extensions.
6099 If such a certificate is found during a verify operation it is
6100 rejected by default: this behaviour can be overridden by either
6101 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6102 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6103 X509_supported_extension() has also been added which returns 1 if a
6104 particular extension is supported.
6107 *) Modify the behaviour of EVP cipher functions in similar way to digests
6108 to retain compatibility with existing code.
6111 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6112 compatibility with existing code. In particular the 'ctx' parameter does
6113 not have to be to be initialized before the call to EVP_DigestInit() and
6114 it is tidied up after a call to EVP_DigestFinal(). New function
6115 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6116 EVP_MD_CTX_copy() changed to not require the destination to be
6117 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6118 requires the destination to be valid.
6120 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6121 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6124 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6125 so that complete 'Handshake' protocol structures are kept in memory
6126 instead of overwriting 'msg_type' and 'length' with 'body' data.
6129 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6130 [Massimo Santin via Richard Levitte]
6132 *) Major restructuring to the underlying ENGINE code. This includes
6133 reduction of linker bloat, separation of pure "ENGINE" manipulation
6134 (initialisation, etc) from functionality dealing with implementations
6135 of specific crypto iterfaces. This change also introduces integrated
6136 support for symmetric ciphers and digest implementations - so ENGINEs
6137 can now accelerate these by providing EVP_CIPHER and EVP_MD
6138 implementations of their own. This is detailed in crypto/engine/README
6139 as it couldn't be adequately described here. However, there are a few
6140 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6141 were changed in the original introduction of ENGINE code have now
6142 reverted back - the hooking from this code to ENGINE is now a good
6143 deal more passive and at run-time, operations deal directly with
6144 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6145 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6146 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6147 they were not being used by the framework as there is no concept of a
6148 BIGNUM_METHOD and they could not be generalised to the new
6149 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6150 ENGINE_cpy() has been removed as it cannot be consistently defined in
6154 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6157 *) Change mkdef.pl to sort symbols that get the same entry number,
6158 and make sure the automatically generated functions ERR_load_*
6159 become part of libeay.num as well.
6162 *) New function SSL_renegotiate_pending(). This returns true once
6163 renegotiation has been requested (either SSL_renegotiate() call
6164 or HelloRequest/ClientHello received from the peer) and becomes
6165 false once a handshake has been completed.
6166 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6167 sends a HelloRequest, but does not ensure that a handshake takes
6168 place. SSL_renegotiate_pending() is useful for checking if the
6169 client has followed the request.)
6172 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6173 By default, clients may request session resumption even during
6174 renegotiation (if session ID contexts permit); with this option,
6175 session resumption is possible only in the first handshake.
6177 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6178 more bits available for options that should not be part of
6179 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6182 *) Add some demos for certificate and certificate request creation.
6185 *) Make maximum certificate chain size accepted from the peer application
6186 settable (SSL*_get/set_max_cert_list()), as proposed by
6187 "Douglas E. Engert" <deengert@anl.gov>.
6190 *) Add support for shared libraries for Unixware-7
6191 (Boyd Lynn Gerber <gerberb@zenez.com>).
6194 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6195 be done prior to destruction. Use this to unload error strings from
6196 ENGINEs that load their own error strings. NB: This adds two new API
6197 functions to "get" and "set" this destroy handler in an ENGINE.
6200 *) Alter all existing ENGINE implementations (except "openssl" and
6201 "openbsd") to dynamically instantiate their own error strings. This
6202 makes them more flexible to be built both as statically-linked ENGINEs
6203 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6204 Also, add stub code to each that makes building them as self-contained
6205 shared-libraries easier (see README.ENGINE).
6208 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6209 implementations into applications that are completely implemented in
6210 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6211 commands that can be used to configure what shared-library to load and
6212 to control aspects of the way it is handled. Also, made an update to
6213 the README.ENGINE file that brings its information up-to-date and
6214 provides some information and instructions on the "dynamic" ENGINE
6215 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6218 *) Make it possible to unload ranges of ERR strings with a new
6219 "ERR_unload_strings" function.
6222 *) Add a copy() function to EVP_MD.
6225 *) Make EVP_MD routines take a context pointer instead of just the
6226 md_data void pointer.
6229 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6230 that the digest can only process a single chunk of data
6231 (typically because it is provided by a piece of
6232 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6233 is only going to provide a single chunk of data, and hence the
6234 framework needn't accumulate the data for oneshot drivers.
6237 *) As with "ERR", make it possible to replace the underlying "ex_data"
6238 functions. This change also alters the storage and management of global
6239 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6240 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6241 index counters. The API functions that use this state have been changed
6242 to take a "class_index" rather than pointers to the class's local STACK
6243 and counter, and there is now an API function to dynamically create new
6244 classes. This centralisation allows us to (a) plug a lot of the
6245 thread-safety problems that existed, and (b) makes it possible to clean
6246 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6247 such data would previously have always leaked in application code and
6248 workarounds were in place to make the memory debugging turn a blind eye
6249 to it. Application code that doesn't use this new function will still
6250 leak as before, but their memory debugging output will announce it now
6251 rather than letting it slide.
6253 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6254 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6255 has a return value to indicate success or failure.
6258 *) Make it possible to replace the underlying "ERR" functions such that the
6259 global state (2 LHASH tables and 2 locks) is only used by the "default"
6260 implementation. This change also adds two functions to "get" and "set"
6261 the implementation prior to it being automatically set the first time
6262 any other ERR function takes place. Ie. an application can call "get",
6263 pass the return value to a module it has just loaded, and that module
6264 can call its own "set" function using that value. This means the
6265 module's "ERR" operations will use (and modify) the error state in the
6266 application and not in its own statically linked copy of OpenSSL code.
6269 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6270 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6271 the operation, and provides a more encapsulated way for external code
6272 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6273 to use these functions rather than manually incrementing the counts.
6275 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6278 *) Add EVP test program.
6281 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6284 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6285 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6286 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6287 These allow a CRL to be built without having to access X509_CRL fields
6288 directly. Modify 'ca' application to use new functions.
6291 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6292 bug workarounds. Rollback attack detection is a security feature.
6293 The problem will only arise on OpenSSL servers when TLSv1 is not
6294 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6295 Software authors not wanting to support TLSv1 will have special reasons
6296 for their choice and can explicitly enable this option.
6297 [Bodo Moeller, Lutz Jaenicke]
6299 *) Rationalise EVP so it can be extended: don't include a union of
6300 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6301 (similar to those existing for EVP_CIPHER_CTX).
6306 EVP_MD_CTX_init(&md); /* new function call */
6307 EVP_DigestInit(&md, EVP_sha1());
6308 EVP_DigestUpdate(&md, in, len);
6309 EVP_DigestFinal(&md, out, NULL);
6310 EVP_MD_CTX_cleanup(&md); /* new function call */
6314 *) Make DES key schedule conform to the usual scheme, as well as
6315 correcting its structure. This means that calls to DES functions
6316 now have to pass a pointer to a des_key_schedule instead of a
6317 plain des_key_schedule (which was actually always a pointer
6320 des_key_schedule ks;
6322 des_set_key_checked(..., &ks);
6323 des_ncbc_encrypt(..., &ks, ...);
6325 (Note that a later change renames 'des_...' into 'DES_...'.)
6328 *) Initial reduction of linker bloat: the use of some functions, such as
6329 PEM causes large amounts of unused functions to be linked in due to
6330 poor organisation. For example pem_all.c contains every PEM function
6331 which has a knock on effect of linking in large amounts of (unused)
6332 ASN1 code. Grouping together similar functions and splitting unrelated
6333 functions prevents this.
6336 *) Cleanup of EVP macros.
6339 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6340 correct _ecb suffix.
6343 *) Add initial OCSP responder support to ocsp application. The
6344 revocation information is handled using the text based index
6345 use by the ca application. The responder can either handle
6346 requests generated internally, supplied in files (for example
6347 via a CGI script) or using an internal minimal server.
6350 *) Add configuration choices to get zlib compression for TLS.
6353 *) Changes to Kerberos SSL for RFC 2712 compliance:
6354 1. Implemented real KerberosWrapper, instead of just using
6355 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6356 2. Implemented optional authenticator field of KerberosWrapper.
6358 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6359 and authenticator structs; see crypto/krb5/.
6361 Generalized Kerberos calls to support multiple Kerberos libraries.
6362 [Vern Staats <staatsvr@asc.hpc.mil>,
6363 Jeffrey Altman <jaltman@columbia.edu>
6364 via Richard Levitte]
6366 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6367 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6368 values for each of the key sizes rather than having just
6369 parameters (and 'speed' generating keys each time).
6372 *) Speed up EVP routines.
6375 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6376 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6377 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6378 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6380 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6381 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6382 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6385 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6387 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6390 *) Added the OS2-EMX target.
6391 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6393 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6394 to support NCONF routines in extension code. New function CONF_set_nconf()
6395 to allow functions which take an NCONF to also handle the old LHASH
6396 structure: this means that the old CONF compatible routines can be
6397 retained (in particular wrt extensions) without having to duplicate the
6398 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6401 *) Enhance the general user interface with mechanisms for inner control
6402 and with possibilities to have yes/no kind of prompts.
6405 *) Change all calls to low level digest routines in the library and
6406 applications to use EVP. Add missing calls to HMAC_cleanup() and
6407 don't assume HMAC_CTX can be copied using memcpy().
6408 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6410 *) Add the possibility to control engines through control names but with
6411 arbitrary arguments instead of just a string.
6412 Change the key loaders to take a UI_METHOD instead of a callback
6413 function pointer. NOTE: this breaks binary compatibility with earlier
6414 versions of OpenSSL [engine].
6415 Adapt the nCipher code for these new conditions and add a card insertion
6419 *) Enhance the general user interface with mechanisms to better support
6420 dialog box interfaces, application-defined prompts, the possibility
6421 to use defaults (for example default passwords from somewhere else)
6422 and interrupts/cancellations.
6425 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6426 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6429 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6430 tidy up some unnecessarily weird code in 'sk_new()').
6431 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6433 *) Change the key loading routines for ENGINEs to use the same kind
6434 callback (pem_password_cb) as all other routines that need this
6438 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6439 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6440 than this minimum value is recommended.
6443 *) New random seeder for OpenVMS, using the system process statistics
6444 that are easily reachable.
6447 *) Windows apparently can't transparently handle global
6448 variables defined in DLLs. Initialisations such as:
6450 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6452 wont compile. This is used by the any applications that need to
6453 declare their own ASN1 modules. This was fixed by adding the option
6454 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6455 needed for static libraries under Win32.
6458 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6459 setting of purpose and trust fields. New X509_STORE trust and
6460 purpose functions and tidy up setting in other SSL functions.
6463 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6464 structure. These are inherited by X509_STORE_CTX when it is
6465 initialised. This allows various defaults to be set in the
6466 X509_STORE structure (such as flags for CRL checking and custom
6467 purpose or trust settings) for functions which only use X509_STORE_CTX
6468 internally such as S/MIME.
6470 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6471 trust settings if they are not set in X509_STORE. This allows X509_STORE
6472 purposes and trust (in S/MIME for example) to override any set by default.
6474 Add command line options for CRL checking to smime, s_client and s_server
6478 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6479 are set then the CRL is looked up in the X509_STORE structure and
6480 its validity and signature checked, then if the certificate is found
6481 in the CRL the verify fails with a revoked error.
6483 Various new CRL related callbacks added to X509_STORE_CTX structure.
6485 Command line options added to 'verify' application to support this.
6487 This needs some additional work, such as being able to handle multiple
6488 CRLs with different times, extension based lookup (rather than just
6489 by subject name) and ultimately more complete V2 CRL extension
6493 *) Add a general user interface API (crypto/ui/). This is designed
6494 to replace things like des_read_password and friends (backward
6495 compatibility functions using this new API are provided).
6496 The purpose is to remove prompting functions from the DES code
6497 section as well as provide for prompting through dialog boxes in
6498 a window system and the like.
6501 *) Add "ex_data" support to ENGINE so implementations can add state at a
6502 per-structure level rather than having to store it globally.
6505 *) Make it possible for ENGINE structures to be copied when retrieved by
6506 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6507 This causes the "original" ENGINE structure to act like a template,
6508 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6509 operational state can be localised to each ENGINE structure, despite the
6510 fact they all share the same "methods". New ENGINE structures returned in
6511 this case have no functional references and the return value is the single
6512 structural reference. This matches the single structural reference returned
6513 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6517 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6518 needs to match any other type at all we need to manually clear the
6522 *) Changes to the "openssl engine" utility to include;
6523 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6524 about an ENGINE's available control commands.
6525 - executing control commands from command line arguments using the
6526 '-pre' and '-post' switches. '-post' is only used if '-t' is
6527 specified and the ENGINE is successfully initialised. The syntax for
6528 the individual commands are colon-separated, for example;
6529 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6532 *) New dynamic control command support for ENGINEs. ENGINEs can now
6533 declare their own commands (numbers), names (strings), descriptions,
6534 and input types for run-time discovery by calling applications. A
6535 subset of these commands are implicitly classed as "executable"
6536 depending on their input type, and only these can be invoked through
6537 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6538 can be based on user input, config files, etc). The distinction is
6539 that "executable" commands cannot return anything other than a boolean
6540 result and can only support numeric or string input, whereas some
6541 discoverable commands may only be for direct use through
6542 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6543 pointers, or other custom uses. The "executable" commands are to
6544 support parameterisations of ENGINE behaviour that can be
6545 unambiguously defined by ENGINEs and used consistently across any
6546 OpenSSL-based application. Commands have been added to all the
6547 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6548 control over shared-library paths without source code alterations.
6551 *) Changed all ENGINE implementations to dynamically allocate their
6552 ENGINEs rather than declaring them statically. Apart from this being
6553 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6554 this also allows the implementations to compile without using the
6555 internal engine_int.h header.
6558 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6559 'const' value. Any code that should be able to modify a RAND_METHOD
6560 should already have non-const pointers to it (ie. they should only
6561 modify their own ones).
6564 *) Made a variety of little tweaks to the ENGINE code.
6565 - "atalla" and "ubsec" string definitions were moved from header files
6566 to C code. "nuron" string definitions were placed in variables
6567 rather than hard-coded - allowing parameterisation of these values
6568 later on via ctrl() commands.
6569 - Removed unused "#if 0"'d code.
6570 - Fixed engine list iteration code so it uses ENGINE_free() to release
6571 structural references.
6572 - Constified the RAND_METHOD element of ENGINE structures.
6573 - Constified various get/set functions as appropriate and added
6574 missing functions (including a catch-all ENGINE_cpy that duplicates
6575 all ENGINE values onto a new ENGINE except reference counts/state).
6576 - Removed NULL parameter checks in get/set functions. Setting a method
6577 or function to NULL is a way of cancelling out a previously set
6578 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6579 and doesn't justify the extra error symbols and code.
6580 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6581 flags from engine_int.h to engine.h.
6582 - Changed prototypes for ENGINE handler functions (init(), finish(),
6583 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6586 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6587 to the algorithm using long division. The binary algorithm can be
6588 used only if the modulus is odd. On 32-bit systems, it is faster
6589 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6590 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6591 up to 450 bits. In 64-bit environments, the binary algorithm
6592 appears to be advantageous for much longer moduli; here we use it
6593 for moduli up to 2048 bits.
6596 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6597 could not support the combine flag in choice fields.
6600 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6601 extensions from a certificate request to the certificate.
6604 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6605 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6606 file: this allows the display of the certificate about to be
6607 signed to be customised, to allow certain fields to be included
6608 or excluded and extension details. The old system didn't display
6609 multicharacter strings properly, omitted fields not in the policy
6610 and couldn't display additional details such as extensions.
6613 *) Function EC_POINTs_mul for multiple scalar multiplication
6614 of an arbitrary number of elliptic curve points
6615 \sum scalars[i]*points[i],
6616 optionally including the generator defined for the EC_GROUP:
6617 scalar*generator + \sum scalars[i]*points[i].
6619 EC_POINT_mul is a simple wrapper function for the typical case
6620 that the point list has just one item (besides the optional
6624 *) First EC_METHODs for curves over GF(p):
6626 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6627 operations and provides various method functions that can also
6628 operate with faster implementations of modular arithmetic.
6630 EC_GFp_mont_method() reuses most functions that are part of
6631 EC_GFp_simple_method, but uses Montgomery arithmetic.
6633 [Bodo Moeller; point addition and point doubling
6634 implementation directly derived from source code provided by
6635 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6637 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6638 crypto/ec/ec_lib.c):
6640 Curves are EC_GROUP objects (with an optional group generator)
6641 based on EC_METHODs that are built into the library.
6643 Points are EC_POINT objects based on EC_GROUP objects.
6645 Most of the framework would be able to handle curves over arbitrary
6646 finite fields, but as there are no obvious types for fields other
6647 than GF(p), some functions are limited to that for now.
6650 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6651 that the file contains a complete HTTP response.
6654 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6655 change the def and num file printf format specifier from "%-40sXXX"
6656 to "%-39s XXX". The latter will always guarantee a space after the
6657 field while the former will cause them to run together if the field
6658 is 40 of more characters long.
6661 *) Constify the cipher and digest 'method' functions and structures
6662 and modify related functions to take constant EVP_MD and EVP_CIPHER
6666 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6667 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6670 *) Modify EVP_Digest*() routines so they now return values. Although the
6671 internal software routines can never fail additional hardware versions
6675 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6677 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6678 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6681 ERR_R_NESTED_ASN1_ERROR
6683 ERR_R_MISSING_ASN1_EOS
6684 were 4 .. 9, conflicting with
6685 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6687 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6688 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6690 Add new error code 'ERR_R_INTERNAL_ERROR'.
6693 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6697 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6698 sets the subject name for a new request or supersedes the
6699 subject name in a given request. Formats that can be parsed are
6700 'CN=Some Name, OU=myOU, C=IT'
6702 'CN=Some Name/OU=myOU/C=IT'.
6704 Add options '-batch' and '-verbose' to 'openssl req'.
6705 [Massimiliano Pala <madwolf@hackmasters.net>]
6707 *) Introduce the possibility to access global variables through
6708 functions on platform were that's the best way to handle exporting
6709 global variables in shared libraries. To enable this functionality,
6710 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6711 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6712 is normally done by Configure or something similar).
6714 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6715 in the source file (foo.c) like this:
6717 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6718 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6720 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6721 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6723 OPENSSL_DECLARE_GLOBAL(int,foo);
6724 #define foo OPENSSL_GLOBAL_REF(foo)
6725 OPENSSL_DECLARE_GLOBAL(double,bar);
6726 #define bar OPENSSL_GLOBAL_REF(bar)
6728 The #defines are very important, and therefore so is including the
6729 header file everywhere where the defined globals are used.
6731 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6732 of ASN.1 items, but that structure is a bit different.
6734 The largest change is in util/mkdef.pl which has been enhanced with
6735 better and easier to understand logic to choose which symbols should
6736 go into the Windows .def files as well as a number of fixes and code
6737 cleanup (among others, algorithm keywords are now sorted
6738 lexicographically to avoid constant rewrites).
6741 *) In BN_div() keep a copy of the sign of 'num' before writing the
6742 result to 'rm' because if rm==num the value will be overwritten
6743 and produce the wrong result if 'num' is negative: this caused
6744 problems with BN_mod() and BN_nnmod().
6747 *) Function OCSP_request_verify(). This checks the signature on an
6748 OCSP request and verifies the signer certificate. The signer
6749 certificate is just checked for a generic purpose and OCSP request
6753 *) Add OCSP_check_validity() function to check the validity of OCSP
6754 responses. OCSP responses are prepared in real time and may only
6755 be a few seconds old. Simply checking that the current time lies
6756 between thisUpdate and nextUpdate max reject otherwise valid responses
6757 caused by either OCSP responder or client clock inaccuracy. Instead
6758 we allow thisUpdate and nextUpdate to fall within a certain period of
6759 the current time. The age of the response can also optionally be
6760 checked. Two new options -validity_period and -status_age added to
6764 *) If signature or public key algorithm is unrecognized print out its
6765 OID rather that just UNKNOWN.
6768 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6769 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6770 ID to be generated from the issuer certificate alone which can then be
6771 passed to OCSP_id_issuer_cmp().
6774 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6775 ASN1 modules to export functions returning ASN1_ITEM pointers
6776 instead of the ASN1_ITEM structures themselves. This adds several
6777 new macros which allow the underlying ASN1 function/structure to
6778 be accessed transparently. As a result code should not use ASN1_ITEM
6779 references directly (such as &X509_it) but instead use the relevant
6780 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6781 use of the new ASN1 code on platforms where exporting structures
6782 is problematical (for example in shared libraries) but exporting
6783 functions returning pointers to structures is not.
6786 *) Add support for overriding the generation of SSL/TLS session IDs.
