5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
8 disabled by default. They can be re-enabled using the
9 enable-weak-ssl-ciphers option to Configure.
12 *) If the server has ALPN configured, but supports no protocols that the
13 client advertises, send a fatal "no_application_protocol" alert.
14 This behaviour is SHALL in RFC 7301, though it isn't universally
15 implemented by other servers.
18 *) Add X25519 support.
19 Integrate support for X25519 into EC library. This includes support
20 for public and private key encoding using the format documented in
21 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
22 OID from that draft, encodes public keys using little endian
23 format in the ECPoint structure and private keys using
24 little endian form in the privateKey field of the ECPrivateKey
25 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
28 Note: the current version supports key generation, public and
29 private key encoding and ECDH key agreement using the EC API.
30 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
34 *) Deprecate SRP_VBASE_get_by_user.
35 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
36 In order to fix an unavoidable memory leak (CVE-2016-0798),
37 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
38 seed, even if the seed is configured.
40 Users should use SRP_VBASE_get1_by_user instead. Note that in
41 SRP_VBASE_get1_by_user, caller must free the returned value. Note
42 also that even though configuring the SRP seed attempts to hide
43 invalid usernames by continuing the handshake with fake
44 credentials, this behaviour is not constant time and no strong
45 guarantees are made that the handshake is indistinguishable from
49 *) Configuration change; it's now possible to build dynamic engines
50 without having to build shared libraries and vice versa. This
51 only applies to the engines in engines/, those in crypto/engine/
52 will always be built into libcrypto (i.e. "static").
54 Building dynamic engines is enabled by default; to disable, use
55 the configuration option "disable-dynamic-engine".
57 The only requirements for building dynamic engines are the
58 presence of the DSO module and building with position independent
59 code, so they will also automatically be disabled if configuring
60 with "disable-dso" or "disable-pic".
62 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
63 are also taken away from openssl/opensslconf.h, as they are
67 *) Configuration change; if there is a known flag to compile
68 position independent code, it will always be applied on the
69 libcrypto and libssl object files, and never on the application
70 object files. This means other libraries that use routines from
71 libcrypto / libssl can be made into shared libraries regardless
72 of how OpenSSL was configured.
74 If this isn't desirable, the configuration options "disable-pic"
75 or "no-pic" can be used to disable the use of PIC. This will
76 also disable building shared libraries and dynamic engines.
79 *) Removed JPAKE code. It was experimental and has no wide use.
82 *) The INSTALL_PREFIX Makefile variable has been renamed to
83 DESTDIR. That makes for less confusion on what this variable
84 is for. Also, the configuration option --install_prefix is
88 *) Heartbeat for TLS has been removed and is disabled by default
89 for DTLS; configure with enable-heartbeats. Code that uses the
90 old #define's might need to be updated.
91 [Emilia Käsper, Rich Salz]
93 *) Rename REF_CHECK to REF_DEBUG.
96 *) New "unified" build system
98 The "unified" build system is aimed to be a common system for all
99 platforms we support. With it comes new support for VMS.
101 This system builds supports building in a different directory tree
102 than the source tree. It produces one Makefile (for unix family
103 or lookalikes), or one descrip.mms (for VMS).
105 The source of information to make the Makefile / descrip.mms is
106 small files called 'build.info', holding the necessary
107 information for each directory with source to compile, and a
108 template in Configurations, like unix-Makefile.tmpl or
111 We rely heavily on the perl module Text::Template.
114 *) Added support for auto-initialisation and de-initialisation of the library.
115 OpenSSL no longer requires explicit init or deinit routines to be called,
116 except in certain circumstances. See the OPENSSL_init_crypto() and
117 OPENSSL_init_ssl() man pages for further information.
120 *) The arguments to the DTLSv1_listen function have changed. Specifically the
121 "peer" argument is now expected to be a BIO_ADDR object.
123 *) Rewrite of BIO networking library. The BIO library lacked consistent
124 support of IPv6, and adding it required some more extensive
125 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
126 which hold all types of addresses and chains of address information.
127 It also introduces a new API, with functions like BIO_socket,
128 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
129 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
130 have been adapted accordingly.
133 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
137 *) CRIME protection: disable compression by default, even if OpenSSL is
138 compiled with zlib enabled. Applications can still enable compression
139 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
140 using the SSL_CONF library to configure compression.