6787 These callbacks can be registered either in an SSL_CTX or per SSL.
6788 The purpose of this is to allow applications to control, if they wish,
6789 the arbitrary values chosen for use as session IDs, particularly as it
6790 can be useful for session caching in multiple-server environments. A
6791 command-line switch for testing this (and any client code that wishes
6792 to use such a feature) has been added to "s_server".
6793 [Geoff Thorpe, Lutz Jaenicke]
6795 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6796 of the form '#if defined(...) || defined(...) || ...' and
6797 '#if !defined(...) && !defined(...) && ...'. This also avoids
6798 the growing number of special cases it was previously handling.
6801 *) Make all configuration macros available for application by making
6802 sure they are available in opensslconf.h, by giving them names starting
6803 with "OPENSSL_" to avoid conflicts with other packages and by making
6804 sure e_os2.h will cover all platform-specific cases together with
6806 Additionally, it is now possible to define configuration/platform-
6807 specific names (called "system identities"). In the C code, these
6808 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6809 macro with the name beginning with "OPENSSL_SYS_", which is determined
6810 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6814 *) New option -set_serial to 'req' and 'x509' this allows the serial
6815 number to use to be specified on the command line. Previously self
6816 signed certificates were hard coded with serial number 0 and the
6817 CA options of 'x509' had to use a serial number in a file which was
6821 *) New options to 'ca' utility to support V2 CRL entry extensions.
6822 Currently CRL reason, invalidity date and hold instruction are
6823 supported. Add new CRL extensions to V3 code and some new objects.
6826 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6827 disable standard block padding (aka PKCS#5 padding) in the EVP
6828 API, which was previously mandatory. This means that the data is
6829 not padded in any way and so the total length much be a multiple
6830 of the block size, otherwise an error occurs.
6833 *) Initial (incomplete) OCSP SSL support.
6836 *) New function OCSP_parse_url(). This splits up a URL into its host,
6837 port and path components: primarily to parse OCSP URLs. New -url
6838 option to ocsp utility.
6841 *) New nonce behavior. The return value of OCSP_check_nonce() now
6842 reflects the various checks performed. Applications can decide
6843 whether to tolerate certain situations such as an absent nonce
6844 in a response when one was present in a request: the ocsp application
6845 just prints out a warning. New function OCSP_add1_basic_nonce()
6846 this is to allow responders to include a nonce in a response even if
6847 the request is nonce-less.
6850 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6851 skipped when using openssl x509 multiple times on a single input file,
6852 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6855 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6856 set string type: to handle setting ASN1_TIME structures. Fix ca
6857 utility to correctly initialize revocation date of CRLs.
6860 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6861 the clients preferred ciphersuites and rather use its own preferences.
6862 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6863 Internet Explorer by ensuring unchanged hash method during stepup.
6864 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6867 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6868 to aes and add a new 'exist' option to print out symbols that don't
6872 *) Additional options to ocsp utility to allow flags to be set and
6873 additional certificates supplied.
6876 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6877 OCSP client a number of certificate to only verify the response
6881 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6882 handle the new API. Currently only ECB, CBC modes supported. Add new
6885 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6886 Encryption Standard (AES) Ciphersuites for Transport Layer
6887 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6888 not enabled by default and were not part of the "ALL" ciphersuite
6889 alias because they were not yet official; they could be
6890 explicitly requested by specifying the "AESdraft" ciphersuite
6891 group alias. In the final release of OpenSSL 0.9.7, the group
6892 alias is called "AES" and is part of "ALL".)
6893 [Ben Laurie, Steve Henson, Bodo Moeller]
6895 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6896 request to response.
6899 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6900 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6901 extract information from a certificate request. OCSP_response_create()
6902 creates a response and optionally adds a basic response structure.
6903 OCSP_basic_add1_status() adds a complete single response to a basic
6904 response and returns the OCSP_SINGLERESP structure just added (to allow
6905 extensions to be included for example). OCSP_basic_add1_cert() adds a
6906 certificate to a basic response and OCSP_basic_sign() signs a basic
6907 response with various flags. New helper functions ASN1_TIME_check()
6908 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6909 (converts ASN1_TIME to GeneralizedTime).
6912 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6913 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6914 structure from a certificate. X509_pubkey_digest() digests the public_key
6915 contents: this is used in various key identifiers.
6918 *) Make sk_sort() tolerate a NULL argument.
6919 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6921 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6922 passed by the function are trusted implicitly. If any of them signed the
6923 response then it is assumed to be valid and is not verified.
6926 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6927 to data. This was previously part of the PKCS7 ASN1 code. This
6928 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6929 [Steve Henson, reported by Kenneth R. Robinette
6930 <support@securenetterm.com>]
6932 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6933 routines: without these tracing memory leaks is very painful.
6934 Fix leaks in PKCS12 and PKCS7 routines.
6937 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6938 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6939 effectively meant GeneralizedTime would never be used. Now it
6940 is initialised to -1 but X509_time_adj() now has to check the value
6941 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6942 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6943 [Steve Henson, reported by Kenneth R. Robinette
6944 <support@securenetterm.com>]
6946 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6947 result in a zero length in the ASN1_INTEGER structure which was
6948 not consistent with the structure when d2i_ASN1_INTEGER() was used
6949 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6950 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6951 where it did not print out a minus for negative ASN1_INTEGER.
6954 *) Add summary printout to ocsp utility. The various functions which
6955 convert status values to strings have been renamed to:
6956 OCSP_response_status_str(), OCSP_cert_status_str() and
6957 OCSP_crl_reason_str() and are no longer static. New options
6958 to verify nonce values and to disable verification. OCSP response
6959 printout format cleaned up.
6962 *) Add additional OCSP certificate checks. These are those specified
6963 in RFC2560. This consists of two separate checks: the CA of the
6964 certificate being checked must either be the OCSP signer certificate
6965 or the issuer of the OCSP signer certificate. In the latter case the
6966 OCSP signer certificate must contain the OCSP signing extended key
6967 usage. This check is performed by attempting to match the OCSP
6968 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6969 in the OCSP_CERTID structures of the response.
6972 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6973 and related routines. This uses the standard OpenSSL certificate
6974 verify routines to perform initial checks (just CA validity) and
6975 to obtain the certificate chain. Then additional checks will be
6976 performed on the chain. Currently the root CA is checked to see
6977 if it is explicitly trusted for OCSP signing. This is used to set
6978 a root CA as a global signing root: that is any certificate that
6979 chains to that CA is an acceptable OCSP signing certificate.
6982 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6983 extensions from a separate configuration file.
6984 As when reading extensions from the main configuration file,
6985 the '-extensions ...' option may be used for specifying the
6987 [Massimiliano Pala <madwolf@comune.modena.it>]
6989 *) New OCSP utility. Allows OCSP requests to be generated or
6990 read. The request can be sent to a responder and the output
6991 parsed, outputed or printed in text form. Not complete yet:
6992 still needs to check the OCSP response validity.
6995 *) New subcommands for 'openssl ca':
6996 'openssl ca -status <serial>' prints the status of the cert with
6997 the given serial number (according to the index file).
6998 'openssl ca -updatedb' updates the expiry status of certificates
7000 [Massimiliano Pala <madwolf@comune.modena.it>]
7002 *) New '-newreq-nodes' command option to CA.pl. This is like
7003 '-newreq', but calls 'openssl req' with the '-nodes' option
7004 so that the resulting key is not encrypted.
7005 [Damien Miller <djm@mindrot.org>]
7007 *) New configuration for the GNU Hurd.
7008 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7010 *) Initial code to implement OCSP basic response verify. This
7011 is currently incomplete. Currently just finds the signer's
7012 certificate and verifies the signature on the response.
7015 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7016 value of OPENSSLDIR. This is available via the new '-d' option
7017 to 'openssl version', and is also included in 'openssl version -a'.
7020 *) Allowing defining memory allocation callbacks that will be given
7021 file name and line number information in additional arguments
7022 (a const char* and an int). The basic functionality remains, as
7023 well as the original possibility to just replace malloc(),
7024 realloc() and free() by functions that do not know about these
7025 additional arguments. To register and find out the current
7026 settings for extended allocation functions, the following
7027 functions are provided:
7029 CRYPTO_set_mem_ex_functions
7030 CRYPTO_set_locked_mem_ex_functions
7031 CRYPTO_get_mem_ex_functions
7032 CRYPTO_get_locked_mem_ex_functions
7034 These work the same way as CRYPTO_set_mem_functions and friends.
7035 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7036 extended allocation function is enabled.
7037 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7038 a conventional allocation function is enabled.
7039 [Richard Levitte, Bodo Moeller]
7041 *) Finish off removing the remaining LHASH function pointer casts.
7042 There should no longer be any prototype-casting required when using
7043 the LHASH abstraction, and any casts that remain are "bugs". See
7044 the callback types and macros at the head of lhash.h for details
7045 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7048 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7049 If /dev/[u]random devices are not available or do not return enough
7050 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7052 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7053 /etc/entropy will be queried once each in this sequence, quering stops
7054 when enough entropy was collected without querying more sockets.
7057 *) Change the Unix RAND_poll() variant to be able to poll several
7058 random devices, as specified by DEVRANDOM, until a sufficient amount
7059 of data has been collected. We spend at most 10 ms on each file
7060 (select timeout) and read in non-blocking mode. DEVRANDOM now
7061 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7062 (previously it was just the string "/dev/urandom"), so on typical
7063 platforms the 10 ms delay will never occur.
7064 Also separate out the Unix variant to its own file, rand_unix.c.
7065 For VMS, there's a currently-empty rand_vms.c.
7068 *) Move OCSP client related routines to ocsp_cl.c. These
7069 provide utility functions which an application needing
7070 to issue a request to an OCSP responder and analyse the
7071 response will typically need: as opposed to those which an
7072 OCSP responder itself would need which will be added later.
7074 OCSP_request_sign() signs an OCSP request with an API similar
7075 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7076 response. OCSP_response_get1_basic() extracts basic response
7077 from response. OCSP_resp_find_status(): finds and extracts status
7078 information from an OCSP_CERTID structure (which will be created
7079 when the request structure is built). These are built from lower
7080 level functions which work on OCSP_SINGLERESP structures but
7081 wont normally be used unless the application wishes to examine
7082 extensions in the OCSP response for example.
7084 Replace nonce routines with a pair of functions.
7085 OCSP_request_add1_nonce() adds a nonce value and optionally
7086 generates a random value. OCSP_check_nonce() checks the
7087 validity of the nonce in an OCSP response.
7090 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7091 This doesn't copy the supplied OCSP_CERTID and avoids the
7092 need to free up the newly created id. Change return type
7093 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7094 This can then be used to add extensions to the request.
7095 Deleted OCSP_request_new(), since most of its functionality
7096 is now in OCSP_REQUEST_new() (and the case insensitive name
7097 clash) apart from the ability to set the request name which
7098 will be added elsewhere.
7101 *) Update OCSP API. Remove obsolete extensions argument from
7102 various functions. Extensions are now handled using the new
7103 OCSP extension code. New simple OCSP HTTP function which
7104 can be used to send requests and parse the response.
7107 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7108 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7109 uses the special reorder version of SET OF to sort the attributes
7110 and reorder them to match the encoded order. This resolves a long
7111 standing problem: a verify on a PKCS7 structure just after signing
7112 it used to fail because the attribute order did not match the
7113 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7114 it uses the received order. This is necessary to tolerate some broken
7115 software that does not order SET OF. This is handled by encoding
7116 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7117 to produce the required SET OF.
7120 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7121 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7122 files to get correct declarations of the ASN.1 item variables.
7125 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7126 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7127 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7128 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7129 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7130 ASN1_ITEM and no wrapper functions.
7133 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7134 replace the old function pointer based I/O routines. Change most of
7135 the *_d2i_bio() and *_d2i_fp() functions to use these.
7138 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7139 lines, recognice more "algorithms" that can be deselected, and make
7140 it complain about algorithm deselection that isn't recognised.
7143 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7144 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7145 to use new functions. Add NO_ASN1_OLD which can be set to remove
7146 some old style ASN1 functions: this can be used to determine if old
7147 code will still work when these eventually go away.
7150 *) New extension functions for OCSP structures, these follow the
7151 same conventions as certificates and CRLs.
7154 *) New function X509V3_add1_i2d(). This automatically encodes and
7155 adds an extension. Its behaviour can be customised with various
7156 flags to append, replace or delete. Various wrappers added for
7157 certifcates and CRLs.
7160 *) Fix to avoid calling the underlying ASN1 print routine when
7161 an extension cannot be parsed. Correct a typo in the
7162 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7165 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7166 entries for variables.
7169 *) Add functionality to apps/openssl.c for detecting locking
7170 problems: As the program is single-threaded, all we have
7171 to do is register a locking callback using an array for
7172 storing which locks are currently held by the program.
7175 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7176 SSL_get_ex_data_X509_STORE_idx(), which is used in
7177 ssl_verify_cert_chain() and thus can be called at any time
7178 during TLS/SSL handshakes so that thread-safety is essential.
7179 Unfortunately, the ex_data design is not at all suited
7180 for multi-threaded use, so it probably should be abolished.
7183 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7184 [Broadcom, tweaked and integrated by Geoff Thorpe]
7186 *) Move common extension printing code to new function
7187 X509V3_print_extensions(). Reorganise OCSP print routines and
7188 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7191 *) New function X509_signature_print() to remove duplication in some
7195 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7196 set (this was treated exactly the same as SET OF previously). This
7197 is used to reorder the STACK representing the structure to match the
7198 encoding. This will be used to get round a problem where a PKCS7
7199 structure which was signed could not be verified because the STACK
7200 order did not reflect the encoded order.
7203 *) Reimplement the OCSP ASN1 module using the new code.
7206 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7207 for its ASN1 operations. The old style function pointers still exist
7208 for now but they will eventually go away.
7211 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7212 completely replaces the old ASN1 functionality with a table driven
7213 encoder and decoder which interprets an ASN1_ITEM structure describing
7214 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7215 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7216 has also been converted to the new form.
7219 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7220 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7221 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7222 for negative moduli.
7225 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7226 of not touching the result's sign bit.
7229 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7233 *) Changed the LHASH code to use prototypes for callbacks, and created
7234 macros to declare and implement thin (optionally static) functions
7235 that provide type-safety and avoid function pointer casting for the
7236 type-specific callbacks.
7239 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7241 [Veers Staats <staatsvr@asc.hpc.mil>,
7242 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7244 *) Reformat the FAQ so the different questions and answers can be divided
7245 in sections depending on the subject.
7248 *) Have the zlib compression code load ZLIB.DLL dynamically under
7252 *) New function BN_mod_sqrt for computing square roots modulo a prime
7253 (using the probabilistic Tonelli-Shanks algorithm unless
7254 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7255 be handled deterministically).
7256 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7258 *) Make BN_mod_inverse faster by explicitly handling small quotients
7259 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7260 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7263 *) New function BN_kronecker.
7266 *) Fix BN_gcd so that it works on negative inputs; the result is
7267 positive unless both parameters are zero.
7268 Previously something reasonably close to an infinite loop was
7269 possible because numbers could be growing instead of shrinking
7270 in the implementation of Euclid's algorithm.
7273 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7274 sign of the number in question.
7276 Fix BN_is_word(a,w) to work correctly for w == 0.
7278 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7279 because its test if the absolute value of 'a' equals 'w'.
7280 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7281 it exists mostly for use in the implementations of BN_is_zero(),
7282 BN_is_one(), and BN_is_word().
7285 *) New function BN_swap.
7288 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7289 the exponentiation functions are more likely to produce reasonable
7290 results on negative inputs.
7293 *) Change BN_mod_mul so that the result is always non-negative.
7294 Previously, it could be negative if one of the factors was negative;
7295 I don't think anyone really wanted that behaviour.
7298 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7299 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7300 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7301 and add new functions:
7310 BN_mod_lshift1_quick
7314 These functions always generate non-negative results.
7316 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7317 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7319 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7320 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7321 be reduced modulo m.
7322 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7325 The following entry accidentally appeared in the CHANGES file
7326 distributed with OpenSSL 0.9.7. The modifications described in
7327 it do *not* apply to OpenSSL 0.9.7.
7329 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7330 was actually never needed) and in BN_mul(). The removal in BN_mul()
7331 required a small change in bn_mul_part_recursive() and the addition
7332 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7333 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7334 bn_sub_words() and bn_add_words() except they take arrays with
7339 *) In 'openssl passwd', verify passwords read from the terminal
7340 unless the '-salt' option is used (which usually means that
7341 verification would just waste user's time since the resulting
7342 hash is going to be compared with some given password hash)
7343 or the new '-noverify' option is used.
7345 This is an incompatible change, but it does not affect
7346 non-interactive use of 'openssl passwd' (passwords on the command
7347 line, '-stdin' option, '-in ...' option) and thus should not
7351 *) Remove all references to RSAref, since there's no more need for it.
7354 *) Make DSO load along a path given through an environment variable
7355 (SHLIB_PATH) with shl_load().
7358 *) Constify the ENGINE code as a result of BIGNUM constification.
7359 Also constify the RSA code and most things related to it. In a
7360 few places, most notable in the depth of the ASN.1 code, ugly
7361 casts back to non-const were required (to be solved at a later
7365 *) Make it so the openssl application has all engines loaded by default.
7368 *) Constify the BIGNUM routines a little more.
7371 *) Add the following functions:
7373 ENGINE_load_cswift()
7375 ENGINE_load_atalla()
7377 ENGINE_load_builtin_engines()
7379 That way, an application can itself choose if external engines that
7380 are built-in in OpenSSL shall ever be used or not. The benefit is
7381 that applications won't have to be linked with libdl or other dso
7382 libraries unless it's really needed.
7384 Changed 'openssl engine' to load all engines on demand.
7385 Changed the engine header files to avoid the duplication of some
7386 declarations (they differed!).
7389 *) 'openssl engine' can now list capabilities.
7392 *) Better error reporting in 'openssl engine'.
7395 *) Never call load_dh_param(NULL) in s_server.
7398 *) Add engine application. It can currently list engines by name and
7399 identity, and test if they are actually available.
7402 *) Improve RPM specification file by forcing symbolic linking and making
7403 sure the installed documentation is also owned by root.root.
7404 [Damien Miller <djm@mindrot.org>]
7406 *) Give the OpenSSL applications more possibilities to make use of
7407 keys (public as well as private) handled by engines.
7410 *) Add OCSP code that comes from CertCo.
7413 *) Add VMS support for the Rijndael code.
7416 *) Added untested support for Nuron crypto accelerator.
7419 *) Add support for external cryptographic devices. This code was
7420 previously distributed separately as the "engine" branch.
7421 [Geoff Thorpe, Richard Levitte]
7423 *) Rework the filename-translation in the DSO code. It is now possible to
7424 have far greater control over how a "name" is turned into a filename
7425 depending on the operating environment and any oddities about the
7426 different shared library filenames on each system.
7429 *) Support threads on FreeBSD-elf in Configure.
7432 *) Fix for SHA1 assembly problem with MASM: it produces
7433 warnings about corrupt line number information when assembling
7434 with debugging information. This is caused by the overlapping
7436 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7439 NCONF_get_number() has no error checking at all. As a replacement,
7440 NCONF_get_number_e() is defined (_e for "error checking") and is
7441 promoted strongly. The old NCONF_get_number is kept around for
7442 binary backward compatibility.
7443 Make it possible for methods to load from something other than a BIO,
7444 by providing a function pointer that is given a name instead of a BIO.
7445 For example, this could be used to load configuration data from an
7449 *) Fix for non blocking accept BIOs. Added new I/O special reason
7450 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7451 with non blocking I/O was not possible because no retry code was
7452 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7456 *) Added the beginnings of Rijndael support.
7459 *) Fix for bug in DirectoryString mask setting. Add support for
7460 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7461 to allow certificate printing to more controllable, additional
7462 'certopt' option to 'x509' to allow new printing options to be
7466 *) Clean old EAY MD5 hack from e_os.h.
7469 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7471 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7472 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7473 [Joe Orton, Steve Henson]
7475 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7477 *) Fix additional bug revealed by the NISCC test suite:
7479 Stop bug triggering large recursion when presented with
7480 certain ASN.1 tags (CVE-2003-0851)
7483 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7485 *) Fix various bugs revealed by running the NISCC test suite:
7487 Stop out of bounds reads in the ASN1 code when presented with
7488 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7490 If verify callback ignores invalid public key errors don't try to check
7491 certificate signature with the NULL public key.
7495 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7496 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7500 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7501 extra data after the compression methods not only for TLS 1.0
7502 but also for SSL 3.0 (as required by the specification).
7503 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7505 *) Change X509_certificate_type() to mark the key as exported/exportable
7506 when it's 512 *bits* long, not 512 bytes.