143 *) The signature of the session callback configured with
144 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
145 was explicitly marked as 'const unsigned char*' instead of
149 *) Always DPURIFY. Remove the use of uninitialized memory in the
150 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
153 *) Removed many obsolete configuration items, including
154 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
155 MD2_CHAR, MD2_INT, MD2_LONG
157 IDEA_SHORT, IDEA_LONG
158 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
159 [Rich Salz, with advice from Andy Polyakov]
161 *) Many BN internals have been moved to an internal header file.
162 [Rich Salz with help from Andy Polyakov]
164 *) Configuration and writing out the results from it has changed.
165 Files such as Makefile include/openssl/opensslconf.h and are now
166 produced through general templates, such as Makefile.in and
167 crypto/opensslconf.h.in and some help from the perl module
170 Also, the center of configuration information is no longer
171 Makefile. Instead, Configure produces a perl module in
172 configdata.pm which holds most of the config data (in the hash
173 table %config), the target data that comes from the target
174 configuration in one of the Configurations/*.conf files (in
178 *) To clarify their intended purposes, the Configure options
179 --prefix and --openssldir change their semantics, and become more
180 straightforward and less interdependent.
182 --prefix shall be used exclusively to give the location INSTALLTOP
183 where programs, scripts, libraries, include files and manuals are
184 going to be installed. The default is now /usr/local.
186 --openssldir shall be used exclusively to give the default
187 location OPENSSLDIR where certificates, private keys, CRLs are
188 managed. This is also where the default openssl.cnf gets
190 If the directory given with this option is a relative path, the
191 values of both the --prefix value and the --openssldir value will
192 be combined to become OPENSSLDIR.
193 The default for --openssldir is INSTALLTOP/ssl.
195 Anyone who uses --openssldir to specify where OpenSSL is to be
196 installed MUST change to use --prefix instead.
199 *) The GOST engine was out of date and therefore it has been removed. An up
200 to date GOST engine is now being maintained in an external repository.
201 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
202 support for GOST ciphersuites (these are only activated if a GOST engine
206 *) EGD is no longer supported by default; use enable-egd when
208 [Ben Kaduk and Rich Salz]
210 *) The distribution now has Makefile.in files, which are used to
211 create Makefile's when Configure is run. *Configure must be run
212 before trying to build now.*
215 *) The return value for SSL_CIPHER_description() for error conditions
219 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
221 Obtaining and performing DNSSEC validation of TLSA records is
222 the application's responsibility. The application provides
223 the TLSA records of its choice to OpenSSL, and these are then
224 used to authenticate the peer.
226 The TLSA records need not even come from DNS. They can, for
227 example, be used to implement local end-entity certificate or
228 trust-anchor "pinning", where the "pin" data takes the form
229 of TLSA records, which can augment or replace verification
230 based on the usual WebPKI public certification authorities.
233 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
234 continues to support deprecated interfaces in default builds.
235 However, applications are strongly advised to compile their
236 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
237 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
238 or the 1.1.0 releases.
240 In environments in which all applications have been ported to
241 not use any deprecated interfaces OpenSSL's Configure script
242 should be used with the --api=1.1.0 option to entirely remove
243 support for the deprecated features from the library and
244 unconditionally disable them in the installed headers.
245 Essentially the same effect can be achieved with the "no-deprecated"
246 argument to Configure, except that this will always restrict
247 the build to just the latest API, rather than a fixed API
250 As applications are ported to future revisions of the API,
251 they should update their compile-time OPENSSL_API_COMPAT define
252 accordingly, but in most cases should be able to continue to
253 compile with later releases.
255 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
256 0x10000000L and 0x00908000L, respectively. However those
257 versions did not support the OPENSSL_API_COMPAT feature, and
258 so applications are not typically tested for explicit support
259 of just the undeprecated features of either release.
262 *) Add support for setting the minimum and maximum supported protocol.
263 It can bet set via the SSL_set_min_proto_version() and
264 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
265 MaxProtcol. It's recommended to use the new APIs to disable
266 protocols instead of disabling individual protocols using
267 SSL_set_options() or SSL_CONF's Protocol. This change also
268 removes support for disabling TLS 1.2 in the OpenSSL TLS
269 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
272 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
275 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
276 and integrates ECDSA and ECDH functionality into EC. Implementations can
277 now redirect key generation and no longer need to convert to or from
280 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
281 include the ec.h header file instead.
284 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
285 ciphers who are no longer supported and drops support the ephemeral RSA key
286 exchange. The LOW ciphers currently doesn't have any ciphers in it.