7509 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7511 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7512 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7513 a protocol version number mismatch like a decryption error
7514 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7517 *) Turn on RSA blinding by default in the default implementation
7518 to avoid a timing attack. Applications that don't want it can call
7519 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7520 They would be ill-advised to do so in most cases.
7521 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7523 *) Change RSA blinding code so that it works when the PRNG is not
7524 seeded (in this case, the secret RSA exponent is abused as
7525 an unpredictable seed -- if it is not unpredictable, there
7526 is no point in blinding anyway). Make RSA blinding thread-safe
7527 by remembering the creator's thread ID in rsa->blinding and
7528 having all other threads use local one-time blinding factors
7529 (this requires more computation than sharing rsa->blinding, but
7530 avoids excessive locking; and if an RSA object is not shared
7531 between threads, blinding will still be very fast).
7534 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7536 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7537 via timing by performing a MAC computation even if incorrrect
7538 block cipher padding has been found. This is a countermeasure
7539 against active attacks where the attacker has to distinguish
7540 between bad padding and a MAC verification error. (CVE-2003-0078)
7542 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7543 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7544 Martin Vuagnoux (EPFL, Ilion)]
7546 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7548 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7549 memory from it's contents. This is done with a counter that will
7550 place alternating values in each byte. This can be used to solve
7551 two issues: 1) the removal of calls to memset() by highly optimizing
7552 compilers, and 2) cleansing with other values than 0, since those can
7553 be read through on certain media, for example a swap space on disk.
7556 *) Bugfix: client side session caching did not work with external caching,
7557 because the session->cipher setting was not restored when reloading
7558 from the external cache. This problem was masked, when
7559 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7560 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7563 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7564 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7565 [Zeev Lieber <zeev-l@yahoo.com>]
7567 *) Undo an undocumented change introduced in 0.9.6e which caused
7568 repeated calls to OpenSSL_add_all_ciphers() and
7569 OpenSSL_add_all_digests() to be ignored, even after calling
7573 *) Change the default configuration reader to deal with last line not
7574 being properly terminated.
7577 *) Change X509_NAME_cmp() so it applies the special rules on handling
7578 DN values that are of type PrintableString, as well as RDNs of type
7579 emailAddress where the value has the type ia5String.
7580 [stefank@valicert.com via Richard Levitte]
7582 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7583 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7584 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7585 the bitwise-OR of the two for use by the majority of applications
7586 wanting this behaviour, and update the docs. The documented
7587 behaviour and actual behaviour were inconsistent and had been
7588 changing anyway, so this is more a bug-fix than a behavioural
7590 [Geoff Thorpe, diagnosed by Nadav Har'El]
7592 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7593 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7596 *) Fix initialization code race conditions in
7597 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7598 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7599 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7600 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7601 ssl2_get_cipher_by_char(),
7602 ssl3_get_cipher_by_char().
7603 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7605 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7606 the cached sessions are flushed, as the remove_cb() might use ex_data
7607 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7608 (see [openssl.org #212]).
7609 [Geoff Thorpe, Lutz Jaenicke]
7611 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7612 length, instead of the encoding length to d2i_ASN1_OBJECT.
7615 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7617 *) [In 0.9.6g-engine release:]
7618 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7619 [Lynn Gazis <lgazis@rainbow.com>]
7621 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7623 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7624 and get fix the header length calculation.
7625 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7626 Alon Kantor <alonk@checkpoint.com> (and others),
7629 *) Use proper error handling instead of 'assertions' in buffer
7630 overflow checks added in 0.9.6e. This prevents DoS (the
7631 assertions could call abort()).
7632 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7634 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7636 *) Add various sanity checks to asn1_get_length() to reject
7637 the ASN1 length bytes if they exceed sizeof(long), will appear
7638 negative or the content length exceeds the length of the
7640 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7642 *) Fix cipher selection routines: ciphers without encryption had no flags
7643 for the cipher strength set and where therefore not handled correctly
7644 by the selection routines (PR #130).
7647 *) Fix EVP_dsa_sha macro.
7651 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7652 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7653 that was added in OpenSSL 0.9.6d.
7655 As the countermeasure turned out to be incompatible with some
7656 broken SSL implementations, the new option is part of SSL_OP_ALL.
7657 SSL_OP_ALL is usually employed when compatibility with weird SSL
7658 implementations is desired (e.g. '-bugs' option to 's_client' and
7659 's_server'), so the new option is automatically set in many
7663 *) Changes in security patch:
7665 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7666 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7667 Air Force Materiel Command, USAF, under agreement number
7670 *) Add various sanity checks to asn1_get_length() to reject
7671 the ASN1 length bytes if they exceed sizeof(long), will appear
7672 negative or the content length exceeds the length of the
7673 supplied buffer. (CVE-2002-0659)
7674 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7676 *) Assertions for various potential buffer overflows, not known to
7678 [Ben Laurie (CHATS)]
7680 *) Various temporary buffers to hold ASCII versions of integers were
7681 too small for 64 bit platforms. (CVE-2002-0655)
7682 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7684 *) Remote buffer overflow in SSL3 protocol - an attacker could
7685 supply an oversized session ID to a client. (CVE-2002-0656)
7686 [Ben Laurie (CHATS)]
7688 *) Remote buffer overflow in SSL2 protocol - an attacker could
7689 supply an oversized client master key. (CVE-2002-0656)
7690 [Ben Laurie (CHATS)]
7692 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7694 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7695 encoded as NULL) with id-dsa-with-sha1.
7696 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7698 *) Check various X509_...() return values in apps/req.c.
7699 [Nils Larsch <nla@trustcenter.de>]
7701 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7702 an end-of-file condition would erronously be flagged, when the CRLF
7703 was just at the end of a processed block. The bug was discovered when
7704 processing data through a buffering memory BIO handing the data to a
7705 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7706 <ptsekov@syntrex.com> and Nedelcho Stanev.
7709 *) Implement a countermeasure against a vulnerability recently found
7710 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7711 before application data chunks to avoid the use of known IVs
7712 with data potentially chosen by the attacker.
7715 *) Fix length checks in ssl3_get_client_hello().
7718 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7719 to prevent ssl3_read_internal() from incorrectly assuming that
7720 ssl3_read_bytes() found application data while handshake
7721 processing was enabled when in fact s->s3->in_read_app_data was
7722 merely automatically cleared during the initial handshake.
7723 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7725 *) Fix object definitions for Private and Enterprise: they were not
7726 recognized in their shortname (=lowercase) representation. Extend
7727 obj_dat.pl to issue an error when using undefined keywords instead
7728 of silently ignoring the problem (Svenning Sorensen
7729 <sss@sss.dnsalias.net>).
7732 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7733 generators, i.e. generators other than 2 and 5. (Previously, the
7734 code did not properly initialise the 'add' and 'rem' values to
7735 BN_generate_prime().)
7737 In the new general case, we do not insist that 'generator' is
7738 actually a primitive root: This requirement is rather pointless;
7739 a generator of the order-q subgroup is just as good, if not
7743 *) Map new X509 verification errors to alerts. Discovered and submitted by
7744 Tom Wu <tom@arcot.com>.
7747 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7748 returning non-zero before the data has been completely received
7749 when using non-blocking I/O.
7750 [Bodo Moeller; problem pointed out by John Hughes]
7752 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7753 [Ben Laurie, Lutz Jaenicke]
7755 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7756 Yoram Zahavi <YoramZ@gilian.com>).
7759 *) Add information about CygWin 1.3 and on, and preserve proper
7760 configuration for the versions before that.
7761 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7763 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7764 check whether we deal with a copy of a session and do not delete from
7765 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7766 <izhar@checkpoint.com>.
7769 *) Do not store session data into the internal session cache, if it
7770 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7771 flag is set). Proposed by Aslam <aslam@funk.com>.
7774 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7778 *) [In 0.9.6d-engine release:]
7779 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7780 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7782 *) Add the configuration target linux-s390x.
7783 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7785 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7786 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7787 variable as an indication that a ClientHello message has been
7788 received. As the flag value will be lost between multiple
7789 invocations of ssl3_accept when using non-blocking I/O, the
7790 function may not be aware that a handshake has actually taken
7791 place, thus preventing a new session from being added to the
7794 To avoid this problem, we now set s->new_session to 2 instead of
7795 using a local variable.
7796 [Lutz Jaenicke, Bodo Moeller]
7798 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7799 if the SSL_R_LENGTH_MISMATCH error is detected.
7800 [Geoff Thorpe, Bodo Moeller]
7802 *) New 'shared_ldflag' column in Configure platform table.
7805 *) Fix EVP_CIPHER_mode macro.
7806 ["Dan S. Camper" <dan@bti.net>]
7808 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7809 type, we must throw them away by setting rr->length to 0.
7810 [D P Chang <dpc@qualys.com>]
7812 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7814 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7815 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7816 worked incorrectly for those cases where range = 10..._2 and
7817 3*range is two bits longer than range.)
7820 *) Only add signing time to PKCS7 structures if it is not already
7824 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7825 OBJ_ld_ce should be OBJ_id_ce.
7826 Also some ip-pda OIDs in crypto/objects/objects.txt were
7827 incorrect (cf. RFC 3039).
7828 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7830 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7831 returns early because it has nothing to do.
7832 [Andy Schneider <andy.schneider@bjss.co.uk>]
7834 *) [In 0.9.6c-engine release:]
7835 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7836 [Andy Schneider <andy.schneider@bjss.co.uk>]
7838 *) [In 0.9.6c-engine release:]
7839 Add support for Cryptographic Appliance's keyserver technology.
7840 (Use engine 'keyclient')
7841 [Cryptographic Appliances and Geoff Thorpe]
7843 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7844 is called via tools/c89.sh because arguments have to be
7845 rearranged (all '-L' options must appear before the first object
7847 [Richard Shapiro <rshapiro@abinitio.com>]
7849 *) [In 0.9.6c-engine release:]
7850 Add support for Broadcom crypto accelerator cards, backported
7852 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7854 *) [In 0.9.6c-engine release:]
7855 Add support for SureWare crypto accelerator cards from
7856 Baltimore Technologies. (Use engine 'sureware')
7857 [Baltimore Technologies and Mark Cox]
7859 *) [In 0.9.6c-engine release:]
7860 Add support for crypto accelerator cards from Accelerated
7861 Encryption Processing, www.aep.ie. (Use engine 'aep')
7862 [AEP Inc. and Mark Cox]
7864 *) Add a configuration entry for gcc on UnixWare.
7865 [Gary Benson <gbenson@redhat.com>]
7867 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7868 messages are stored in a single piece (fixed-length part and
7869 variable-length part combined) and fix various bugs found on the way.
7872 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7873 instead. BIO_gethostbyname() does not know what timeouts are
7874 appropriate, so entries would stay in cache even when they have
7876 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7878 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7879 faced with a pathologically small ClientHello fragment that does
7880 not contain client_version: Instead of aborting with an error,
7881 simply choose the highest available protocol version (i.e.,
7882 TLS 1.0 unless it is disabled). In practice, ClientHello
7883 messages are never sent like this, but this change gives us
7884 strictly correct behaviour at least for TLS.
7887 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7888 never resets s->method to s->ctx->method when called from within
7889 one of the SSL handshake functions.
7890 [Bodo Moeller; problem pointed out by Niko Baric]
7892 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7893 (sent using the client's version number) if client_version is
7894 smaller than the protocol version in use. Also change
7895 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7896 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7897 the client will at least see that alert.
7900 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7904 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7905 client receives HelloRequest while in a handshake.
7906 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7908 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7909 should end in 'break', not 'goto end' which circumvents various
7910 cleanups done in state SSL_ST_OK. But session related stuff
7911 must be disabled for SSL_ST_OK in the case that we just sent a
7914 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7915 before just sending a HelloRequest.
7916 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7918 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7919 reveal whether illegal block cipher padding was found or a MAC
7920 verification error occurred. (Neither SSLerr() codes nor alerts
7921 are directly visible to potential attackers, but the information
7922 may leak via logfiles.)
7924 Similar changes are not required for the SSL 2.0 implementation
7925 because the number of padding bytes is sent in clear for SSL 2.0,
7926 and the extra bytes are just ignored. However ssl/s2_pkt.c
7927 failed to verify that the purported number of padding bytes is in
7931 *) Add OpenUNIX-8 support including shared libraries
7932 (Boyd Lynn Gerber <gerberb@zenez.com>).
7935 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7936 'wristwatch attack' using huge encoding parameters (cf.
7937 James H. Manger's CRYPTO 2001 paper). Note that the
7938 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7939 encoding parameters and hence was not vulnerable.
7942 *) BN_sqr() bug fix.
7943 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7945 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7946 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7947 followed by modular reduction.
7948 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7950 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7951 equivalent based on BN_pseudo_rand() instead of BN_rand().
7954 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7955 This function was broken, as the check for a new client hello message
7956 to handle SGC did not allow these large messages.
7957 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7960 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7963 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7964 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7967 *) Rework the configuration and shared library support for Tru64 Unix.
7968 The configuration part makes use of modern compiler features and
7969 still retains old compiler behavior for those that run older versions
7970 of the OS. The shared library support part includes a variant that
7971 uses the RPATH feature, and is available through the special
7972 configuration target "alpha-cc-rpath", which will never be selected
7974 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7976 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7977 with the same message size as in ssl3_get_certificate_request().
7978 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7979 messages might inadvertently be reject as too long.
7980 [Petr Lampa <lampa@fee.vutbr.cz>]
7982 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7985 *) Modified SSL library such that the verify_callback that has been set
7986 specificly for an SSL object with SSL_set_verify() is actually being
7987 used. Before the change, a verify_callback set with this function was
7988 ignored and the verify_callback() set in the SSL_CTX at the time of
7989 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7990 to allow the necessary settings.
7993 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7994 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7995 done automatically (in contradiction to the requirements of the C
7996 standard). This made problems when used from OpenSSH.
7999 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8000 dh->length and always used
8002 BN_rand_range(priv_key, dh->p).
8004 BN_rand_range() is not necessary for Diffie-Hellman, and this
8005 specific range makes Diffie-Hellman unnecessarily inefficient if
8006 dh->length (recommended exponent length) is much smaller than the
8007 length of dh->p. We could use BN_rand_range() if the order of
8008 the subgroup was stored in the DH structure, but we only have
8013 BN_rand(priv_key, l, ...)
8015 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8021 RSA_eay_public_encrypt
8022 RSA_eay_private_decrypt
8023 RSA_eay_private_encrypt (signing)
8024 RSA_eay_public_decrypt (signature verification)
8026 (default implementations for RSA_public_encrypt,
8027 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8028 always reject numbers >= n.
8031 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8032 to synchronize access to 'locking_thread'. This is necessary on
8033 systems where access to 'locking_thread' (an 'unsigned long'
8034 variable) is not atomic.
8037 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8038 *before* setting the 'crypto_lock_rand' flag. The previous code had
8039 a race condition if 0 is a valid thread ID.
8040 [Travis Vitek <vitek@roguewave.com>]
8042 *) Add support for shared libraries under Irix.
8043 [Albert Chin-A-Young <china@thewrittenword.com>]
8045 *) Add configuration option to build on Linux on both big-endian and
8047 [Ralf Baechle <ralf@uni-koblenz.de>]
8049 *) Add the possibility to create shared libraries on HP-UX.
8052 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8054 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8055 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8056 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8057 PRNG state recovery was possible based on the output of
8058 one PRNG request appropriately sized to gain knowledge on
8059 'md' followed by enough consecutive 1-byte PRNG requests
8060 to traverse all of 'state'.
8062 1. When updating 'md_local' (the current thread's copy of 'md')
8063 during PRNG output generation, hash all of the previous
8064 'md_local' value, not just the half used for PRNG output.
8066 2. Make the number of bytes from 'state' included into the hash
8067 independent from the number of PRNG bytes requested.
8069 The first measure alone would be sufficient to avoid
8070 Markku-Juhani's attack. (Actually it had never occurred
8071 to me that the half of 'md_local' used for chaining was the
8072 half from which PRNG output bytes were taken -- I had always
8073 assumed that the secret half would be used.) The second
8074 measure makes sure that additional data from 'state' is never
8075 mixed into 'md_local' in small portions; this heuristically
8076 further strengthens the PRNG.
8079 *) Fix crypto/bn/asm/mips3.s.
8082 *) When only the key is given to "enc", the IV is undefined. Print out
8083 an error message in this case.
8086 *) Handle special case when X509_NAME is empty in X509 printing routines.
8089 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8090 positive and less than q.
8093 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8094 used: it isn't thread safe and the add_lock_callback should handle
8096 [Paul Rose <Paul.Rose@bridge.com>]
8098 *) Verify that incoming data obeys the block size in
8099 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8103 [Ulf Möller, Bodo Möller]
8105 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8106 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8107 when fixing the server behaviour for backwards-compatible 'client
8108 hello' messages. (Note that the attack is impractical against
8109 SSL 3.0 and TLS 1.0 anyway because length and version checking
8110 means that the probability of guessing a valid ciphertext is
8111 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8114 Before 0.9.5, the countermeasure (hide the error by generating a
8115 random 'decryption result') did not work properly because
8116 ERR_clear_error() was missing, meaning that SSL_get_error() would
8117 detect the supposedly ignored error.
8119 Both problems are now fixed.
8122 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8123 (previously it was 1024).
8126 *) Fix for compatibility mode trust settings: ignore trust settings
8127 unless some valid trust or reject settings are present.
8130 *) Fix for blowfish EVP: its a variable length cipher.
8133 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8134 parameters in DSA public key structures and return an error in the
8135 DSA routines if parameters are absent.
8138 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8139 in the current directory if neither $RANDFILE nor $HOME was set.
8140 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8141 caused some confusion to Windows users who haven't defined $HOME.
8142 Thus RAND_file_name() is changed again: e_os.h can define a
8143 DEFAULT_HOME, which will be used if $HOME is not set.
8144 For Windows, we use "C:"; on other platforms, we still require
8145 environment variables.
8147 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8148 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8149 having multiple threads call RAND_poll() concurrently.
8152 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8153 combination of a flag and a thread ID variable.
8154 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8155 flag), *other* threads can enter ssleay_add_bytes without obeying
8156 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8157 that they do not hold after the first thread unsets add_do_not_lock).
8160 *) Change bctest again: '-x' expressions are not available in all
8164 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8166 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8167 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8169 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8170 the default extension for executables, if any. Also, make the perl
8171 scripts that use symlink() to test if it really exists and use "cp"
8172 if it doesn't. All this made OpenSSL compilable and installable in
8176 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8177 If SEQUENCE is length is indefinite just set c->slen to the total
8178 amount of data available.
8179 [Steve Henson, reported by shige@FreeBSD.org]
8180 [This change does not apply to 0.9.7.]
8182 *) Change bctest to avoid here-documents inside command substitution
8183 (workaround for FreeBSD /bin/sh bug).
8184 For compatibility with Ultrix, avoid shell functions (introduced
8185 in the bctest version that searches along $PATH).
8188 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8189 with des_encrypt() defined on some operating systems, like Solaris
8193 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8194 On the Importance of Eliminating Errors in Cryptographic
8195 Computations, J. Cryptology 14 (2001) 2, 101-119,
8196 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8199 *) MIPS assembler BIGNUM division bug fix.
8202 *) Disabled incorrect Alpha assembler code.
8205 *) Fix PKCS#7 decode routines so they correctly update the length
8206 after reading an EOC for the EXPLICIT tag.
8208 [This change does not apply to 0.9.7.]
8210 *) Fix bug in PKCS#12 key generation routines. This was triggered
8211 if a 3DES key was generated with a 0 initial byte. Include
8212 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8213 (but broken) behaviour.
8216 *) Enhance bctest to search for a working bc along $PATH and print
8218 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8220 *) Fix memory leaks in err.c: free err_data string if necessary;
8221 don't write to the wrong index in ERR_set_error_data.
8224 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8228 *) Replace rdtsc with _emit statements for VC++ version 5.
8229 [Jeremy Cooper <jeremy@baymoo.org>]
8231 *) Make it possible to reuse SSLv2 sessions.
8234 *) In copy_email() check for >= 0 as a return value for
8235 X509_NAME_get_index_by_NID() since 0 is a valid index.
8236 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8238 *) Avoid coredump with unsupported or invalid public keys by checking if
8239 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8240 PKCS7_verify() fails with non detached data.
8243 *) Don't use getenv in library functions when run as setuid/setgid.
8244 New function OPENSSL_issetugid().
8247 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8248 due to incorrect handling of multi-threading:
8250 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8252 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8254 3. Count how many times MemCheck_off() has been called so that
8255 nested use can be treated correctly. This also avoids
8256 inband-signalling in the previous code (which relied on the
8257 assumption that thread ID 0 is impossible).