289 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
290 opaque. For HMAC_CTX, the following constructors and destructors
293 HMAC_CTX *HMAC_CTX_new(void);
294 void HMAC_CTX_free(HMAC_CTX *ctx);
296 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
297 destroy such methods has been added. See EVP_MD_meth_new(3) and
298 EVP_CIPHER_meth_new(3) for documentation.
301 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
302 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
303 EVP_MD_CTX_reset() should be called instead to reinitialise
304 an already created structure.
305 2) For consistency with the majority of our object creators and
306 destructors, EVP_MD_CTX_(create|destroy) were renamed to
307 EVP_MD_CTX_(new|free). The old names are retained as macros
308 for deprecated builds.
311 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
312 cryptographic operations to be performed asynchronously as long as an
313 asynchronous capable engine is used. See the ASYNC_start_job() man page for
314 further details. Libssl has also had this capability integrated with the
315 introduction of the new mode SSL_MODE_ASYNC and associated error
316 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
317 pages. This work was developed in partnership with Intel Corp.
320 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
321 always enabled now. If you want to disable the support you should
322 exclude it using the list of supported ciphers. This also means that the
323 "-no_ecdhe" option has been removed from s_server.
326 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
327 SSL_{CTX_}set1_curves() which can set a list.
330 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
331 curve you want to support using SSL_{CTX_}set1_curves().
334 *) State machine rewrite. The state machine code has been significantly
335 refactored in order to remove much duplication of code and solve issues
336 with the old code (see ssl/statem/README for further details). This change
337 does have some associated API changes. Notably the SSL_state() function
338 has been removed and replaced by SSL_get_state which now returns an
339 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
340 altogether. The previous handshake states defined in ssl.h and ssl3.h have
344 *) All instances of the string "ssleay" in the public API were replaced
345 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
346 Some error codes related to internal RSA_eay API's were renamed.
349 *) The demo files in crypto/threads were moved to demo/threads.
352 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
354 [Matt Caswell, Rich Salz]
356 *) New ASN.1 embed macro.
358 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
359 structure is not allocated: it is part of the parent. That is instead of
367 This reduces memory fragmentation and make it impossible to accidentally
368 set a mandatory field to NULL.
370 This currently only works for some fields specifically a SEQUENCE, CHOICE,
371 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
372 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
376 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
379 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
380 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
381 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
382 DES and RC4 ciphersuites.
385 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
386 This changes the decoding behaviour for some invalid messages,
387 though the change is mostly in the more lenient direction, and
388 legacy behaviour is preserved as much as possible.
391 *) Fix no-stdio build.
392 [ David Woodhouse <David.Woodhouse@intel.com> and also
393 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
395 *) New testing framework
396 The testing framework has been largely rewritten and is now using
397 perl and the perl modules Test::Harness and an extended variant of
398 Test::More called OpenSSL::Test to do its work. All test scripts in
399 test/ have been rewritten into test recipes, and all direct calls to
400 executables in test/Makefile have become individual recipes using the
401 simplified testing OpenSSL::Test::Simple.
403 For documentation on our testing modules, do:
405 perldoc test/testlib/OpenSSL/Test/Simple.pm
406 perldoc test/testlib/OpenSSL/Test.pm
410 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
411 are used; the latter aborts on memory leaks (usually checked on exit).
412 Some undocumented "set malloc, etc., hooks" functions were removed
413 and others were changed. All are now documented.
416 *) In DSA_generate_parameters_ex, if the provided seed is too short,
418 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
420 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
421 from RFC4279, RFC4785, RFC5487, RFC5489.
423 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
424 original RSA_PSK patch.
427 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
428 era flag was never set throughout the codebase (only read). Also removed
429 SSL3_FLAGS_POP_BUFFER which was only used if
430 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
433 *) Changed the default name options in the "ca", "crl", "req" and "x509"
434 to be "oneline" instead of "compat".
437 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
438 not aware of clients that still exhibit this bug, and the workaround
439 hasn't been working properly for a while.
442 *) The return type of BIO_number_read() and BIO_number_written() as well as
443 the corresponding num_read and num_write members in the BIO structure has
444 changed from unsigned long to uint64_t. On platforms where an unsigned
445 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
449 *) Given the pervasive nature of TLS extensions it is inadvisable to run
450 OpenSSL without support for them. It also means that maintaining
451 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
452 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
455 *) Removed support for the two export grade static DH ciphersuites
456 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
457 were newly added (along with a number of other static DH ciphersuites) to
458 1.0.2. However the two export ones have *never* worked since they were
459 introduced. It seems strange in any case to be adding new export
460 ciphersuites, and given "logjam" it also does not seem correct to fix them.
463 *) Version negotiation has been rewritten. In particular SSLv23_method(),
464 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
465 and turned into macros which simply call the new preferred function names
466 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
467 should use the new names instead. Also as part of this change the ssl23.h
468 header file has been removed.
471 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
472 code and the associated standard is no longer considered fit-for-purpose.
475 *) RT2547 was closed. When generating a private key, try to make the
476 output file readable only by the owner. This behavior change might
477 be noticeable when interacting with other software.
479 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
483 *) Added HTTP GET support to the ocsp command.
486 *) Changed default digest for the dgst and enc commands from MD5 to
490 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
493 *) Added support for TLS extended master secret from
494 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
495 initial patch which was a great help during development.
498 *) All libssl internal structures have been removed from the public header
499 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
500 now redundant). Users should not attempt to access internal structures
501 directly. Instead they should use the provided API functions.
504 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
505 Access to deprecated functions can be re-enabled by running config with
506 "enable-deprecated". In addition applications wishing to use deprecated
507 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
508 will, by default, disable some transitive includes that previously existed
509 in the header files (e.g. ec.h will no longer, by default, include bn.h)
512 *) Added support for OCB mode. OpenSSL has been granted a patent license
513 compatible with the OpenSSL license for use of OCB. Details are available
514 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
515 for OCB can be removed by calling config with no-ocb.
518 *) SSLv2 support has been removed. It still supports receiving a SSLv2
519 compatible client hello.
522 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
523 done while fixing the error code for the key-too-small case.
524 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
526 *) CA.sh has been removmed; use CA.pl instead.
529 *) Removed old DES API.
532 *) Remove various unsupported platforms:
538 Sinix/ReliantUNIX RM400
543 16-bit platforms such as WIN16
546 *) Clean up OPENSSL_NO_xxx #define's
547 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
548 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
549 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
550 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
551 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
552 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
553 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
554 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
555 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
556 Remove MS_STATIC; it's a relic from platforms <32 bits.
559 *) Cleaned up dead code
560 Remove all but one '#ifdef undef' which is to be looked at.
563 *) Clean up calling of xxx_free routines.
564 Just like free(), fix most of the xxx_free routines to accept
565 NULL. Remove the non-null checks from callers. Save much code.
568 *) Add secure heap for storage of private keys (when possible).
569 Add BIO_s_secmem(), CBIGNUM, etc.
570 Contributed by Akamai Technologies under our Corporate CLA.
573 *) Experimental support for a new, fast, unbiased prime candidate generator,
574 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
575 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
577 *) New output format NSS in the sess_id command line tool. This allows
578 exporting the session id and the master key in NSS keylog format.
579 [Martin Kaiser <martin@kaiser.cx>]
581 *) Harmonize version and its documentation. -f flag is used to display
583 [mancha <mancha1@zoho.com>]
585 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
586 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
587 [mancha <mancha1@zoho.com>]
589 *) Fix some double frees. These are not thought to be exploitable.
590 [mancha <mancha1@zoho.com>]
592 *) A missing bounds check in the handling of the TLS heartbeat extension
593 can be used to reveal up to 64k of memory to a connected client or
596 Thanks for Neel Mehta of Google Security for discovering this bug and to
597 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
598 preparing the fix (CVE-2014-0160)
599 [Adam Langley, Bodo Moeller]
601 *) Fix for the attack described in the paper "Recovering OpenSSL
602 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
603 by Yuval Yarom and Naomi Benger. Details can be obtained from:
604 http://eprint.iacr.org/2014/140
606 Thanks to Yuval Yarom and Naomi Benger for discovering this
607 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
608 [Yuval Yarom and Naomi Benger]
610 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
611 this fixes a limitation in previous versions of OpenSSL.
614 *) Experimental encrypt-then-mac support.
616 Experimental support for encrypt then mac from
617 draft-gutmann-tls-encrypt-then-mac-02.txt
619 To enable it set the appropriate extension number (0x42 for the test
620 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
622 For non-compliant peers (i.e. just about everything) this should have no
625 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
629 *) Add EVP support for key wrapping algorithms, to avoid problems with
630 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
631 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
632 algorithms and include tests cases.