8260 *) Add "-rand" option also to s_client and s_server.
8263 *) Fix CPU detection on Irix 6.x.
8264 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8265 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8267 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8270 [This change does not apply to 0.9.7.]
8272 *) Use the cached encoding of an X509_NAME structure rather than
8273 copying it. This is apparently the reason for the libsafe "errors"
8274 but the code is actually correct.
8277 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8278 Bleichenbacher's DSA attack.
8279 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8280 to be set and top=0 forces the highest bit to be set; top=-1 is new
8281 and leaves the highest bit random.
8282 [Ulf Moeller, Bodo Moeller]
8284 *) In the NCONF_...-based implementations for CONF_... queries
8285 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8286 a temporary CONF structure with the data component set to NULL
8287 (which gives segmentation faults in lh_retrieve).
8288 Instead, use NULL for the CONF pointer in CONF_get_string and
8289 CONF_get_number (which may use environment variables) and directly
8290 return NULL from CONF_get_section.
8293 *) Fix potential buffer overrun for EBCDIC.
8296 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8297 keyUsage if basicConstraints absent for a CA.
8300 *) Make SMIME_write_PKCS7() write mail header values with a format that
8301 is more generally accepted (no spaces before the semicolon), since
8302 some programs can't parse those values properly otherwise. Also make
8303 sure BIO's that break lines after each write do not create invalid
8307 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8308 macros previously used would not encode an empty SEQUENCE OF
8309 and break the signature.
8311 [This change does not apply to 0.9.7.]
8313 *) Zero the premaster secret after deriving the master secret in
8317 *) Add some EVP_add_digest_alias registrations (as found in
8318 OpenSSL_add_all_digests()) to SSL_library_init()
8319 aka OpenSSL_add_ssl_algorithms(). This provides improved
8320 compatibility with peers using X.509 certificates
8321 with unconventional AlgorithmIdentifier OIDs.
8324 *) Fix for Irix with NO_ASM.
8325 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8327 *) ./config script fixes.
8328 [Ulf Moeller, Richard Levitte]
8330 *) Fix 'openssl passwd -1'.
8333 *) Change PKCS12_key_gen_asc() so it can cope with non null
8334 terminated strings whose length is passed in the passlen
8335 parameter, for example from PEM callbacks. This was done
8336 by adding an extra length parameter to asc2uni().
8337 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8339 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8340 call failed, free the DSA structure.
8343 *) Fix to uni2asc() to cope with zero length Unicode strings.
8344 These are present in some PKCS#12 files.
8347 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8348 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8349 when writing a 32767 byte record.
8350 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8352 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8353 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8355 (RSA objects have a reference count access to which is protected
8356 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8357 so they are meant to be shared between threads.)
8358 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8359 "Reddie, Steven" <Steven.Reddie@ca.com>]
8361 *) Fix a deadlock in CRYPTO_mem_leaks().
8364 *) Use better test patterns in bntest.
8367 *) rand_win.c fix for Borland C.
8370 *) BN_rshift bugfix for n == 0.
8373 *) Add a 'bctest' script that checks for some known 'bc' bugs
8374 so that 'make test' does not abort just because 'bc' is broken.
8377 *) Store verify_result within SSL_SESSION also for client side to
8378 avoid potential security hole. (Re-used sessions on the client side
8379 always resulted in verify_result==X509_V_OK, not using the original
8380 result of the server certificate verification.)
8383 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8384 SSL3_RT_APPLICATION_DATA, return 0.
8385 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8389 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8390 releases, have been re-implemented by renaming the previous
8391 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8392 and ssl3_read_internal, respectively, and adding 'peek' parameters
8393 to them. The new ssl[23]_{read,peek} functions are calls to
8394 ssl[23]_read_internal with the 'peek' flag set appropriately.
8395 A 'peek' parameter has also been added to ssl3_read_bytes, which
8396 does the actual work for ssl3_read_internal.
8399 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8400 the method-specific "init()" handler. Also clean up ex_data after
8401 calling the method-specific "finish()" handler. Previously, this was
8402 happening the other way round.
8405 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8406 The previous value, 12, was not always sufficient for BN_mod_exp().
8409 *) Make sure that shared libraries get the internal name engine with
8410 the full version number and not just 0. This should mark the
8411 shared libraries as not backward compatible. Of course, this should
8412 be changed again when we can guarantee backward binary compatibility.
8415 *) Fix typo in get_cert_by_subject() in by_dir.c
8416 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8418 *) Rework the system to generate shared libraries:
8420 - Make note of the expected extension for the shared libraries and
8421 if there is a need for symbolic links from for example libcrypto.so.0
8422 to libcrypto.so.0.9.7. There is extended info in Configure for
8425 - Make as few rebuilds of the shared libraries as possible.
8427 - Still avoid linking the OpenSSL programs with the shared libraries.
8429 - When installing, install the shared libraries separately from the
8433 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8435 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8436 and not in SSL_clear because the latter is also used by the
8437 accept/connect functions; previously, the settings made by
8438 SSL_set_read_ahead would be lost during the handshake.
8439 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8441 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8442 Previously, it would create entries for disabled algorithms no
8446 *) Added several new manual pages for SSL_* function.
8449 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8451 *) In ssl23_get_client_hello, generate an error message when faced
8452 with an initial SSL 3.0/TLS record that is too small to contain the
8453 first two bytes of the ClientHello message, i.e. client_version.
8454 (Note that this is a pathologic case that probably has never happened
8455 in real life.) The previous approach was to use the version number
8456 from the record header as a substitute; but our protocol choice
8457 should not depend on that one because it is not authenticated
8458 by the Finished messages.
8461 *) More robust randomness gathering functions for Windows.
8462 [Jeffrey Altman <jaltman@columbia.edu>]
8464 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8465 not set then we don't setup the error code for issuer check errors
8466 to avoid possibly overwriting other errors which the callback does
8467 handle. If an application does set the flag then we assume it knows
8468 what it is doing and can handle the new informational codes
8472 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8473 a general "ANY" type, as such it should be able to decode anything
8474 including tagged types. However it didn't check the class so it would
8475 wrongly interpret tagged types in the same way as their universal
8476 counterpart and unknown types were just rejected. Changed so that the
8477 tagged and unknown types are handled in the same way as a SEQUENCE:
8478 that is the encoding is stored intact. There is also a new type
8479 "V_ASN1_OTHER" which is used when the class is not universal, in this
8480 case we have no idea what the actual type is so we just lump them all
8484 *) On VMS, stdout may very well lead to a file that is written to
8485 in a record-oriented fashion. That means that every write() will
8486 write a separate record, which will be read separately by the
8487 programs trying to read from it. This can be very confusing.
8489 The solution is to put a BIO filter in the way that will buffer
8490 text until a linefeed is reached, and then write everything a
8491 line at a time, so every record written will be an actual line,
8492 not chunks of lines and not (usually doesn't happen, but I've
8493 seen it once) several lines in one record. BIO_f_linebuffer() is
8496 Currently, it's a VMS-only method, because that's where it has
8497 been tested well enough.
8500 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8501 it can return incorrect results.
8502 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8503 but it was in 0.9.6-beta[12].)
8506 *) Disable the check for content being present when verifying detached
8507 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8508 include zero length content when signing messages.
8511 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8512 BIO_ctrl (for BIO pairs).
8515 *) Add DSO method for VMS.
8518 *) Bug fix: Montgomery multiplication could produce results with the
8522 *) Add RPM specification openssl.spec and modify it to build three
8523 packages. The default package contains applications, application
8524 documentation and run-time libraries. The devel package contains
8525 include files, static libraries and function documentation. The
8526 doc package contains the contents of the doc directory. The original
8527 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8530 *) Add a large number of documentation files for many SSL routines.
8531 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8533 *) Add a configuration entry for Sony News 4.
8534 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8536 *) Don't set the two most significant bits to one when generating a
8537 random number < q in the DSA library.
8540 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8541 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8542 the underlying transport is blocking) if a handshake took place.
8543 (The default behaviour is needed by applications such as s_client
8544 and s_server that use select() to determine when to use SSL_read;
8545 but for applications that know in advance when to expect data, it
8546 just makes things more complicated.)
8549 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8553 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8554 work better on such systems.
8555 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8557 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8558 Update PKCS12_parse() so it copies the friendlyName and the
8559 keyid to the certificates aux info.
8562 *) Fix bug in PKCS7_verify() which caused an infinite loop
8563 if there was more than one signature.
8564 [Sven Uszpelkat <su@celocom.de>]
8566 *) Major change in util/mkdef.pl to include extra information
8567 about each symbol, as well as presenting variables as well
8568 as functions. This change means that there's n more need
8569 to rebuild the .num files when some algorithms are excluded.
8572 *) Allow the verify time to be set by an application,
8573 rather than always using the current time.
8576 *) Phase 2 verify code reorganisation. The certificate
8577 verify code now looks up an issuer certificate by a
8578 number of criteria: subject name, authority key id
8579 and key usage. It also verifies self signed certificates
8580 by the same criteria. The main comparison function is
8581 X509_check_issued() which performs these checks.
8583 Lot of changes were necessary in order to support this
8584 without completely rewriting the lookup code.
8586 Authority and subject key identifier are now cached.
8588 The LHASH 'certs' is X509_STORE has now been replaced
8589 by a STACK_OF(X509_OBJECT). This is mainly because an
8590 LHASH can't store or retrieve multiple objects with
8591 the same hash value.
8593 As a result various functions (which were all internal
8594 use only) have changed to handle the new X509_STORE
8595 structure. This will break anything that messed round
8596 with X509_STORE internally.
8598 The functions X509_STORE_add_cert() now checks for an
8599 exact match, rather than just subject name.
8601 The X509_STORE API doesn't directly support the retrieval
8602 of multiple certificates matching a given criteria, however
8603 this can be worked round by performing a lookup first
8604 (which will fill the cache with candidate certificates)
8605 and then examining the cache for matches. This is probably
8606 the best we can do without throwing out X509_LOOKUP
8607 entirely (maybe later...).
8609 The X509_VERIFY_CTX structure has been enhanced considerably.
8611 All certificate lookup operations now go via a get_issuer()
8612 callback. Although this currently uses an X509_STORE it
8613 can be replaced by custom lookups. This is a simple way
8614 to bypass the X509_STORE hackery necessary to make this
8615 work and makes it possible to use more efficient techniques
8616 in future. A very simple version which uses a simple
8617 STACK for its trusted certificate store is also provided
8618 using X509_STORE_CTX_trusted_stack().
8620 The verify_cb() and verify() callbacks now have equivalents
8621 in the X509_STORE_CTX structure.
8623 X509_STORE_CTX also has a 'flags' field which can be used
8624 to customise the verify behaviour.
8627 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8628 excludes S/MIME capabilities.
8631 *) When a certificate request is read in keep a copy of the
8632 original encoding of the signed data and use it when outputing
8633 again. Signatures then use the original encoding rather than
8634 a decoded, encoded version which may cause problems if the
8635 request is improperly encoded.
8638 *) For consistency with other BIO_puts implementations, call
8639 buffer_write(b, ...) directly in buffer_puts instead of calling
8642 In BIO_puts, increment b->num_write as in BIO_write.
8643 [Peter.Sylvester@EdelWeb.fr]
8645 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8646 BN_zero, we may not return a BIGNUM with an array consisting of
8650 *) Avoid calling abort() from within the library when problems are
8651 detected, except if preprocessor symbols have been defined
8652 (such as REF_CHECK, BN_DEBUG etc.).
8655 *) New openssl application 'rsautl'. This utility can be
8656 used for low level RSA operations. DER public key
8657 BIO/fp routines also added.
8660 *) New Configure entry and patches for compiling on QNX 4.
8661 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8663 *) A demo state-machine implementation was sponsored by
8664 Nuron (http://www.nuron.com/) and is now available in
8665 demos/state_machine.
8668 *) New options added to the 'dgst' utility for signature
8669 generation and verification.
8672 *) Unrecognized PKCS#7 content types are now handled via a
8673 catch all ASN1_TYPE structure. This allows unsupported
8674 types to be stored as a "blob" and an application can
8675 encode and decode it manually.
8678 *) Fix various signed/unsigned issues to make a_strex.c
8680 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8682 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8683 length if passed a buffer. ASN1_INTEGER_to_BN failed
8684 if passed a NULL BN and its argument was negative.
8685 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8687 *) Modification to PKCS#7 encoding routines to output definite
8688 length encoding. Since currently the whole structures are in
8689 memory there's not real point in using indefinite length
8690 constructed encoding. However if OpenSSL is compiled with
8691 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8694 *) Added BIO_vprintf() and BIO_vsnprintf().
8697 *) Added more prefixes to parse for in the the strings written
8698 through a logging bio, to cover all the levels that are available
8699 through syslog. The prefixes are now:
8701 PANIC, EMERG, EMR => LOG_EMERG
8702 ALERT, ALR => LOG_ALERT
8703 CRIT, CRI => LOG_CRIT
8704 ERROR, ERR => LOG_ERR
8705 WARNING, WARN, WAR => LOG_WARNING
8706 NOTICE, NOTE, NOT => LOG_NOTICE
8707 INFO, INF => LOG_INFO
8708 DEBUG, DBG => LOG_DEBUG
8710 and as before, if none of those prefixes are present at the
8711 beginning of the string, LOG_ERR is chosen.
8713 On Win32, the LOG_* levels are mapped according to this:
8715 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8716 LOG_WARNING => EVENTLOG_WARNING_TYPE
8717 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8721 *) Made it possible to reconfigure with just the configuration
8722 argument "reconf" or "reconfigure". The command line arguments
8723 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8724 and are retrieved from there when reconfiguring.
8728 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8730 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8733 *) The obj_dat.pl script was messing up the sorting of object
8734 names. The reason was that it compared the quoted version
8735 of strings as a result "OCSP" > "OCSP Signing" because
8736 " > SPACE. Changed script to store unquoted versions of
8737 names and add quotes on output. It was also omitting some
8738 names from the lookup table if they were given a default
8739 value (that is if SN is missing it is given the same
8740 value as LN and vice versa), these are now added on the
8741 grounds that if an object has a name we should be able to
8742 look it up. Finally added warning output when duplicate
8743 short or long names are found.
8746 *) Changes needed for Tandem NSK.
8747 [Scott Uroff <scott@xypro.com>]
8749 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8750 RSA_padding_check_SSLv23(), special padding was never detected
8751 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8752 version rollback attacks was not effective.
8754 In s23_clnt.c, don't use special rollback-attack detection padding
8755 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8756 client; similarly, in s23_srvr.c, don't do the rollback check if
8757 SSL 2.0 is the only protocol enabled in the server.
8760 *) Make it possible to get hexdumps of unprintable data with 'openssl
8761 asn1parse'. By implication, the functions ASN1_parse_dump() and
8762 BIO_dump_indent() are added.
8765 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8766 these print out strings and name structures based on various
8767 flags including RFC2253 support and proper handling of
8768 multibyte characters. Added options to the 'x509' utility
8769 to allow the various flags to be set.
8772 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8773 Also change the functions X509_cmp_current_time() and
8774 X509_gmtime_adj() work with an ASN1_TIME structure,
8775 this will enable certificates using GeneralizedTime in validity
8776 dates to be checked.
8779 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8780 negative public key encodings) on by default,
8781 NO_NEG_PUBKEY_BUG can be set to disable it.
8784 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8785 content octets. An i2c_ASN1_OBJECT is unnecessary because
8786 the encoding can be trivially obtained from the structure.
8789 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8790 not read locks (CRYPTO_r_[un]lock).
8793 *) A first attempt at creating official support for shared
8794 libraries through configuration. I've kept it so the
8795 default is static libraries only, and the OpenSSL programs
8796 are always statically linked for now, but there are
8797 preparations for dynamic linking in place.
8798 This has been tested on Linux and Tru64.
8801 *) Randomness polling function for Win9x, as described in:
8802 Peter Gutmann, Software Generation of Practically Strong
8806 *) Fix so PRNG is seeded in req if using an already existing
8810 *) New options to smime application. -inform and -outform
8811 allow alternative formats for the S/MIME message including
8812 PEM and DER. The -content option allows the content to be
8813 specified separately. This should allow things like Netscape
8814 form signing output easier to verify.
8817 *) Fix the ASN1 encoding of tags using the 'long form'.
8820 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8821 STRING types. These convert content octets to and from the
8822 underlying type. The actual tag and length octets are
8823 already assumed to have been read in and checked. These
8824 are needed because all other string types have virtually
8825 identical handling apart from the tag. By having versions
8826 of the ASN1 functions that just operate on content octets
8827 IMPLICIT tagging can be handled properly. It also allows
8828 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8829 and ASN1_INTEGER are identical apart from the tag.
8832 *) Change the handling of OID objects as follows:
8834 - New object identifiers are inserted in objects.txt, following
8835 the syntax given in objects.README.
8836 - objects.pl is used to process obj_mac.num and create a new
8838 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8841 This is currently kind of a hack, and the perl code in objects.pl
8842 isn't very elegant, but it works as I intended. The simplest way
8843 to check that it worked correctly is to look in obj_dat.h and
8844 check the array nid_objs and make sure the objects haven't moved
8845 around (this is important!). Additions are OK, as well as
8846 consistent name changes.
8849 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8852 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8853 The given file adds to whatever has already been seeded into the
8854 random pool through the RANDFILE configuration file option or
8855 environment variable, or the default random state file.
8858 *) mkstack.pl now sorts each macro group into lexical order.
8859 Previously the output order depended on the order the files
8860 appeared in the directory, resulting in needless rewriting
8864 *) Patches to make OpenSSL compile under Win32 again. Mostly
8865 work arounds for the VC++ problem that it treats func() as
8866 func(void). Also stripped out the parts of mkdef.pl that
8867 added extra typesafe functions: these no longer exist.
8870 *) Reorganisation of the stack code. The macros are now all
8871 collected in safestack.h . Each macro is defined in terms of
8872 a "stack macro" of the form SKM_<name>(type, a, b). The
8873 DEBUG_SAFESTACK is now handled in terms of function casts,
8874 this has the advantage of retaining type safety without the
8875 use of additional functions. If DEBUG_SAFESTACK is not defined
8876 then the non typesafe macros are used instead. Also modified the
8877 mkstack.pl script to handle the new form. Needs testing to see
8878 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8879 the default if no major problems. Similar behaviour for ASN1_SET_OF
8880 and PKCS12_STACK_OF.
8883 *) When some versions of IIS use the 'NET' form of private key the
8884 key derivation algorithm is different. Normally MD5(password) is
8885 used as a 128 bit RC4 key. In the modified case
8886 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8887 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8888 as the old Netscape_RSA functions except they have an additional
8889 'sgckey' parameter which uses the modified algorithm. Also added
8890 an -sgckey command line option to the rsa utility. Thanks to
8891 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8892 algorithm to openssl-dev.
8895 *) The evp_local.h macros were using 'c.##kname' which resulted in
8896 invalid expansion on some systems (SCO 5.0.5 for example).
8897 Corrected to 'c.kname'.
8898 [Phillip Porch <root@theporch.com>]
8900 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8901 a STACK of email addresses from a certificate or request, these look
8902 in the subject name and the subject alternative name extensions and
8903 omit any duplicate addresses.
8906 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8907 This makes DSA verification about 2 % faster.
8910 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8911 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8912 plus overhead for 1024 bit moduli).
8913 This makes exponentiations about 0.5 % faster for 1024 bit
8914 exponents (as measured by "openssl speed rsa2048").
8917 *) Rename memory handling macros to avoid conflicts with other
8919 Malloc => OPENSSL_malloc
8920 Malloc_locked => OPENSSL_malloc_locked
8921 Realloc => OPENSSL_realloc
8922 Free => OPENSSL_free
8925 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8926 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8929 *) CygWin32 support.
8930 [John Jarvie <jjarvie@newsguy.com>]
8932 *) The type-safe stack code has been rejigged. It is now only compiled
8933 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8934 by default all type-specific stack functions are "#define"d back to
8935 standard stack functions. This results in more streamlined output
8936 but retains the type-safety checking possibilities of the original
8940 *) The STACK code has been cleaned up, and certain type declarations
8941 that didn't make a lot of sense have been brought in line. This has
8942 also involved a cleanup of sorts in safestack.h to more correctly
8943 map type-safe stack functions onto their plain stack counterparts.
8944 This work has also resulted in a variety of "const"ifications of
8945 lots of the code, especially "_cmp" operations which should normally
8946 be prototyped with "const" parameters anyway.
8949 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8950 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8951 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8952 where all of 'md' is used each time the PRNG is used, but 'state'
8953 is used only indexed by a cyclic counter. As entropy may not be
8954 well distributed from the beginning, 'md' is important as a
8955 chaining variable. However, the output function chains only half
8956 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8957 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8958 in all of 'state' being rewritten, with the new values depending
8959 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8962 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8963 the handshake is continued after ssl_verify_cert_chain();
8964 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8965 can lead to 'unexplainable' connection aborts later.
8966 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8968 *) Major EVP API cipher revision.
8969 Add hooks for extra EVP features. This allows various cipher
8970 parameters to be set in the EVP interface. Support added for variable
8971 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8972 setting of RC2 and RC5 parameters.
8974 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8977 Remove lots of duplicated code from the EVP library. For example *every*
8978 cipher init() function handles the 'iv' in the same way according to the
8979 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8980 for CFB and OFB modes they zero ctx->num.
8982 New functionality allows removal of S/MIME code RC2 hack.
8984 Most of the routines have the same form and so can be declared in terms
8987 By shifting this to the top level EVP_CipherInit() it can be removed from
8988 all individual ciphers. If the cipher wants to handle IVs or keys
8989 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8992 Change lots of functions like EVP_EncryptUpdate() to now return a
8993 value: although software versions of the algorithms cannot fail
8994 any installed hardware versions can.
8997 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8998 this option is set, tolerate broken clients that send the negotiated
8999 protocol version number instead of the requested protocol version
9003 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9004 i.e. non-zero for export ciphersuites, zero otherwise.
9005 Previous versions had this flag inverted, inconsistent with
9006 rsa_tmp_cb (..._TMP_RSA_CB).
9007 [Bodo Moeller; problem reported by Amit Chopra]
9009 *) Add missing DSA library text string. Work around for some IIS
9010 key files with invalid SEQUENCE encoding.
9013 *) Add a document (doc/standards.txt) that list all kinds of standards
9014 and so on that are implemented in OpenSSL.
9017 *) Enhance c_rehash script. Old version would mishandle certificates
9018 with the same subject name hash and wouldn't handle CRLs at all.
9019 Added -fingerprint option to crl utility, to support new c_rehash
9023 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9026 *) Fix for SSL server purpose checking. Server checking was
9027 rejecting certificates which had extended key usage present
9028 but no ssl client purpose.
9029 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9031 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9032 is a little unclear about how a blank password is handled.
9033 Since the password in encoded as a BMPString with terminating
9034 double NULL a zero length password would end up as just the
9035 double NULL. However no password at all is different and is
9036 handled differently in the PKCS#12 key generation code. NS
9037 treats a blank password as zero length. MSIE treats it as no
9038 password on export: but it will try both on import. We now do
9039 the same: PKCS12_parse() tries zero length and no password if
9040 the password is set to "" or NULL (NULL is now a valid password:
9041 it wasn't before) as does the pkcs12 application.
9044 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9045 perror when PEM_read_bio_X509_REQ fails, the error message must
9046 be obtained from the error queue.
9049 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9050 it in ERR_remove_state if appropriate, and change ERR_get_state
9051 accordingly to avoid race conditions (this is necessary because
9052 thread_hash is no longer constant once set).
9055 *) Bugfix for linux-elf makefile.one.
9058 *) RSA_get_default_method() will now cause a default
9059 RSA_METHOD to be chosen if one doesn't exist already.
9060 Previously this was only set during a call to RSA_new()
9061 or RSA_new_method(NULL) meaning it was possible for
9062 RSA_get_default_method() to return NULL.
9065 *) Added native name translation to the existing DSO code
9066 that will convert (if the flag to do so is set) filenames
9067 that are sufficiently small and have no path information
9068 into a canonical native form. Eg. "blah" converted to
9069 "libblah.so" or "blah.dll" etc.
9072 *) New function ERR_error_string_n(e, buf, len) which is like
9073 ERR_error_string(e, buf), but writes at most 'len' bytes
9074 including the 0 terminator. For ERR_error_string_n, 'buf'
9076 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9078 *) CONF library reworked to become more general. A new CONF
9079 configuration file reader "class" is implemented as well as a
9080 new functions (NCONF_*, for "New CONF") to handle it. The now
9081 old CONF_* functions are still there, but are reimplemented to
9082 work in terms of the new functions. Also, a set of functions
9083 to handle the internal storage of the configuration data is
9084 provided to make it easier to write new configuration file
9085 reader "classes" (I can definitely see something reading a
9086 configuration file in XML format, for example), called _CONF_*,
9087 or "the configuration storage API"...
9089 The new configuration file reading functions are:
9091 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9092 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9094 NCONF_default, NCONF_WIN32
9096 NCONF_dump_fp, NCONF_dump_bio
9098 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9099 NCONF_new creates a new CONF object. This works in the same way
9100 as other interfaces in OpenSSL, like the BIO interface.
9101 NCONF_dump_* dump the internal storage of the configuration file,
9102 which is useful for debugging. All other functions take the same
9103 arguments as the old CONF_* functions wth the exception of the
9104 first that must be a `CONF *' instead of a `LHASH *'.
9106 To make it easer to use the new classes with the old CONF_* functions,
9107 the function CONF_set_default_method is provided.
9110 *) Add '-tls1' option to 'openssl ciphers', which was already
9111 mentioned in the documentation but had not been implemented.
9112 (This option is not yet really useful because even the additional
9113 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9116 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9117 OpenSSL-based applications) load shared libraries and bind to
9118 them in a portable way.
9119 [Geoff Thorpe, with contributions from Richard Levitte]
9121 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9123 *) Make sure _lrotl and _lrotr are only used with MSVC.
9125 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9126 (the default implementation of RAND_status).
9128 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9129 to '-clrext' (= clear extensions), as intended and documented.
9130 [Bodo Moeller; inconsistency pointed out by Michael Attili
9133 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9134 was larger than the MD block size.
9135 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9137 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9138 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9139 using the passed key: if the passed key was a private key the result
9140 of X509_print(), for example, would be to print out all the private key
9144 *) des_quad_cksum() byte order bug fix.
9145 [Ulf Möller, using the problem description in krb4-0.9.7, where
9146 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9148 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9150 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9152 *) For easily testing in shell scripts whether some command
9153 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9154 returns with exit code 0 iff no command of the given name is available.
9155 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9156 the output goes to stdout and nothing is printed to stderr.
9157 Additional arguments are always ignored.
9159 Since for each cipher there is a command of the same name,
9160 the 'no-cipher' compilation switches can be tested this way.
9162 ('openssl no-XXX' is not able to detect pseudo-commands such
9163 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9166 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9169 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9170 is set; it will be thrown away anyway because each handshake creates
9172 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9173 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9174 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9175 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9178 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9179 'Q' and 'R' lose their special meanings (quit/renegotiate).
9180 This is part of what -quiet does; unlike -quiet, -ign_eof
9181 does not suppress any output.
9184 *) Add compatibility options to the purpose and trust code. The
9185 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9186 accepts a certificate or CA, this was the previous behaviour,
9187 with all the associated security issues.
9189 X509_TRUST_COMPAT is the old trust behaviour: only and
9190 automatically trust self signed roots in certificate store. A
9191 new trust setting X509_TRUST_DEFAULT is used to specify that
9192 a purpose has no associated trust setting and it should instead
9193 use the value in the default purpose.
9196 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9197 and fix a memory leak.
9200 *) In util/mkerr.pl (which implements 'make errors'), preserve
9201 reason strings from the previous version of the .c file, as
9202 the default to have only downcase letters (and digits) in
9203 automatically generated reasons codes is not always appropriate.
9206 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9207 using strerror. Previously, ERR_reason_error_string() returned
9208 library names as reason strings for SYSerr; but SYSerr is a special
9209 case where small numbers are errno values, not library numbers.
9212 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9213 converts DSA parameters into DH parameters. (When creating parameters,
9214 DSA_generate_parameters is used.)
9217 *) Include 'length' (recommended exponent length) in C code generated
9218 by 'openssl dhparam -C'.
9221 *) The second argument to set_label in perlasm was already being used
9222 so couldn't be used as a "file scope" flag. Moved to third argument
9226 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9227 instead of RAND_bytes for encryption IVs and salts.
9230 *) Include RAND_status() into RAND_METHOD instead of implementing
9231 it only for md_rand.c Otherwise replacing the PRNG by calling
9232 RAND_set_rand_method would be impossible.
9235 *) Don't let DSA_generate_key() enter an infinite loop if the random
9236 number generation fails.
9239 *) New 'rand' application for creating pseudo-random output.
9242 *) Added configuration support for Linux/IA64
9243 [Rolf Haberrecker <rolf@suse.de>]
9245 *) Assembler module support for Mingw32.
9248 *) Shared library support for HPUX (in shlib/).
9249 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9251 *) Shared library support for Solaris gcc.
9252 [Lutz Behnke <behnke@trustcenter.de>]
9254 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9256 *) PKCS7_encrypt() was adding text MIME headers twice because they
9257 were added manually and by SMIME_crlf_copy().
9260 *) In bntest.c don't call BN_rand with zero bits argument.
9261 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9263 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9264 case was implemented. This caused BN_div_recp() to fail occasionally.
9267 *) Add an optional second argument to the set_label() in the perl
9268 assembly language builder. If this argument exists and is set
9269 to 1 it signals that the assembler should use a symbol whose
9270 scope is the entire file, not just the current function. This
9271 is needed with MASM which uses the format label:: for this scope.
9272 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9274 *) Change the ASN1 types so they are typedefs by default. Before
9275 almost all types were #define'd to ASN1_STRING which was causing
9276 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9280 *) Change names of new functions to the new get1/get0 naming
9281 convention: After 'get1', the caller owns a reference count
9282 and has to call ..._free; 'get0' returns a pointer to some
9283 data structure without incrementing reference counters.
9284 (Some of the existing 'get' functions increment a reference
9285 counter, some don't.)
9286 Similarly, 'set1' and 'add1' functions increase reference
9287 counters or duplicate objects.
9290 *) Allow for the possibility of temp RSA key generation failure:
9291 the code used to assume it always worked and crashed on failure.
9294 *) Fix potential buffer overrun problem in BIO_printf().
9295 [Ulf Möller, using public domain code by Patrick Powell; problem
9296 pointed out by David Sacerdote <das33@cornell.edu>]
9298 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9299 RAND_egd() and RAND_status(). In the command line application,
9300 the EGD socket can be specified like a seed file using RANDFILE
9304 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9305 Some CAs (e.g. Verisign) distribute certificates in this form.
9308 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9309 list to exclude them. This means that no special compilation option
9310 is needed to use anonymous DH: it just needs to be included in the
9314 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9315 EVP_MD_type. The old functionality is available in a new macro called
9316 EVP_MD_md(). Change code that uses it and update docs.
9319 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9320 where the 'void *' argument is replaced by a function pointer argument.
9321 Previously 'void *' was abused to point to functions, which works on
9322 many platforms, but is not correct. As these functions are usually
9323 called by macros defined in OpenSSL header files, most source code
9324 should work without changes.
9327 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9328 sections with information on -D... compiler switches used for
9329 compiling the library so that applications can see them. To enable
9330 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9331 must be defined. E.g.,
9332 #define OPENSSL_ALGORITHM_DEFINES
9333 #include <openssl/opensslconf.h>
9334 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9335 [Richard Levitte, Ulf and Bodo Möller]
9337 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9341 *) Change the 'other' type in certificate aux info to a STACK_OF
9342 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9343 the required ASN1 format: arbitrary types determined by an OID.
9346 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9347 argument to 'req'. This is not because the function is newer or
9348 better than others it just uses the work 'NEW' in the certificate
9349 request header lines. Some software needs this.
9352 *) Reorganise password command line arguments: now passwords can be
9353 obtained from various sources. Delete the PEM_cb function and make
9354 it the default behaviour: i.e. if the callback is NULL and the
9355 usrdata argument is not NULL interpret it as a null terminated pass
9356 phrase. If usrdata and the callback are NULL then the pass phrase
9357 is prompted for as usual.
9360 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9361 the support is automatically enabled. The resulting binaries will
9362 autodetect the card and use it if present.
9363 [Ben Laurie and Compaq Inc.]
9365 *) Work around for Netscape hang bug. This sends certificate request
9366 and server done in one record. Since this is perfectly legal in the
9367 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9368 the bugs/SSLv3 entry for more info.
9371 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9374 *) Add -rand argument to smime and pkcs12 applications and read/write
9378 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9381 *) Add command line password options to the remaining applications.
9384 *) Bug fix for BN_div_recp() for numerators with an even number of
9388 *) More tests in bntest.c, and changed test_bn output.
9391 *) ./config recognizes MacOS X now.
9394 *) Bug fix for BN_div() when the first words of num and divsor are
9395 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9398 *) Add support for various broken PKCS#8 formats, and command line
9399 options to produce them.
9402 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9403 get temporary BIGNUMs from a BN_CTX.
9406 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9410 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9411 include a #define from the old name to the new. The original intent
9412 was that statically linked binaries could for example just call
9413 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9414 link with digests. This never worked because SSLeay_add_all_digests()
9415 and SSLeay_add_all_ciphers() were in the same source file so calling
9416 one would link with the other. They are now in separate source files.
9419 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9422 *) Use a less unusual form of the Miller-Rabin primality test (it used
9423 a binary algorithm for exponentiation integrated into the Miller-Rabin
9424 loop, our standard modexp algorithms are faster).
9427 *) Support for the EBCDIC character set completed.
9428 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9430 *) Source code cleanups: use const where appropriate, eliminate casts,
9431 use void * instead of char * in lhash.
9434 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9435 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9436 this the server could overwrite ephemeral keys that the client
9440 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9441 using 50 iterations of the Rabin-Miller test.
9443 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9444 iterations of the Rabin-Miller test as required by the appendix
9445 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9446 As BN_is_prime_fasttest includes trial division, DSA parameter
9447 generation becomes much faster.
9449 This implies a change for the callback functions in DSA_is_prime
9450 and DSA_generate_parameters: The callback function is called once
9451 for each positive witness in the Rabin-Miller test, not just
9452 occasionally in the inner loop; and the parameters to the
9453 callback function now provide an iteration count for the outer
9454 loop rather than for the current invocation of the inner loop.
9455 DSA_generate_parameters additionally can call the callback
9456 function with an 'iteration count' of -1, meaning that a
9457 candidate has passed the trial division test (when q is generated
9458 from an application-provided seed, trial division is skipped).
9461 *) New function BN_is_prime_fasttest that optionally does trial
9462 division before starting the Rabin-Miller test and has
9463 an additional BN_CTX * argument (whereas BN_is_prime always
9464 has to allocate at least one BN_CTX).
9465 'callback(1, -1, cb_arg)' is called when a number has passed the
9466 trial division stage.
9469 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9473 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9476 *) New function BN_pseudo_rand().
9479 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9480 bignum version of BN_from_montgomery() with the working code from
9481 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9485 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9486 made it impossible to use the same SSL_SESSION data structure in
9487 SSL2 clients in multiple threads.
9490 *) The return value of RAND_load_file() no longer counts bytes obtained
9491 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9492 to seed the PRNG (previously an explicit byte count was required).
9493 [Ulf Möller, Bodo Möller]
9495 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9496 used (char *) instead of (void *) and had casts all over the place.
9499 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9502 *) Retain source code compatibility for BN_prime_checks macro:
9503 BN_is_prime(..., BN_prime_checks, ...) now uses
9504 BN_prime_checks_for_size to determine the appropriate number of
9505 Rabin-Miller iterations.
9508 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9509 DH_CHECK_P_NOT_SAFE_PRIME.
9510 (Check if this is true? OpenPGP calls them "strong".)
9513 *) Merge the functionality of "dh" and "gendh" programs into a new program
9514 "dhparam". The old programs are retained for now but will handle DH keys
9515 (instead of parameters) in future.
9518 *) Make the ciphers, s_server and s_client programs check the return values
9519 when a new cipher list is set.
9522 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9523 ciphers. Before when the 56bit ciphers were enabled the sorting was
9526 The syntax for the cipher sorting has been extended to support sorting by
9527 cipher-strength (using the strength_bits hard coded in the tables).
9528 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9530 Fix a bug in the cipher-command parser: when supplying a cipher command
9531 string with an "undefined" symbol (neither command nor alphanumeric
9532 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9533 an error is flagged.
9535 Due to the strength-sorting extension, the code of the
9536 ssl_create_cipher_list() function was completely rearranged. I hope that
9537 the readability was also increased :-)
9538 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9540 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9541 for the first serial number and places 2 in the serial number file. This
9542 avoids problems when the root CA is created with serial number zero and
9543 the first user certificate has the same issuer name and serial number
9547 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9548 the new code. Add documentation for this stuff.
9551 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9552 X509_*() to X509at_*() on the grounds that they don't handle X509
9553 structures and behave in an analogous way to the X509v3 functions:
9554 they shouldn't be called directly but wrapper functions should be used
9557 So we also now have some wrapper functions that call the X509at functions
9558 when passed certificate requests. (TO DO: similar things can be done with
9559 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9560 things. Some of these need some d2i or i2d and print functionality
9561 because they handle more complex structures.)
9564 *) Add missing #ifndefs that caused missing symbols when building libssl
9565 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9566 NO_RSA in ssl/s2*.c.
9567 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9569 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9570 has a return value which indicates the quality of the random data
9571 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9572 error queue. New function RAND_pseudo_bytes() generates output that is
9573 guaranteed to be unique but not unpredictable. RAND_add is like
9574 RAND_seed, but takes an extra argument for an entropy estimate
9575 (RAND_seed always assumes full entropy).
9578 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9579 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9580 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9581 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9582 false-positive rate of at most 2^-80 for random input.
9585 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9588 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9589 in the 0.9.5 release), this returns the chain
9590 from an X509_CTX structure with a dup of the stack and all
9591 the X509 reference counts upped: so the stack will exist
9592 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9595 Also make SSL_SESSION_print() print out the verify return
9599 *) Add manpage for the pkcs12 command. Also change the default
9600 behaviour so MAC iteration counts are used unless the new
9601 -nomaciter option is used. This improves file security and
9602 only older versions of MSIE (4.0 for example) need it.
9605 *) Honor the no-xxx Configure options when creating .DEF files.
9608 *) Add PKCS#10 attributes to field table: challengePassword,
9609 unstructuredName and unstructuredAddress. These are taken from
9610 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9611 international characters are used.
9613 More changes to X509_ATTRIBUTE code: allow the setting of types
9614 based on strings. Remove the 'loc' parameter when adding
9615 attributes because these will be a SET OF encoding which is sorted
9619 *) Initial changes to the 'req' utility to allow request generation
9620 automation. This will allow an application to just generate a template
9621 file containing all the field values and have req construct the
9624 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9625 used all over the place including certificate requests and PKCS#7
9626 structures. They are currently handled manually where necessary with
9627 some primitive wrappers for PKCS#7. The new functions behave in a
9628 manner analogous to the X509 extension functions: they allow
9629 attributes to be looked up by NID and added.
9631 Later something similar to the X509V3 code would be desirable to
9632 automatically handle the encoding, decoding and printing of the
9633 more complex types. The string types like challengePassword can
9634 be handled by the string table functions.
9636 Also modified the multi byte string table handling. Now there is
9637 a 'global mask' which masks out certain types. The table itself
9638 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9639 is useful when for example there is only one permissible type
9640 (as in countryName) and using the mask might result in no valid
9644 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9645 SSL_get_peer_finished to allow applications to obtain the latest
9646 Finished messages sent to the peer or expected from the peer,
9647 respectively. (SSL_get_peer_finished is usually the Finished message
9648 actually received from the peer, otherwise the protocol will be aborted.)
9650 As the Finished message are message digests of the complete handshake
9651 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9652 be used for external authentication procedures when the authentication
9653 provided by SSL/TLS is not desired or is not enough.
9656 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9657 the host supports BWX extension and if Compaq C is present on the
9658 $PATH. Just exploiting of the BWX extension results in 20-30%
9659 performance kick for some algorithms, e.g. DES and RC4 to mention
9660 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9664 *) Add support for MS "fast SGC". This is arguably a violation of the
9665 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9666 weak crypto and after checking the certificate is SGC a second one
9667 with strong crypto. MS SGC stops the first handshake after receiving
9668 the server certificate message and sends a second client hello. Since
9669 a server will typically do all the time consuming operations before
9670 expecting any further messages from the client (server key exchange
9671 is the most expensive) there is little difference between the two.
9673 To get OpenSSL to support MS SGC we have to permit a second client
9674 hello message after we have sent server done. In addition we have to
9675 reset the MAC if we do get this second client hello.
9678 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9679 if a DER encoded private key is RSA or DSA traditional format. Changed
9680 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9681 format DER encoded private key. Newer code should use PKCS#8 format which
9682 has the key type encoded in the ASN1 structure. Added DER private key
9683 support to pkcs8 application.
9686 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9687 ciphersuites has been selected (as required by the SSL 3/TLS 1
9688 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9689 is set, we interpret this as a request to violate the specification
9690 (the worst that can happen is a handshake failure, and 'correct'
9691 behaviour would result in a handshake failure anyway).