635 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
639 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
640 MGF1 digest and OAEP label.
643 *) Make openssl verify return errors.
644 [Chris Palmer <palmer@google.com> and Ben Laurie]
646 *) New function ASN1_TIME_diff to calculate the difference between two
647 ASN1_TIME structures or one structure and the current time.
650 *) Update fips_test_suite to support multiple command line options. New
651 test to induce all self test errors in sequence and check expected
655 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
656 sign or verify all in one operation.
659 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
660 test programs and fips_test_suite. Includes functionality to parse
661 the minimal script output of fipsalgest.pl directly.
664 *) Add authorisation parameter to FIPS_module_mode_set().
667 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
670 *) Use separate DRBG fields for internal and external flags. New function
671 FIPS_drbg_health_check() to perform on demand health checking. Add
672 generation tests to fips_test_suite with reduced health check interval to
673 demonstrate periodic health checking. Add "nodh" option to
674 fips_test_suite to skip very slow DH test.
677 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
681 *) More extensive health check for DRBG checking many more failure modes.
682 New function FIPS_selftest_drbg_all() to handle every possible DRBG
683 combination: call this in fips_test_suite.
686 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
687 and POST to handle Dual EC cases.
690 *) Add support for canonical generation of DSA parameter 'g'. See
693 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
694 POST to handle HMAC cases.
697 *) Add functions FIPS_module_version() and FIPS_module_version_text()
698 to return numerical and string versions of the FIPS module number.
701 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
702 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
703 outside the validated module in the FIPS capable OpenSSL.
706 *) Minor change to DRBG entropy callback semantics. In some cases
707 there is no multiple of the block length between min_len and
708 max_len. Allow the callback to return more than max_len bytes
709 of entropy but discard any extra: it is the callback's responsibility
710 to ensure that the extra data discarded does not impact the
711 requested amount of entropy.
714 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
715 information in FIPS186-3, SP800-57 and SP800-131A.
718 *) CCM support via EVP. Interface is very similar to GCM case except we
719 must supply all data in one chunk (i.e. no update, final) and the
720 message length must be supplied if AAD is used. Add algorithm test
724 *) Initial version of POST overhaul. Add POST callback to allow the status
725 of POST to be monitored and/or failures induced. Modify fips_test_suite
726 to use callback. Always run all selftests even if one fails.
729 *) XTS support including algorithm test driver in the fips_gcmtest program.
730 Note: this does increase the maximum key length from 32 to 64 bytes but
731 there should be no binary compatibility issues as existing applications
732 will never use XTS mode.
735 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
736 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
737 performs algorithm blocking for unapproved PRNG types. Also do not
738 set PRNG type in FIPS_mode_set(): leave this to the application.
739 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
740 the standard OpenSSL PRNG: set additional data to a date time vector.
743 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
744 This shouldn't present any incompatibility problems because applications
745 shouldn't be using these directly and any that are will need to rethink
746 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
749 *) Extensive self tests and health checking required by SP800-90 DRBG.
750 Remove strength parameter from FIPS_drbg_instantiate and always
751 instantiate at maximum supported strength.
754 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
757 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
760 *) New function DH_compute_key_padded() to compute a DH key and pad with
761 leading zeroes if needed: this complies with SP800-56A et al.
764 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
765 anything, incomplete, subject to change and largely untested at present.
768 *) Modify fipscanisteronly build option to only build the necessary object
769 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
772 *) Add experimental option FIPSSYMS to give all symbols in
773 fipscanister.o and FIPS or fips prefix. This will avoid
774 conflicts with future versions of OpenSSL. Add perl script
775 util/fipsas.pl to preprocess assembly language source files
776 and rename any affected symbols.
779 *) Add selftest checks and algorithm block of non-fips algorithms in
780 FIPS mode. Remove DES2 from selftests.
783 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
784 return internal method without any ENGINE dependencies. Add new
785 tiny fips sign and verify functions.
788 *) New build option no-ec2m to disable characteristic 2 code.
791 *) New build option "fipscanisteronly". This only builds fipscanister.o
792 and (currently) associated fips utilities. Uses the file Makefile.fips
793 instead of Makefile.org as the prototype.