9694 *) In SSL_CTX_add_session, take into account that there might be multiple
9695 SSL_SESSION structures with the same session ID (e.g. when two threads
9696 concurrently obtain them from an external cache).
9697 The internal cache can handle only one SSL_SESSION with a given ID,
9698 so if there's a conflict, we now throw out the old one to achieve
9702 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9703 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9704 some routines that use cipher OIDs: some ciphers do not have OIDs
9705 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9709 *) Simplify the trust setting structure and code. Now we just have
9710 two sequences of OIDs for trusted and rejected settings. These will
9711 typically have values the same as the extended key usage extension
9712 and any application specific purposes.
9714 The trust checking code now has a default behaviour: it will just
9715 check for an object with the same NID as the passed id. Functions can
9716 be provided to override either the default behaviour or the behaviour
9717 for a given id. SSL client, server and email already have functions
9718 in place for compatibility: they check the NID and also return "trusted"
9719 if the certificate is self signed.
9722 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9723 traditional format into an EVP_PKEY structure.
9726 *) Add a password callback function PEM_cb() which either prompts for
9727 a password if usr_data is NULL or otherwise assumes it is a null
9728 terminated password. Allow passwords to be passed on command line
9729 environment or config files in a few more utilities.
9732 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9733 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9734 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9735 Update documentation.
9738 *) Support for ASN1 "NULL" type. This could be handled before by using
9739 ASN1_TYPE but there wasn't any function that would try to read a NULL
9740 and produce an error if it couldn't. For compatibility we also have
9741 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9742 don't allocate anything because they don't need to.
9745 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9747 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9749 *) Rebuild of the memory allocation routines used by OpenSSL code and
9750 possibly others as well. The purpose is to make an interface that
9751 provide hooks so anyone can build a separate set of allocation and
9752 deallocation routines to be used by OpenSSL, for example memory
9753 pool implementations, or something else, which was previously hard
9754 since Malloc(), Realloc() and Free() were defined as macros having
9755 the values malloc, realloc and free, respectively (except for Win32
9756 compilations). The same is provided for memory debugging code.
9757 OpenSSL already comes with functionality to find memory leaks, but
9758 this gives people a chance to debug other memory problems.
9760 With these changes, a new set of functions and macros have appeared:
9762 CRYPTO_set_mem_debug_functions() [F]
9763 CRYPTO_get_mem_debug_functions() [F]
9764 CRYPTO_dbg_set_options() [F]
9765 CRYPTO_dbg_get_options() [F]
9766 CRYPTO_malloc_debug_init() [M]
9768 The memory debug functions are NULL by default, unless the library
9769 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9770 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9771 gives the standard debugging functions that come with OpenSSL) or
9772 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9773 provided by the library user) must be used. When the standard
9774 debugging functions are used, CRYPTO_dbg_set_options can be used to
9775 request additional information:
9776 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9777 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9779 Also, things like CRYPTO_set_mem_functions will always give the
9780 expected result (the new set of functions is used for allocation
9781 and deallocation) at all times, regardless of platform and compiler
9784 To finish it up, some functions that were never use in any other
9785 way than through macros have a new API and new semantic:
9788 CRYPTO_dbg_realloc()
9791 All macros of value have retained their old syntax.
9792 [Richard Levitte and Bodo Moeller]
9794 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9795 ordering of SMIMECapabilities wasn't in "strength order" and there
9796 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9800 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9801 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9802 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9804 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9805 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9806 functionality to handle multipart/signed properly) and a utility
9807 called 'smime' to call all this stuff. This is based on code I
9808 originally wrote for Celo who have kindly allowed it to be
9809 included in OpenSSL.
9812 *) Add variants des_set_key_checked and des_set_key_unchecked of
9813 des_set_key (aka des_key_sched). Global variable des_check_key
9814 decides which of these is called by des_set_key; this way
9815 des_check_key behaves as it always did, but applications and
9816 the library itself, which was buggy for des_check_key == 1,
9817 have a cleaner way to pick the version they need.
9820 *) New function PKCS12_newpass() which changes the password of a
9824 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9825 dynamic mix. In both cases the ids can be used as an index into the
9826 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9827 functions so they accept a list of the field values and the
9828 application doesn't need to directly manipulate the X509_TRUST
9832 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9836 *) Modify the way the V3 extension code looks up extensions. This now
9837 works in a similar way to the object code: we have some "standard"
9838 extensions in a static table which is searched with OBJ_bsearch()
9839 and the application can add dynamic ones if needed. The file
9840 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9841 updated whenever a new extension is added to the core code and kept
9842 in ext_nid order. There is a simple program 'tabtest.c' which checks
9843 this. New extensions are not added too often so this file can readily
9844 be maintained manually.
9846 There are two big advantages in doing things this way. The extensions
9847 can be looked up immediately and no longer need to be "added" using
9848 X509V3_add_standard_extensions(): this function now does nothing.
9849 [Side note: I get *lots* of email saying the extension code doesn't
9850 work because people forget to call this function]
9851 Also no dynamic allocation is done unless new extensions are added:
9852 so if we don't add custom extensions there is no need to call
9853 X509V3_EXT_cleanup().
9856 *) Modify enc utility's salting as follows: make salting the default. Add a
9857 magic header, so unsalted files fail gracefully instead of just decrypting
9858 to garbage. This is because not salting is a big security hole, so people
9859 should be discouraged from doing it.
9862 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9863 digest to be passed on the command line but it only used this
9864 parameter when signing a certificate. Modified so all relevant
9865 operations are affected by the digest parameter including the
9866 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9867 DSA key was used because it didn't fix the digest.
9870 *) Initial certificate chain verify code. Currently tests the untrusted
9871 certificates for consistency with the verify purpose (which is set
9872 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9874 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9875 this is because it will reject chains with invalid extensions whereas
9876 every previous version of OpenSSL and SSLeay made no checks at all.
9878 Trust code: checks the root CA for the relevant trust settings. Trust
9879 settings have an initial value consistent with the verify purpose: e.g.
9880 if the verify purpose is for SSL client use it expects the CA to be
9881 trusted for SSL client use. However the default value can be changed to
9882 permit custom trust settings: one example of this would be to only trust
9883 certificates from a specific "secure" set of CAs.
9885 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9886 which should be used for version portability: especially since the
9887 verify structure is likely to change more often now.
9889 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9890 to set them. If not set then assume SSL clients will verify SSL servers
9893 Two new options to the verify program: -untrusted allows a set of
9894 untrusted certificates to be passed in and -purpose which sets the
9895 intended purpose of the certificate. If a purpose is set then the
9896 new chain verify code is used to check extension consistency.
9899 *) Support for the authority information access extension.
9902 *) Modify RSA and DSA PEM read routines to transparently handle
9903 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9904 public keys in a format compatible with certificate
9905 SubjectPublicKeyInfo structures. Unfortunately there were already
9906 functions called *_PublicKey_* which used various odd formats so
9907 these are retained for compatibility: however the DSA variants were
9908 never in a public release so they have been deleted. Changed dsa/rsa
9909 utilities to handle the new format: note no releases ever handled public
9910 keys so we should be OK.
9912 The primary motivation for this change is to avoid the same fiasco
9913 that dogs private keys: there are several incompatible private key
9914 formats some of which are standard and some OpenSSL specific and
9915 require various evil hacks to allow partial transparent handling and
9916 even then it doesn't work with DER formats. Given the option anything
9917 other than PKCS#8 should be dumped: but the other formats have to
9918 stay in the name of compatibility.
9920 With public keys and the benefit of hindsight one standard format
9921 is used which works with EVP_PKEY, RSA or DSA structures: though
9922 it clearly returns an error if you try to read the wrong kind of key.
9924 Added a -pubkey option to the 'x509' utility to output the public key.
9925 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9926 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9927 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9928 that do the same as the EVP_PKEY_assign_*() except they up the
9929 reference count of the added key (they don't "swallow" the
9933 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9934 CRLs would fail if the file contained no certificates or no CRLs:
9935 added a new function to read in both types and return the number
9936 read: this means that if none are read it will be an error. The
9937 DER versions of the certificate and CRL reader would always fail
9938 because it isn't possible to mix certificates and CRLs in DER format
9939 without choking one or the other routine. Changed this to just read
9940 a certificate: this is the best we can do. Also modified the code
9941 in apps/verify.c to take notice of return codes: it was previously
9942 attempting to read in certificates from NULL pointers and ignoring
9943 any errors: this is one reason why the cert and CRL reader seemed
9944 to work. It doesn't check return codes from the default certificate
9945 routines: these may well fail if the certificates aren't installed.
9948 *) Code to support otherName option in GeneralName.
9951 *) First update to verify code. Change the verify utility
9952 so it warns if it is passed a self signed certificate:
9953 for consistency with the normal behaviour. X509_verify
9954 has been modified to it will now verify a self signed
9955 certificate if *exactly* the same certificate appears
9956 in the store: it was previously impossible to trust a
9957 single self signed certificate. This means that:
9958 openssl verify ss.pem
9959 now gives a warning about a self signed certificate but
9960 openssl verify -CAfile ss.pem ss.pem
9964 *) For servers, store verify_result in SSL_SESSION data structure
9965 (and add it to external session representation).
9966 This is needed when client certificate verifications fails,
9967 but an application-provided verification callback (set by
9968 SSL_CTX_set_cert_verify_callback) allows accepting the session
9969 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9970 but returns 1): When the session is reused, we have to set
9971 ssl->verify_result to the appropriate error code to avoid
9973 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9975 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9976 case in PKCS7_dataInit() where the signed PKCS7 structure
9977 didn't contain any existing data because it was being created.
9978 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9980 *) Add a salt to the key derivation routines in enc.c. This
9981 forms the first 8 bytes of the encrypted file. Also add a
9982 -S option to allow a salt to be input on the command line.
9985 *) New function X509_cmp(). Oddly enough there wasn't a function
9986 to compare two certificates. We do this by working out the SHA1
9987 hash and comparing that. X509_cmp() will be needed by the trust
9991 *) SSL_get1_session() is like SSL_get_session(), but increments
9992 the reference count in the SSL_SESSION returned.
9993 [Geoff Thorpe <geoff@eu.c2.net>]
9995 *) Fix for 'req': it was adding a null to request attributes.
9996 Also change the X509_LOOKUP and X509_INFO code to handle
9997 certificate auxiliary information.
10000 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10004 *) Add the possibility to add extra information to the memory leak
10005 detecting output, to form tracebacks, showing from where each
10006 allocation was originated: CRYPTO_push_info("constant string") adds
10007 the string plus current file name and line number to a per-thread
10008 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10009 is like calling CYRPTO_pop_info() until the stack is empty.
10010 Also updated memory leak detection code to be multi-thread-safe.
10013 *) Add options -text and -noout to pkcs7 utility and delete the
10014 encryption options which never did anything. Update docs.
10017 *) Add options to some of the utilities to allow the pass phrase
10018 to be included on either the command line (not recommended on
10019 OSes like Unix) or read from the environment. Update the
10020 manpages and fix a few bugs.
10023 *) Add a few manpages for some of the openssl commands.
10026 *) Fix the -revoke option in ca. It was freeing up memory twice,
10027 leaking and not finding already revoked certificates.
10030 *) Extensive changes to support certificate auxiliary information.
10031 This involves the use of X509_CERT_AUX structure and X509_AUX
10032 functions. An X509_AUX function such as PEM_read_X509_AUX()
10033 can still read in a certificate file in the usual way but it
10034 will also read in any additional "auxiliary information". By
10035 doing things this way a fair degree of compatibility can be
10036 retained: existing certificates can have this information added
10037 using the new 'x509' options.
10039 Current auxiliary information includes an "alias" and some trust
10040 settings. The trust settings will ultimately be used in enhanced
10041 certificate chain verification routines: currently a certificate
10042 can only be trusted if it is self signed and then it is trusted
10046 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10047 The problem was that one of the replacement routines had not been working
10048 since SSLeay releases. For now the offending routine has been replaced
10049 with non-optimised assembler. Even so, this now gives around 95%
10050 performance improvement for 1024 bit RSA signs.
10053 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10054 handling. Most clients have the effective key size in bits equal to
10055 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10056 A few however don't do this and instead use the size of the decrypted key
10057 to determine the RC2 key length and the AlgorithmIdentifier to determine
10058 the effective key length. In this case the effective key length can still
10059 be 40 bits but the key length can be 168 bits for example. This is fixed
10060 by manually forcing an RC2 key into the EVP_PKEY structure because the
10061 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10062 the key length and effective key length are equal.
10065 *) Add a bunch of functions that should simplify the creation of
10066 X509_NAME structures. Now you should be able to do:
10067 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10068 and have it automatically work out the correct field type and fill in
10069 the structures. The more adventurous can try:
10070 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10071 and it will (hopefully) work out the correct multibyte encoding.
10074 *) Change the 'req' utility to use the new field handling and multibyte
10075 copy routines. Before the DN field creation was handled in an ad hoc
10076 way in req, ca, and x509 which was rather broken and didn't support
10077 BMPStrings or UTF8Strings. Since some software doesn't implement
10078 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10079 using the dirstring_type option. See the new comment in the default
10080 openssl.cnf for more info.
10083 *) Make crypto/rand/md_rand.c more robust:
10084 - Assure unique random numbers after fork().
10085 - Make sure that concurrent threads access the global counter and
10086 md serializably so that we never lose entropy in them
10087 or use exactly the same state in multiple threads.
10088 Access to the large state is not always serializable because
10089 the additional locking could be a performance killer, and
10090 md should be large enough anyway.
10093 *) New file apps/app_rand.c with commonly needed functionality
10094 for handling the random seed file.
10096 Use the random seed file in some applications that previously did not:
10098 dsaparam -genkey (which also ignored its '-rand' option),
10101 x509 (when signing).
10102 Except on systems with /dev/urandom, it is crucial to have a random
10103 seed file at least for key creation, DSA signing, and for DH exchanges;
10104 for RSA signatures we could do without one.
10106 gendh and gendsa (unlike genrsa) used to read only the first byte
10107 of each file listed in the '-rand' option. The function as previously
10108 found in genrsa is now in app_rand.c and is used by all programs
10109 that support '-rand'.
10112 *) In RAND_write_file, use mode 0600 for creating files;
10113 don't just chmod when it may be too late.
10116 *) Report an error from X509_STORE_load_locations
10117 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10120 *) New function ASN1_mbstring_copy() this copies a string in either
10121 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10122 into an ASN1_STRING type. A mask of permissible types is passed
10123 and it chooses the "minimal" type to use or an error if not type
10127 *) Add function equivalents to the various macros in asn1.h. The old
10128 macros are retained with an M_ prefix. Code inside the library can
10129 use the M_ macros. External code (including the openssl utility)
10130 should *NOT* in order to be "shared library friendly".
10133 *) Add various functions that can check a certificate's extensions
10134 to see if it usable for various purposes such as SSL client,
10135 server or S/MIME and CAs of these types. This is currently
10136 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10137 verification. Also added a -purpose flag to x509 utility to
10138 print out all the purposes.
10141 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10145 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10146 for, obtain and decode and extension and obtain its critical flag.
10147 This allows all the necessary extension code to be handled in a
10148 single function call.
10151 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10152 platforms. See crypto/rc4/rc4_enc.c for further details.
10155 *) New -noout option to asn1parse. This causes no output to be produced
10156 its main use is when combined with -strparse and -out to extract data
10157 from a file (which may not be in ASN.1 format).
10160 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10161 when producing the local key id.
10162 [Richard Levitte <levitte@stacken.kth.se>]
10164 *) New option -dhparam in s_server. This allows a DH parameter file to be
10165 stated explicitly. If it is not stated then it tries the first server
10166 certificate file. The previous behaviour hard coded the filename
10170 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10171 a public key to be input or output. For example:
10172 openssl rsa -in key.pem -pubout -out pubkey.pem
10173 Also added necessary DSA public key functions to handle this.
10176 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10177 in the message. This was handled by allowing
10178 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10179 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10181 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10182 to the end of the strings whereas this didn't. This would cause problems
10183 if strings read with d2i_ASN1_bytes() were later modified.
10184 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10186 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10187 data and it contains EOF it will end up returning an error. This is
10188 caused by input 46 bytes long. The cause is due to the way base64
10189 BIOs find the start of base64 encoded data. They do this by trying a
10190 trial decode on each line until they find one that works. When they
10191 do a flag is set and it starts again knowing it can pass all the
10192 data directly through the decoder. Unfortunately it doesn't reset
10193 the context it uses. This means that if EOF is reached an attempt
10194 is made to pass two EOFs through the context and this causes the
10195 resulting error. This can also cause other problems as well. As is
10196 usual with these problems it takes *ages* to find and the fix is
10197 trivial: move one line.
10198 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10200 *) Ugly workaround to get s_client and s_server working under Windows. The
10201 old code wouldn't work because it needed to select() on sockets and the
10202 tty (for keypresses and to see if data could be written). Win32 only
10203 supports select() on sockets so we select() with a 1s timeout on the
10204 sockets and then see if any characters are waiting to be read, if none
10205 are present then we retry, we also assume we can always write data to
10206 the tty. This isn't nice because the code then blocks until we've
10207 received a complete line of data and it is effectively polling the
10208 keyboard at 1s intervals: however it's quite a bit better than not
10209 working at all :-) A dedicated Windows application might handle this
10210 with an event loop for example.
10213 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10214 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10215 will be called when RSA_sign() and RSA_verify() are used. This is useful
10216 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10217 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10218 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10219 This necessitated the support of an extra signature type NID_md5_sha1
10220 for SSL signatures and modifications to the SSL library to use it instead
10221 of calling RSA_public_decrypt() and RSA_private_encrypt().
10224 *) Add new -verify -CAfile and -CApath options to the crl program, these
10225 will lookup a CRL issuers certificate and verify the signature in a
10226 similar way to the verify program. Tidy up the crl program so it
10227 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10228 less strict. It will now permit CRL extensions even if it is not
10229 a V2 CRL: this will allow it to tolerate some broken CRLs.
10232 *) Initialize all non-automatic variables each time one of the openssl
10233 sub-programs is started (this is necessary as they may be started
10234 multiple times from the "OpenSSL>" prompt).
10235 [Lennart Bang, Bodo Moeller]
10237 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10238 removing all other RSA functionality (this is what NO_RSA does). This
10239 is so (for example) those in the US can disable those operations covered
10240 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10244 *) Non-copying interface to BIO pairs.
10245 (still largely untested)
10248 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10249 ASCII string. This was handled independently in various places before.
10252 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10253 UTF8 strings a character at a time.
10256 *) Use client_version from client hello to select the protocol
10257 (s23_srvr.c) and for RSA client key exchange verification
10258 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10261 *) Add various utility functions to handle SPKACs, these were previously
10262 handled by poking round in the structure internals. Added new function
10263 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10264 print, verify and generate SPKACs. Based on an original idea from
10265 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10268 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10271 *) Allow the config file extension section to be overwritten on the
10272 command line. Based on an original idea from Massimiliano Pala
10273 <madwolf@comune.modena.it>. The new option is called -extensions
10274 and can be applied to ca, req and x509. Also -reqexts to override
10275 the request extensions in req and -crlexts to override the crl extensions
10279 *) Add new feature to the SPKAC handling in ca. Now you can include
10280 the same field multiple times by preceding it by "XXXX." for example:
10283 this is the same syntax as used in the req config file.
10286 *) Allow certificate extensions to be added to certificate requests. These
10287 are specified in a 'req_extensions' option of the req section of the
10288 config file. They can be printed out with the -text option to req but
10289 are otherwise ignored at present.
10292 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10293 data read consists of only the final block it would not decrypted because
10294 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10295 A misplaced 'break' also meant the decrypted final block might not be
10296 copied until the next read.
10299 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10300 a few extra parameters to the DH structure: these will be useful if
10301 for example we want the value of 'q' or implement X9.42 DH.
10304 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10305 provides hooks that allow the default DSA functions or functions on a
10306 "per key" basis to be replaced. This allows hardware acceleration and
10307 hardware key storage to be handled without major modification to the
10308 library. Also added low level modexp hooks and CRYPTO_EX structure and
10309 associated functions.
10312 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10313 as "read only": it can't be written to and the buffer it points to will
10314 not be freed. Reading from a read only BIO is much more efficient than
10315 a normal memory BIO. This was added because there are several times when
10316 an area of memory needs to be read from a BIO. The previous method was
10317 to create a memory BIO and write the data to it, this results in two
10318 copies of the data and an O(n^2) reading algorithm. There is a new
10319 function BIO_new_mem_buf() which creates a read only memory BIO from
10320 an area of memory. Also modified the PKCS#7 routines to use read only
10324 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10325 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10326 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10327 but a retry condition occurred while trying to read the rest.
10330 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10331 NID_pkcs7_encrypted by default: this was wrong since this should almost
10332 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10333 the encrypted data type: this is a more sensible place to put it and it
10334 allows the PKCS#12 code to be tidied up that duplicated this
10338 *) Changed obj_dat.pl script so it takes its input and output files on
10339 the command line. This should avoid shell escape redirection problems
10343 *) Initial support for certificate extension requests, these are included
10344 in things like Xenroll certificate requests. Included functions to allow
10345 extensions to be obtained and added.
10348 *) -crlf option to s_client and s_server for sending newlines as
10349 CRLF (as required by many protocols).
10352 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10354 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10355 [Ralf S. Engelschall]
10357 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10358 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10360 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10364 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10365 DH parameters/keys (q is lost during that conversion, but the resulting
10366 DH parameters contain its length).
10368 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10369 much faster than DH_generate_parameters (which creates parameters
10370 where p = 2*q + 1), and also the smaller q makes DH computations
10371 much more efficient (160-bit exponentiation instead of 1024-bit
10372 exponentiation); so this provides a convenient way to support DHE
10373 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10374 utter importance to use
10375 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10377 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10378 when such DH parameters are used, because otherwise small subgroup
10379 attacks may become possible!
10382 *) Avoid memory leak in i2d_DHparams.
10385 *) Allow the -k option to be used more than once in the enc program:
10386 this allows the same encrypted message to be read by multiple recipients.
10389 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10390 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10391 it will always use the numerical form of the OID, even if it has a short
10395 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10396 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10397 otherwise bn_mod_exp was called. In the case of hardware keys for example
10398 no private key components need be present and it might store extra data
10399 in the RSA structure, which cannot be accessed from bn_mod_exp.
10400 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10401 private key operations.
10404 *) Added support for SPARC Linux.
10407 *) pem_password_cb function type incompatibly changed from
10408 typedef int pem_password_cb(char *buf, int size, int rwflag);
10410 ....(char *buf, int size, int rwflag, void *userdata);
10411 so that applications can pass data to their callbacks:
10412 The PEM[_ASN1]_{read,write}... functions and macros now take an
10413 additional void * argument, which is just handed through whenever
10414 the password callback is called.
10415 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10417 New function SSL_CTX_set_default_passwd_cb_userdata.
10419 Compatibility note: As many C implementations push function arguments
10420 onto the stack in reverse order, the new library version is likely to
10421 interoperate with programs that have been compiled with the old
10422 pem_password_cb definition (PEM_whatever takes some data that
10423 happens to be on the stack as its last argument, and the callback
10424 just ignores this garbage); but there is no guarantee whatsoever that
10427 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10428 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10429 problems not only on Windows, but also on some Unix platforms.
10430 To avoid problematic command lines, these definitions are now in an
10431 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10432 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10435 *) MIPS III/IV assembler module is reimplemented.
10438 *) More DES library cleanups: remove references to srand/rand and
10439 delete an unused file.
10442 *) Add support for the the free Netwide assembler (NASM) under Win32,
10443 since not many people have MASM (ml) and it can be hard to obtain.
10444 This is currently experimental but it seems to work OK and pass all
10445 the tests. Check out INSTALL.W32 for info.
10448 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10449 without temporary keys kept an extra copy of the server key,
10450 and connections with temporary keys did not free everything in case
10454 *) New function RSA_check_key and new openssl rsa option -check
10455 for verifying the consistency of RSA keys.
10456 [Ulf Moeller, Bodo Moeller]
10458 *) Various changes to make Win32 compile work:
10459 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10460 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10461 comparison" warnings.
10462 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10465 *) Add a debugging option to PKCS#5 v2 key generation function: when
10466 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10467 derived keys are printed to stderr.
10470 *) Copy the flags in ASN1_STRING_dup().
10471 [Roman E. Pavlov <pre@mo.msk.ru>]
10473 *) The x509 application mishandled signing requests containing DSA
10474 keys when the signing key was also DSA and the parameters didn't match.
10476 It was supposed to omit the parameters when they matched the signing key:
10477 the verifying software was then supposed to automatically use the CA's
10478 parameters if they were absent from the end user certificate.
10480 Omitting parameters is no longer recommended. The test was also
10481 the wrong way round! This was probably due to unusual behaviour in
10482 EVP_cmp_parameters() which returns 1 if the parameters match.
10483 This meant that parameters were omitted when they *didn't* match and
10484 the certificate was useless. Certificates signed with 'ca' didn't have
10486 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10488 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10489 The interface is as follows:
10490 Applications can use
10491 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10492 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10493 "off" is now the default.
10494 The library internally uses
10495 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10496 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10497 to disable memory-checking temporarily.
10499 Some inconsistent states that previously were possible (and were
10500 even the default) are now avoided.
10502 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10503 with each memory chunk allocated; this is occasionally more helpful
10504 than just having a counter.
10506 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10508 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10512 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10513 which largely parallels "options", but is for changing API behaviour,
10514 whereas "options" are about protocol behaviour.
10515 Initial "mode" flags are:
10517 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10518 a single record has been written.
10519 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10520 retries use the same buffer location.
10521 (But all of the contents must be
10525 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10528 *) Fix problems with no-hmac etc.
10529 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10531 *) New functions RSA_get_default_method(), RSA_set_method() and
10532 RSA_get_method(). These allows replacement of RSA_METHODs without having
10533 to mess around with the internals of an RSA structure.
10536 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10537 Also really enable memory leak checks in openssl.c and in some
10539 [Chad C. Mulligan, Bodo Moeller]
10541 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10542 up the length of negative integers. This has now been simplified to just
10543 store the length when it is first determined and use it later, rather
10544 than trying to keep track of where data is copied and updating it to
10546 [Steve Henson, reported by Brien Wheeler
10547 <bwheeler@authentica-security.com>]
10549 *) Add a new function PKCS7_signatureVerify. This allows the verification
10550 of a PKCS#7 signature but with the signing certificate passed to the
10551 function itself. This contrasts with PKCS7_dataVerify which assumes the
10552 certificate is present in the PKCS#7 structure. This isn't always the
10553 case: certificates can be omitted from a PKCS#7 structure and be
10554 distributed by "out of band" means (such as a certificate database).
10557 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10558 function prototypes in pem.h, also change util/mkdef.pl to add the
10559 necessary function names.
10562 *) mk1mf.pl (used by Windows builds) did not properly read the
10563 options set by Configure in the top level Makefile, and Configure
10564 was not even able to write more than one option correctly.
10565 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10568 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10569 file to be loaded from a BIO or FILE pointer. The BIO version will
10570 for example allow memory BIOs to contain config info.
10573 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10574 Whoever hopes to achieve shared-library compatibility across versions
10575 must use this, not the compile-time macro.
10576 (Exercise 0.9.4: Which is the minimum library version required by
10578 Note: All this applies only to multi-threaded programs, others don't
10582 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10583 through a BIO pair triggered the default case, i.e.
10584 SSLerr(...,SSL_R_UNKNOWN_STATE).
10587 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10588 can use the SSL library even if none of the specific BIOs is
10592 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10593 for the encoded length.
10594 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10596 *) Add initial documentation of the X509V3 functions.
10599 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10600 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10601 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10602 secure PKCS#8 private key format with a high iteration count.
10605 *) Fix determination of Perl interpreter: A perl or perl5
10606 _directory_ in $PATH was also accepted as the interpreter.
10607 [Ralf S. Engelschall]
10609 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10610 wrong with it but it was very old and did things like calling
10611 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10612 unusual formatting.
10615 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10616 to use the new extension code.
10619 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10620 with macros. This should make it easier to change their form, add extra
10621 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10625 *) Add to configuration table a new entry that can specify an alternative
10626 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10627 according to Mark Crispin <MRC@Panda.COM>.
10631 *) DES CBC did not update the IV. Weird.
10634 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10635 Changing the behaviour of the former might break existing programs --
10636 where IV updating is needed, des_ncbc_encrypt can be used.
10639 *) When bntest is run from "make test" it drives bc to check its
10640 calculations, as well as internally checking them. If an internal check
10641 fails, it needs to cause bc to give a non-zero result or make test carries
10642 on without noticing the failure. Fixed.
10645 *) DES library cleanups.
10648 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10649 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10650 ciphers. NOTE: although the key derivation function has been verified
10651 against some published test vectors it has not been extensively tested
10652 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10656 *) Instead of "mkdir -p", which is not fully portable, use new
10657 Perl script "util/mkdir-p.pl".
10660 *) Rewrite the way password based encryption (PBE) is handled. It used to
10661 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10662 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10663 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10664 the 'parameter' field of the AlgorithmIdentifier is passed to the
10665 underlying key generation function so it must do its own ASN1 parsing.
10666 This has also changed the EVP_PBE_CipherInit() function which now has a
10667 'parameter' argument instead of literal salt and iteration count values
10668 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10671 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10672 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10673 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10674 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10675 value was just used as a "magic string" and not used directly its
10676 value doesn't matter.
10679 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10683 *) "linux-sparc64" configuration (ultrapenguin).
10684 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10685 "linux-sparc" configuration.
10686 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10688 *) config now generates no-xxx options for missing ciphers.
10691 *) Support the EBCDIC character set (work in progress).
10692 File ebcdic.c not yet included because it has a different license.
10693 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10695 *) Support BS2000/OSD-POSIX.
10696 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10698 *) Make callbacks for key generation use void * instead of char *.
10701 *) Make S/MIME samples compile (not yet tested).
10704 *) Additional typesafe stacks.
10707 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10711 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10713 *) New configuration variant "sco5-gcc".
10715 *) Updated some demos.
10716 [Sean O Riordain, Wade Scholine]
10718 *) Add missing BIO_free at exit of pkcs12 application.
10721 *) Fix memory leak in conf.c.
10724 *) Updates for Win32 to assembler version of MD5.
10727 *) Set #! path to perl in apps/der_chop to where we found it
10728 instead of using a fixed path.
10731 *) SHA library changes for irix64-mips4-cc.
10734 *) Improvements for VMS support.
10738 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10740 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10741 This also avoids the problems with SC4.2 and unpatched SC5.
10742 [Andy Polyakov <appro@fy.chalmers.se>]
10744 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10745 These are required because of the typesafe stack would otherwise break
10746 existing code. If old code used a structure member which used to be STACK
10747 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10748 sk_num or sk_value it would produce an error because the num, data members
10749 are not present in STACK_OF. Now it just produces a warning. sk_set
10750 replaces the old method of assigning a value to sk_value
10751 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10752 that does this will no longer work (and should use sk_set instead) but
10753 this could be regarded as a "questionable" behaviour anyway.
10756 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10757 correctly handle encrypted S/MIME data.
10760 *) Change type of various DES function arguments from des_cblock
10761 (which means, in function argument declarations, pointer to char)
10762 to des_cblock * (meaning pointer to array with 8 char elements),
10763 which allows the compiler to do more typechecking; it was like
10764 that back in SSLeay, but with lots of ugly casts.
10766 Introduce new type const_des_cblock.
10769 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10770 problems: find RecipientInfo structure that matches recipient certificate
10771 and initialise the ASN1 structures properly based on passed cipher.
10774 *) Belatedly make the BN tests actually check the results.
10777 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10778 to and from BNs: it was completely broken. New compilation option
10779 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10780 key elements as negative integers.
10783 *) Reorganize and speed up MD5.
10784 [Andy Polyakov <appro@fy.chalmers.se>]
10787 [Richard Levitte <richard@levitte.org>]
10789 *) New option -out to asn1parse to allow the parsed structure to be
10790 output to a file. This is most useful when combined with the -strparse
10791 option to examine the output of things like OCTET STRINGS.
10794 *) Make SSL library a little more fool-proof by not requiring any longer
10795 that SSL_set_{accept,connect}_state be called before
10796 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10797 in many applications because usually everything *appeared* to work as
10798 intended anyway -- now it really works as intended).
10801 *) Move openssl.cnf out of lib/.
10804 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10805 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10806 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10807 [Ralf S. Engelschall]
10809 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10810 handle PKCS#7 enveloped data properly.
10811 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10813 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10814 copying pointers. The cert_st handling is changed by this in
10815 various ways (and thus what used to be known as ctx->default_cert
10816 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10817 any longer when s->cert does not give us what we need).
10818 ssl_cert_instantiate becomes obsolete by this change.
10819 As soon as we've got the new code right (possibly it already is?),
10820 we have solved a couple of bugs of the earlier code where s->cert
10821 was used as if it could not have been shared with other SSL structures.
10823 Note that using the SSL API in certain dirty ways now will result
10824 in different behaviour than observed with earlier library versions:
10825 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10826 does not influence s as it used to.
10828 In order to clean up things more thoroughly, inside SSL_SESSION
10829 we don't use CERT any longer, but a new structure SESS_CERT
10830 that holds per-session data (if available); currently, this is
10831 the peer's certificate chain and, for clients, the server's certificate
10832 and temporary key. CERT holds only those values that can have
10833 meaningful defaults in an SSL_CTX.
10836 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10837 from the internal representation. Various PKCS#7 fixes: remove some
10838 evil casts and set the enc_dig_alg field properly based on the signing
10842 *) Allow PKCS#12 password to be set from the command line or the
10843 environment. Let 'ca' get its config file name from the environment
10844 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10848 *) Allow certificate policies extension to use an IA5STRING for the
10849 organization field. This is contrary to the PKIX definition but
10850 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10854 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10855 without disallowing inline assembler and the like for non-pedantic builds.
10858 *) Support Borland C++ builder.
10859 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10861 *) Support Mingw32.
10864 *) SHA-1 cleanups and performance enhancements.
10865 [Andy Polyakov <appro@fy.chalmers.se>]
10867 *) Sparc v8plus assembler for the bignum library.
10868 [Andy Polyakov <appro@fy.chalmers.se>]
10870 *) Accept any -xxx and +xxx compiler options in Configure.
10873 *) Update HPUX configuration.
10876 *) Add missing sk_<type>_unshift() function to safestack.h
10877 [Ralf S. Engelschall]
10879 *) New function SSL_CTX_use_certificate_chain_file that sets the
10880 "extra_cert"s in addition to the certificate. (This makes sense
10881 only for "PEM" format files, as chains as a whole are not
10885 *) Support verify_depth from the SSL API.
10886 x509_vfy.c had what can be considered an off-by-one-error:
10887 Its depth (which was not part of the external interface)
10888 was actually counting the number of certificates in a chain;
10889 now it really counts the depth.
10892 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10893 instead of X509err, which often resulted in confusing error
10894 messages since the error codes are not globally unique
10895 (e.g. an alleged error in ssl3_accept when a certificate
10896 didn't match the private key).
10898 *) New function SSL_CTX_set_session_id_context that allows to set a default
10899 value (so that you don't need SSL_set_session_id_context for each
10900 connection using the SSL_CTX).
10903 *) OAEP decoding bug fix.
10906 *) Support INSTALL_PREFIX for package builders, as proposed by
10910 *) New Configure options "threads" and "no-threads". For systems
10911 where the proper compiler options are known (currently Solaris
10912 and Linux), "threads" is the default.
10915 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10918 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10919 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10920 such as /usr/local/bin.
10923 *) "make linux-shared" to build shared libraries.
10924 [Niels Poppe <niels@netbox.org>]
10926 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10929 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10930 extension adding in x509 utility.
10933 *) Remove NOPROTO sections and error code comments.
10936 *) Partial rewrite of the DEF file generator to now parse the ANSI
10940 *) New Configure options --prefix=DIR and --openssldir=DIR.
10943 *) Complete rewrite of the error code script(s). It is all now handled
10944 by one script at the top level which handles error code gathering,
10945 header rewriting and C source file generation. It should be much better
10946 than the old method: it now uses a modified version of Ulf's parser to
10947 read the ANSI prototypes in all header files (thus the old K&R definitions
10948 aren't needed for error creation any more) and do a better job of
10949 translating function codes into names. The old 'ASN1 error code imbedded
10950 in a comment' is no longer necessary and it doesn't use .err files which
10951 have now been deleted. Also the error code call doesn't have to appear all
10952 on one line (which resulted in some large lines...).
10955 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10958 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10959 0 (which usually indicates a closed connection), but continue reading.
10962 *) Fix some race conditions.
10965 *) Add support for CRL distribution points extension. Add Certificate
10966 Policies and CRL distribution points documentation.
10969 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10972 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10973 8 of keying material. Merlin has also confirmed interop with this fix
10974 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10975 [Merlin Hughes <merlin@baltimore.ie>]
10977 *) Fix lots of warnings.
10978 [Richard Levitte <levitte@stacken.kth.se>]
10980 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10981 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10982 [Richard Levitte <levitte@stacken.kth.se>]
10984 *) Fix problems with sizeof(long) == 8.
10985 [Andy Polyakov <appro@fy.chalmers.se>]
10987 *) Change functions to ANSI C.
10990 *) Fix typos in error codes.
10991 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10993 *) Remove defunct assembler files from Configure.
10996 *) SPARC v8 assembler BIGNUM implementation.
10997 [Andy Polyakov <appro@fy.chalmers.se>]
10999 *) Support for Certificate Policies extension: both print and set.
11000 Various additions to support the r2i method this uses.
11003 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11004 return a const string when you are expecting an allocated buffer.
11007 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11008 types DirectoryString and DisplayText.
11011 *) Add code to allow r2i extensions to access the configuration database,
11012 add an LHASH database driver and add several ctx helper functions.
11015 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11016 fail when they extended the size of a BIGNUM.
11019 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11020 support typesafe stack.
11023 *) Fix typo in SSL_[gs]et_options().
11024 [Nils Frostberg <nils@medcom.se>]
11026 *) Delete various functions and files that belonged to the (now obsolete)
11027 old X509V3 handling code.
11030 *) New Configure option "rsaref".
11033 *) Don't auto-generate pem.h.
11036 *) Introduce type-safe ASN.1 SETs.
11039 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11040 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11042 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11043 that links with OpenSSL (well at least cause lots of warnings), but fear
11044 not: the conversion is trivial, and it eliminates loads of evil casts. A
11045 few STACKed things have been converted already. Feel free to convert more.
11046 In the fullness of time, I'll do away with the STACK type altogether.
11049 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11050 specified in <certfile> by updating the entry in the index.txt file.
11051 This way one no longer has to edit the index.txt file manually for
11052 revoking a certificate. The -revoke option does the gory details now.
11053 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11055 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11056 `-text' option at all and this way the `-noout -text' combination was
11057 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11058 [Ralf S. Engelschall]
11060 *) Make sure a corresponding plain text error message exists for the
11061 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11062 verify callback function determined that a certificate was revoked.
11063 [Ralf S. Engelschall]
11065 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11066 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11067 all available cipers including rc5, which was forgotten until now.
11068 In order to let the testing shell script know which algorithms
11069 are available, a new (up to now undocumented) command
11070 "openssl list-cipher-commands" is used.
11073 *) Bugfix: s_client occasionally would sleep in select() when
11074 it should have checked SSL_pending() first.
11077 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11078 the raw DSA values prior to ASN.1 encoding.
11081 *) Tweaks to Configure
11082 [Niels Poppe <niels@netbox.org>]
11084 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11088 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11091 *) New config option to avoid instructions that are illegal on the 80386.
11092 The default code is faster, but requires at least a 486.
11095 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11096 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11097 same as SSL2_VERSION anyway.
11100 *) New "-showcerts" option for s_client.
11103 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11104 application. Various cleanups and fixes.
11107 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11108 modify error routines to work internally. Add error codes and PBE init
11109 to library startup routines.
11112 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11113 packing functions to asn1 and evp. Changed function names and error
11114 codes along the way.
11117 *) PKCS12 integration: and so it begins... First of several patches to
11118 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11119 objects to objects.h
11122 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11123 and display support for Thawte strong extranet extension.
11126 *) Add LinuxPPC support.
11127 [Jeff Dubrule <igor@pobox.org>]
11129 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11130 bn_div_words in alpha.s.
11131 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11133 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11134 OAEP isn't supported when OpenSSL is built with RSAref.
11135 [Ulf Moeller <ulf@fitug.de>]
11137 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11138 so they no longer are missing under -DNOPROTO.
11139 [Soren S. Jorvang <soren@t.dk>]
11142 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11144 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11145 doesn't work when the session is reused. Coming soon!
11148 *) Fix a security hole, that allows sessions to be reused in the wrong
11149 context thus bypassing client cert protection! All software that uses
11150 client certs and session caches in multiple contexts NEEDS PATCHING to
11151 allow session reuse! A fuller solution is in the works.
11152 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11154 *) Some more source tree cleanups (removed obsolete files
11155 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11156 permission on "config" script to be executable) and a fix for the INSTALL
11158 [Ulf Moeller <ulf@fitug.de>]
11160 *) Remove some legacy and erroneous uses of malloc, free instead of
11162 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11164 *) Make rsa_oaep_test return non-zero on error.