796 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
797 Update fips_gcmtest to use IV generator.
800 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
801 setting output buffer to NULL. The *Final function must be
802 called although it will not retrieve any additional data. The tag
803 can be set or retrieved with a ctrl. The IV length is by default 12
804 bytes (96 bits) but can be set to an alternative value. If the IV
805 length exceeds the maximum IV length (currently 16 bytes) it cannot be
809 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
810 underlying do_cipher function handles all cipher semantics itself
811 including padding and finalisation. This is useful if (for example)
812 an ENGINE cipher handles block padding itself. The behaviour of
813 do_cipher is subtly changed if this flag is set: the return value
814 is the number of characters written to the output buffer (zero is
815 no longer an error code) or a negative error code. Also if the
816 input buffer is NULL and length 0 finalisation should be performed.
819 *) If a candidate issuer certificate is already part of the constructed
820 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
823 *) Improve forward-security support: add functions
825 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
826 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
828 for use by SSL/TLS servers; the callback function will be called whenever a
829 new session is created, and gets to decide whether the session may be
830 cached to make it resumable (return 0) or not (return 1). (As by the
831 SSL/TLS protocol specifications, the session_id sent by the server will be
832 empty to indicate that the session is not resumable; also, the server will
833 not generate RFC 4507 (RFC 5077) session tickets.)
835 A simple reasonable callback implementation is to return is_forward_secure.
836 This parameter will be set to 1 or 0 depending on the ciphersuite selected
837 by the SSL/TLS server library, indicating whether it can provide forward
839 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
841 *) New -verify_name option in command line utilities to set verification
845 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
846 Add CMAC pkey methods.
849 *) Experimental renegotiation in s_server -www mode. If the client
850 browses /reneg connection is renegotiated. If /renegcert it is
851 renegotiated requesting a certificate.
854 *) Add an "external" session cache for debugging purposes to s_server. This
855 should help trace issues which normally are only apparent in deployed
856 multi-process servers.
859 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
860 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
861 BIO_set_cipher() and some obscure PEM functions were changed so they
862 can now return an error. The RAND changes required a change to the
863 RAND_METHOD structure.
866 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
867 a gcc attribute to warn if the result of a function is ignored. This
868 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
869 whose return value is often ignored.
872 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
874 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
875 Builds that are not configured with "enable-weak-ssl-ciphers" will not
876 provide any "EXPORT" or "LOW" strength ciphers.
879 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
880 is by default disabled at build-time. Builds that are not configured with
881 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
882 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
883 will need to explicitly call either of:
885 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
887 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
889 as appropriate. Even if either of those is used, or the application
890 explicitly uses the version-specific SSLv2_method() or its client and
891 server variants, SSLv2 ciphers vulnerable to exhaustive search key
892 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
893 ciphers, and SSLv2 56-bit DES are no longer available.
897 *) Fix a double-free in DSA code
899 A double free bug was discovered when OpenSSL parses malformed DSA private
900 keys and could lead to a DoS attack or memory corruption for applications
901 that receive DSA private keys from untrusted sources. This scenario is
904 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
909 *) Disable SRP fake user seed to address a server memory leak.
911 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
913 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
914 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
915 was changed to ignore the "fake user" SRP seed, even if the seed
918 Users should use SRP_VBASE_get1_by_user instead. Note that in
919 SRP_VBASE_get1_by_user, caller must free the returned value. Note
920 also that even though configuring the SRP seed attempts to hide
921 invalid usernames by continuing the handshake with fake
922 credentials, this behaviour is not constant time and no strong
923 guarantees are made that the handshake is indistinguishable from
924 that of a valid user.
928 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
930 In the BN_hex2bn function the number of hex digits is calculated using an
931 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
932 large values of |i| this can result in |bn_expand| not allocating any
933 memory because |i * 4| is negative. This can leave the internal BIGNUM data
934 field as NULL leading to a subsequent NULL ptr deref. For very large values
935 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
936 In this case memory is allocated to the internal BIGNUM data field, but it
937 is insufficiently sized leading to heap corruption. A similar issue exists
938 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
939 is ever called by user applications with very large untrusted hex/dec data.
940 This is anticipated to be a rare occurrence.
942 All OpenSSL internal usage of these functions use data that is not expected
943 to be untrusted, e.g. config file data or application command line
944 arguments. If user developed applications generate config file data based
945 on untrusted data then it is possible that this could also lead to security
946 consequences. This is also anticipated to be rare.
948 This issue was reported to OpenSSL by Guido Vranken.
952 *) Fix memory issues in BIO_*printf functions
954 The internal |fmtstr| function used in processing a "%s" format string in
955 the BIO_*printf functions could overflow while calculating the length of a
956 string and cause an OOB read when printing very long strings.
958 Additionally the internal |doapr_outch| function can attempt to write to an
959 OOB memory location (at an offset from the NULL pointer) in the event of a
960 memory allocation failure. In 1.0.2 and below this could be caused where
961 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
962 could be in processing a very long "%s" format string. Memory leaks can
965 The first issue may mask the second issue dependent on compiler behaviour.
966 These problems could enable attacks where large amounts of untrusted data
967 is passed to the BIO_*printf functions. If applications use these functions
968 in this way then they could be vulnerable. OpenSSL itself uses these
969 functions when printing out human-readable dumps of ASN.1 data. Therefore
970 applications that print this data could be vulnerable if the data is from
971 untrusted sources. OpenSSL command line applications could also be
972 vulnerable where they print out ASN.1 data, or if untrusted data is passed
973 as command line arguments.
975 Libssl is not considered directly vulnerable. Additionally certificates etc
976 received via remote connections via libssl are also unlikely to be able to
977 trigger these issues because of message size limits enforced within libssl.
979 This issue was reported to OpenSSL Guido Vranken.
983 *) Side channel attack on modular exponentiation
985 A side-channel attack was found which makes use of cache-bank conflicts on
986 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
987 of RSA keys. The ability to exploit this issue is limited as it relies on
988 an attacker who has control of code in a thread running on the same
989 hyper-threaded core as the victim thread which is performing decryptions.
991 This issue was reported to OpenSSL by Yuval Yarom, The University of
992 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
993 Nadia Heninger, University of Pennsylvania with more information at
994 http://cachebleed.info.
998 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
999 if no keysize is specified with default_bits. This fixes an
1000 omission in an earlier change that changed all RSA/DSA key generation
1001 apps to use 2048 bits by default.
1004 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1005 *) DH small subgroups
1007 Historically OpenSSL only ever generated DH parameters based on "safe"
1008 primes. More recently (in version 1.0.2) support was provided for
1009 generating X9.42 style parameter files such as those required for RFC 5114
1010 support. The primes used in such files may not be "safe". Where an
1011 application is using DH configured with parameters based on primes that are
1012 not "safe" then an attacker could use this fact to find a peer's private
1013 DH exponent. This attack requires that the attacker complete multiple
1014 handshakes in which the peer uses the same private DH exponent. For example
1015 this could be used to discover a TLS server's private DH exponent if it's
1016 reusing the private DH exponent or it's using a static DH ciphersuite.
1018 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1019 TLS. It is not on by default. If the option is not set then the server
1020 reuses the same private DH exponent for the life of the server process and
1021 would be vulnerable to this attack. It is believed that many popular
1022 applications do set this option and would therefore not be at risk.
1024 The fix for this issue adds an additional check where a "q" parameter is
1025 available (as is the case in X9.42 based parameters). This detects the
1026 only known attack, and is the only possible defense for static DH
1027 ciphersuites. This could have some performance impact.
1029 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1030 default and cannot be disabled. This could have some performance impact.
1032 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1036 *) SSLv2 doesn't block disabled ciphers
1038 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1039 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1040 been disabled, provided that the SSLv2 protocol was not also disabled via
1043 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1044 and Sebastian Schinzel.
1048 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1050 *) BN_mod_exp may produce incorrect results on x86_64
1052 There is a carry propagating bug in the x86_64 Montgomery squaring
1053 procedure. No EC algorithms are affected. Analysis suggests that attacks
1054 against RSA and DSA as a result of this defect would be very difficult to
1055 perform and are not believed likely. Attacks against DH are considered just
1056 feasible (although very difficult) because most of the work necessary to
1057 deduce information about a private key may be performed offline. The amount
1058 of resources required for such an attack would be very significant and
1059 likely only accessible to a limited number of attackers. An attacker would
1060 additionally need online access to an unpatched system using the target
1061 private key in a scenario with persistent DH parameters and a private
1062 key that is shared between multiple clients. For example this can occur by
1063 default in OpenSSL DHE based SSL/TLS ciphersuites.