11165 [Ulf Moeller <ulf@fitug.de>]
11167 *) Add support for native Solaris shared libraries. Configure
11168 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11169 if someone would make that last step automatic.
11170 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11172 *) ctx_size was not built with the right compiler during "make links". Fixed.
11175 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11176 except NULL ciphers". This means the default cipher list will no longer
11177 enable NULL ciphers. They need to be specifically enabled e.g. with
11178 the string "DEFAULT:eNULL".
11181 *) Fix to RSA private encryption routines: if p < q then it would
11182 occasionally produce an invalid result. This will only happen with
11183 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11186 *) Be less restrictive and allow also `perl util/perlpath.pl
11187 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11188 because this way one can also use an interpreter named `perl5' (which is
11189 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11190 installed as `perl').
11191 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11193 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11194 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11196 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11197 advapi32.lib to Win32 build and change the pem test comparison
11198 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11199 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11200 and crypto/des/ede_cbcm_enc.c.
11203 *) DES quad checksum was broken on big-endian architectures. Fixed.
11206 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11207 Win32 test batch file so it (might) work again. The Win32 test batch file
11208 is horrible: I feel ill....
11211 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11212 in e_os.h. Audit of header files to check ANSI and non ANSI
11213 sections: 10 functions were absent from non ANSI section and not exported
11214 from Windows DLLs. Fixed up libeay.num for new functions.
11217 *) Make `openssl version' output lines consistent.
11218 [Ralf S. Engelschall]
11220 *) Fix Win32 symbol export lists for BIO functions: Added
11221 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11222 to ms/libeay{16,32}.def.
11223 [Ralf S. Engelschall]
11225 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11226 fine under Unix and passes some trivial tests I've now added. But the
11227 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11228 added to make sure no one expects that this stuff really works in the
11229 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11230 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11232 [Ralf S. Engelschall]
11234 *) Fix the generation of two part addresses in perl.
11235 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11237 *) Add config entry for Linux on MIPS.
11238 [John Tobey <jtobey@channel1.com>]
11240 *) Make links whenever Configure is run, unless we are on Windoze.
11243 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11244 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11248 *) Add a useful kludge to allow package maintainers to specify compiler and
11249 other platforms details on the command line without having to patch the
11250 Configure script everytime: One now can use ``perl Configure
11251 <id>:<details>'', i.e. platform ids are allowed to have details appended
11252 to them (separated by colons). This is treated as there would be a static
11253 pre-configured entry in Configure's %table under key <id> with value
11254 <details> and ``perl Configure <id>'' is called. So, when you want to
11255 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11256 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11257 now, which overrides the FreeBSD-elf entry on-the-fly.
11258 [Ralf S. Engelschall]
11260 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11263 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11264 on the `perl Configure ...' command line. This way one can compile
11265 OpenSSL libraries with Position Independent Code (PIC) which is needed
11266 for linking it into DSOs.
11267 [Ralf S. Engelschall]
11269 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11273 *) Cleaned up the LICENSE document: The official contact for any license
11274 questions now is the OpenSSL core team under openssl-core@openssl.org.
11275 And add a paragraph about the dual-license situation to make sure people
11276 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11277 to the OpenSSL toolkit.
11278 [Ralf S. Engelschall]
11280 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11281 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11282 Additionally cleaned up the `make links' target: Remove unnecessary
11283 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11284 to speed processing and no longer clutter the display with confusing
11285 stuff. Instead only the actually done links are displayed.
11286 [Ralf S. Engelschall]
11288 *) Permit null encryption ciphersuites, used for authentication only. It used
11289 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11290 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11294 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11295 signed attributes when verifying signatures (this would break them),
11296 the detached data encoding was wrong and public keys obtained using
11297 X509_get_pubkey() weren't freed.
11300 *) Add text documentation for the BUFFER functions. Also added a work around
11301 to a Win95 console bug. This was triggered by the password read stuff: the
11302 last character typed gets carried over to the next fread(). If you were
11303 generating a new cert request using 'req' for example then the last
11304 character of the passphrase would be CR which would then enter the first
11308 *) Added the new `Includes OpenSSL Cryptography Software' button as
11309 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11310 button and can be used by applications based on OpenSSL to show the
11311 relationship to the OpenSSL project.
11312 [Ralf S. Engelschall]
11314 *) Remove confusing variables in function signatures in files
11315 ssl/ssl_lib.c and ssl/ssl.h.
11316 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11318 *) Don't install bss_file.c under PREFIX/include/
11319 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11321 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11322 functions that return function pointers and has support for NT specific
11323 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11324 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11325 unsigned to signed types: this was killing the Win32 compile.
11328 *) Add new certificate file to stack functions,
11329 SSL_add_dir_cert_subjects_to_stack() and
11330 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11331 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11332 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11333 This means that Apache-SSL and similar packages don't have to mess around
11334 to add as many CAs as they want to the preferred list.
11337 *) Experiment with doxygen documentation. Currently only partially applied to
11339 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11340 openssl.doxy as the configuration file.
11343 *) Get rid of remaining C++-style comments which strict C compilers hate.
11344 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11346 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11347 compiled in by default: it has problems with large keys.
11350 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11351 DH private keys and/or callback functions which directly correspond to
11352 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11353 is needed for applications which have to configure certificates on a
11354 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11356 For the RSA certificate situation is makes no difference, but
11357 for the DSA certificate situation this fixes the "no shared cipher"
11358 problem where the OpenSSL cipher selection procedure failed because the
11359 temporary keys were not overtaken from the context and the API provided
11360 no way to reconfigure them.
11361 The new functions now let applications reconfigure the stuff and they
11362 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11363 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11364 non-public-API function ssl_cert_instantiate() is used as a helper
11365 function and also to reduce code redundancy inside ssl_rsa.c.
11366 [Ralf S. Engelschall]
11368 *) Move s_server -dcert and -dkey options out of the undocumented feature
11369 area because they are useful for the DSA situation and should be
11370 recognized by the users.
11371 [Ralf S. Engelschall]
11373 *) Fix the cipher decision scheme for export ciphers: the export bits are
11374 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11375 SSL_EXP_MASK. So, the original variable has to be used instead of the
11376 already masked variable.
11377 [Richard Levitte <levitte@stacken.kth.se>]
11379 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11380 [Richard Levitte <levitte@stacken.kth.se>]
11382 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11383 from `int' to `unsigned int' because it's a length and initialized by
11384 EVP_DigestFinal() which expects an `unsigned int *'.
11385 [Richard Levitte <levitte@stacken.kth.se>]
11387 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11388 script. Instead use the usual Shell->Perl transition trick.
11389 [Ralf S. Engelschall]
11391 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11392 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11393 -noout -modulus' as it's already the case for `openssl rsa -noout
11394 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11395 currently the public key is printed (a decision which was already done by
11396 `openssl dsa -modulus' in the past) which serves a similar purpose.
11397 Additionally the NO_RSA no longer completely removes the whole -modulus
11398 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11400 [Ralf S. Engelschall]
11402 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11403 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11404 [Arne Ansper <arne@ats.cyber.ee>]
11406 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11407 to be added. Now both 'req' and 'ca' can use new objects defined in the
11411 *) Add cool BIO that does syslog (or event log on NT).
11412 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11414 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11415 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11416 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11417 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11420 *) Add preliminary config info for new extension code.
11423 *) Make RSA_NO_PADDING really use no padding.
11424 [Ulf Moeller <ulf@fitug.de>]
11426 *) Generate errors when private/public key check is done.
11429 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11430 for some CRL extensions and new objects added.
11433 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11434 key usage extension and fuller support for authority key id.
11437 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11438 padding method for RSA, which is recommended for new applications in PKCS
11439 #1 v2.0 (RFC 2437, October 1998).
11440 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11441 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11442 against Bleichbacher's attack on RSA.
11443 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11446 *) Updates to the new SSL compression code
11447 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11449 *) Fix so that the version number in the master secret, when passed
11450 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11451 (because the server will not accept higher), that the version number
11452 is 0x03,0x01, not 0x03,0x00
11453 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11455 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11456 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11457 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11460 *) Support for RAW extensions where an arbitrary extension can be
11461 created by including its DER encoding. See apps/openssl.cnf for
11465 *) Make sure latest Perl versions don't interpret some generated C array
11466 code as Perl array code in the crypto/err/err_genc.pl script.
11467 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11469 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11470 not many people have the assembler. Various Win32 compilation fixes and
11471 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11472 build instructions.
11475 *) Modify configure script 'Configure' to automatically create crypto/date.h
11476 file under Win32 and also build pem.h from pem.org. New script
11477 util/mkfiles.pl to create the MINFO file on environments that can't do a
11478 'make files': perl util/mkfiles.pl >MINFO should work.
11481 *) Major rework of DES function declarations, in the pursuit of correctness
11482 and purity. As a result, many evil casts evaporated, and some weirdness,
11483 too. You may find this causes warnings in your code. Zapping your evil
11484 casts will probably fix them. Mostly.
11487 *) Fix for a typo in asn1.h. Bug fix to object creation script
11488 obj_dat.pl. It considered a zero in an object definition to mean
11489 "end of object": none of the objects in objects.h have any zeros
11490 so it wasn't spotted.
11491 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11493 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11494 Masking (CBCM). In the absence of test vectors, the best I have been able
11495 to do is check that the decrypt undoes the encrypt, so far. Send me test
11496 vectors if you have them.
11499 *) Correct calculation of key length for export ciphers (too much space was
11500 allocated for null ciphers). This has not been tested!
11503 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11504 message is now correct (it understands "crypto" and "ssl" on its
11505 command line). There is also now an "update" option. This will update
11506 the util/ssleay.num and util/libeay.num files with any new functions.
11508 perl util/mkdef.pl crypto ssl update
11509 it will update them.
11512 *) Overhauled the Perl interface (perl/*):
11513 - ported BN stuff to OpenSSL's different BN library
11514 - made the perl/ source tree CVS-aware
11515 - renamed the package from SSLeay to OpenSSL (the files still contain
11516 their history because I've copied them in the repository)
11517 - removed obsolete files (the test scripts will be replaced
11518 by better Test::Harness variants in the future)
11519 [Ralf S. Engelschall]
11521 *) First cut for a very conservative source tree cleanup:
11522 1. merge various obsolete readme texts into doc/ssleay.txt
11523 where we collect the old documents and readme texts.
11524 2. remove the first part of files where I'm already sure that we no
11525 longer need them because of three reasons: either they are just temporary
11526 files which were left by Eric or they are preserved original files where
11527 I've verified that the diff is also available in the CVS via "cvs diff
11528 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11529 the crypto/md/ stuff).
11530 [Ralf S. Engelschall]
11532 *) More extension code. Incomplete support for subject and issuer alt
11533 name, issuer and authority key id. Change the i2v function parameters
11534 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11535 what that's for :-) Fix to ASN1 macro which messed up
11536 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11539 *) Preliminary support for ENUMERATED type. This is largely copied from the
11543 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11544 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11546 *) Make sure `make rehash' target really finds the `openssl' program.
11547 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11549 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11550 like to hear about it if this slows down other processors.
11553 *) Add CygWin32 platform information to Configure script.
11554 [Alan Batie <batie@aahz.jf.intel.com>]
11556 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11557 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11559 *) New program nseq to manipulate netscape certificate sequences
11562 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11566 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11567 but the BN code had some problems that would cause failures when
11568 doing certificate verification and some other functions.
11569 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11571 *) Add ASN1 and PEM code to support netscape certificate sequences.
11574 *) Add ASN1 and PEM code to support netscape certificate sequences.
11577 *) Add several PKIX and private extended key usage OIDs.
11580 *) Modify the 'ca' program to handle the new extension code. Modify
11581 openssl.cnf for new extension format, add comments.
11584 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11585 and add a sample to openssl.cnf so req -x509 now adds appropriate
11589 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11590 error code, add initial support to X509_print() and x509 application.
11593 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11594 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11595 stuff is currently isolated and isn't even compiled yet.
11598 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11599 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11600 Removed the versions check from X509 routines when loading extensions:
11601 this allows certain broken certificates that don't set the version
11602 properly to be processed.
11605 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11606 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11607 can still be regenerated with "make depend".
11610 *) Spelling mistake in C version of CAST-128.
11611 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11613 *) Changes to the error generation code. The perl script err-code.pl
11614 now reads in the old error codes and retains the old numbers, only
11615 adding new ones if necessary. It also only changes the .err files if new
11616 codes are added. The makefiles have been modified to only insert errors
11617 when needed (to avoid needlessly modifying header files). This is done
11618 by only inserting errors if the .err file is newer than the auto generated
11619 C file. To rebuild all the error codes from scratch (the old behaviour)
11620 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11621 or delete all the .err files.
11624 *) CAST-128 was incorrectly implemented for short keys. The C version has
11625 been fixed, but is untested. The assembler versions are also fixed, but
11626 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11627 to regenerate it if needed.
11628 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11629 Hagino <itojun@kame.net>]
11631 *) File was opened incorrectly in randfile.c.
11632 [Ulf Möller <ulf@fitug.de>]
11634 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11635 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11636 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11637 al: it's just almost always a UTCTime. Note this patch adds new error
11638 codes so do a "make errors" if there are problems.
11641 *) Correct Linux 1 recognition in config.
11642 [Ulf Möller <ulf@fitug.de>]
11644 *) Remove pointless MD5 hash when using DSA keys in ca.
11645 [Anonymous <nobody@replay.com>]
11647 *) Generate an error if given an empty string as a cert directory. Also
11648 generate an error if handed NULL (previously returned 0 to indicate an
11649 error, but didn't set one).
11650 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11652 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11655 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11656 parameters. This was causing a warning which killed off the Win32 compile.
11659 *) Remove C++ style comments from crypto/bn/bn_local.h.
11660 [Neil Costigan <neil.costigan@celocom.com>]
11662 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11663 based on a text string, looking up short and long names and finally
11664 "dot" format. The "dot" format stuff didn't work. Added new function
11665 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11666 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11667 OID is not part of the table.
11670 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11671 X509_LOOKUP_by_alias().
11674 *) Sort openssl functions by name.
11677 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11678 encryption from sample DSA keys (in case anyone is interested the password
11682 *) Make _all_ *_free functions accept a NULL pointer.
11683 [Frans Heymans <fheymans@isaserver.be>]
11685 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11687 [Anonymous <nobody@replay.com>]
11689 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11690 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11692 *) Don't blow it for numeric -newkey arguments to apps/req.
11693 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11695 *) Temp key "for export" tests were wrong in s3_srvr.c.
11696 [Anonymous <nobody@replay.com>]
11698 *) Add prototype for temp key callback functions
11699 SSL_CTX_set_tmp_{rsa,dh}_callback().
11702 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11703 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11706 *) X509_name_add_entry() freed the wrong thing after an error.
11707 [Arne Ansper <arne@ats.cyber.ee>]
11709 *) rsa_eay.c would attempt to free a NULL context.
11710 [Arne Ansper <arne@ats.cyber.ee>]
11712 *) BIO_s_socket() had a broken should_retry() on Windoze.
11713 [Arne Ansper <arne@ats.cyber.ee>]
11715 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11716 [Arne Ansper <arne@ats.cyber.ee>]
11718 *) Make sure the already existing X509_STORE->depth variable is initialized
11719 in X509_STORE_new(), but document the fact that this variable is still
11720 unused in the certificate verification process.
11721 [Ralf S. Engelschall]
11723 *) Fix the various library and apps files to free up pkeys obtained from
11724 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11727 *) Fix reference counting in X509_PUBKEY_get(). This makes
11728 demos/maurice/example2.c work, amongst others, probably.
11729 [Steve Henson and Ben Laurie]
11731 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11732 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11733 are no longer created. This way we have a single and consistent command
11734 line interface `openssl <command>', similar to `cvs <command>'.
11735 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11737 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11738 BIT STRING wrapper always have zero unused bits.
11741 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11744 *) Make the top-level INSTALL documentation easier to understand.
11747 *) Makefiles updated to exit if an error occurs in a sub-directory
11748 make (including if user presses ^C) [Paul Sutton]
11750 *) Make Montgomery context stuff explicit in RSA data structure.
11753 *) Fix build order of pem and err to allow for generated pem.h.
11756 *) Fix renumbering bug in X509_NAME_delete_entry().
11759 *) Enhanced the err-ins.pl script so it makes the error library number
11760 global and can add a library name. This is needed for external ASN1 and
11761 other error libraries.
11764 *) Fixed sk_insert which never worked properly.
11767 *) Fix ASN1 macros so they can handle indefinite length constructed
11768 EXPLICIT tags. Some non standard certificates use these: they can now
11772 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11773 into a single doc/ssleay.txt bundle. This way the information is still
11774 preserved but no longer messes up this directory. Now it's new room for
11775 the new set of documentation files.
11776 [Ralf S. Engelschall]
11778 *) SETs were incorrectly DER encoded. This was a major pain, because they
11779 shared code with SEQUENCEs, which aren't coded the same. This means that
11780 almost everything to do with SETs or SEQUENCEs has either changed name or
11781 number of arguments.
11782 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11784 *) Fix test data to work with the above.
11787 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11788 was already fixed by Eric for 0.9.1 it seems.
11789 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11791 *) Autodetect FreeBSD3.
11794 *) Fix various bugs in Configure. This affects the following platforms:
11798 unixware-2.0-pentium
11802 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11803 before they are needed.
11806 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11810 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11812 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11813 changed SSLeay to OpenSSL in version strings.
11814 [Ralf S. Engelschall]
11816 *) Some fixups to the top-level documents.
11819 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11820 because the symlink to include/ was missing.
11821 [Ralf S. Engelschall]
11823 *) Incorporated the popular no-RSA/DSA-only patches
11824 which allow to compile a RSA-free SSLeay.
11825 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11827 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11828 when "ssleay" is still not found.
11829 [Ralf S. Engelschall]
11831 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11832 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11834 *) Updated the README file.
11835 [Ralf S. Engelschall]
11837 *) Added various .cvsignore files in the CVS repository subdirs
11838 to make a "cvs update" really silent.
11839 [Ralf S. Engelschall]
11841 *) Recompiled the error-definition header files and added
11842 missing symbols to the Win32 linker tables.
11843 [Ralf S. Engelschall]
11845 *) Cleaned up the top-level documents;
11846 o new files: CHANGES and LICENSE
11847 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11848 o merged COPYRIGHT into LICENSE
11849 o removed obsolete TODO file
11850 o renamed MICROSOFT to INSTALL.W32
11851 [Ralf S. Engelschall]
11853 *) Removed dummy files from the 0.9.1b source tree:
11854 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11855 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11856 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11857 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11858 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11859 [Ralf S. Engelschall]
11861 *) Added various platform portability fixes.
11864 *) The Genesis of the OpenSSL rpject:
11865 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11866 Young and Tim J. Hudson created while they were working for C2Net until
11868 [The OpenSSL Project]
11871 Changes between 0.9.0b and 0.9.1b [not released]
11873 *) Updated a few CA certificates under certs/
11876 *) Changed some BIGNUM api stuff.
11879 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11880 DGUX x86, Linux Alpha, etc.
11883 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11884 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11888 *) Add -strparse option to asn1pars program which parses nested
11890 [Dr Stephen Henson <shenson@bigfoot.com>]
11892 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11895 *) DSA fix for "ca" program.
11898 *) Added "-genkey" option to "dsaparam" program.
11901 *) Added RIPE MD160 (rmd160) message digest.
11904 *) Added -a (all) option to "ssleay version" command.
11907 *) Added PLATFORM define which is the id given to Configure.
11910 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11913 *) Extended the ASN.1 parser routines.
11916 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11919 *) Added a BN_CTX to the BN library.
11922 *) Fixed the weak key values in DES library
11925 *) Changed API in EVP library for cipher aliases.
11928 *) Added support for RC2/64bit cipher.
11931 *) Converted the lhash library to the crypto/mem.c functions.
11934 *) Added more recognized ASN.1 object ids.
11937 *) Added more RSA padding checks for SSL/TLS.
11940 *) Added BIO proxy/filter functionality.
11943 *) Added extra_certs to SSL_CTX which can be used
11944 send extra CA certificates to the client in the CA cert chain sending
11945 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11948 *) Now Fortezza is denied in the authentication phase because
11949 this is key exchange mechanism is not supported by SSLeay at all.
11952 *) Additional PKCS1 checks.
11955 *) Support the string "TLSv1" for all TLS v1 ciphers.
11958 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11959 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11962 *) Fixed a few memory leaks.
11965 *) Fixed various code and comment typos.
11968 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11969 bytes sent in the client random.
11970 [Edward Bishop <ebishop@spyglass.com>]