1065 This issue was reported to OpenSSL by Hanno Böck.
1069 *) Certificate verify crash with missing PSS parameter
1071 The signature verification routines will crash with a NULL pointer
1072 dereference if presented with an ASN.1 signature using the RSA PSS
1073 algorithm and absent mask generation function parameter. Since these
1074 routines are used to verify certificate signature algorithms this can be
1075 used to crash any certificate verification operation and exploited in a
1076 DoS attack. Any application which performs certificate verification is
1077 vulnerable including OpenSSL clients and servers which enable client
1080 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1084 *) X509_ATTRIBUTE memory leak
1086 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1087 memory. This structure is used by the PKCS#7 and CMS routines so any
1088 application which reads PKCS#7 or CMS data from untrusted sources is
1089 affected. SSL/TLS is not affected.
1091 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1096 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1097 This changes the decoding behaviour for some invalid messages,
1098 though the change is mostly in the more lenient direction, and
1099 legacy behaviour is preserved as much as possible.
1102 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1104 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1106 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1108 *) Alternate chains certificate forgery
1110 During certificate verfification, OpenSSL will attempt to find an
1111 alternative certificate chain if the first attempt to build such a chain
1112 fails. An error in the implementation of this logic can mean that an
1113 attacker could cause certain checks on untrusted certificates to be
1114 bypassed, such as the CA flag, enabling them to use a valid leaf
1115 certificate to act as a CA and "issue" an invalid certificate.
1117 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1121 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1123 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1124 incompatibility in the handling of HMAC. The previous ABI has now been
1128 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1130 *) Malformed ECParameters causes infinite loop
1132 When processing an ECParameters structure OpenSSL enters an infinite loop
1133 if the curve specified is over a specially malformed binary polynomial
1136 This can be used to perform denial of service against any
1137 system which processes public keys, certificate requests or
1138 certificates. This includes TLS clients and TLS servers with
1139 client authentication enabled.
1141 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1145 *) Exploitable out-of-bounds read in X509_cmp_time
1147 X509_cmp_time does not properly check the length of the ASN1_TIME
1148 string and can read a few bytes out of bounds. In addition,
1149 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1152 An attacker can use this to craft malformed certificates and CRLs of
1153 various sizes and potentially cause a segmentation fault, resulting in
1154 a DoS on applications that verify certificates or CRLs. TLS clients
1155 that verify CRLs are affected. TLS clients and servers with client
1156 authentication enabled may be affected if they use custom verification
1159 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1160 independently by Hanno Böck.
1164 *) PKCS7 crash with missing EnvelopedContent
1166 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1167 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1168 with missing content and trigger a NULL pointer dereference on parsing.
1170 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1171 structures from untrusted sources are affected. OpenSSL clients and
1172 servers are not affected.
1174 This issue was reported to OpenSSL by Michal Zalewski (Google).
1178 *) CMS verify infinite loop with unknown hash function
1180 When verifying a signedData message the CMS code can enter an infinite loop
1181 if presented with an unknown hash function OID. This can be used to perform
1182 denial of service against any system which verifies signedData messages using
1184 This issue was reported to OpenSSL by Johannes Bauer.
1188 *) Race condition handling NewSessionTicket
1190 If a NewSessionTicket is received by a multi-threaded client when attempting to
1191 reuse a previous ticket then a race condition can occur potentially leading to
1192 a double free of the ticket data.
1196 *) Only support 256-bit or stronger elliptic curves with the
1197 'ecdh_auto' setting (server) or by default (client). Of supported
1198 curves, prefer P-256 (both).
1201 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1203 *) ClientHello sigalgs DoS fix
1205 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1206 invalid signature algorithms extension a NULL pointer dereference will
1207 occur. This can be exploited in a DoS attack against the server.
1209 This issue was was reported to OpenSSL by David Ramos of Stanford
1212 [Stephen Henson and Matt Caswell]
1214 *) Multiblock corrupted pointer fix
1216 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1217 feature only applies on 64 bit x86 architecture platforms that support AES
1218 NI instructions. A defect in the implementation of "multiblock" can cause
1219 OpenSSL's internal write buffer to become incorrectly set to NULL when
1220 using non-blocking IO. Typically, when the user application is using a
1221 socket BIO for writing, this will only result in a failed connection.
1222 However if some other BIO is used then it is likely that a segmentation
1223 fault will be triggered, thus enabling a potential DoS attack.
1225 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1229 *) Segmentation fault in DTLSv1_listen fix
1231 The DTLSv1_listen function is intended to be stateless and processes the
1232 initial ClientHello from many peers. It is common for user code to loop
1233 over the call to DTLSv1_listen until a valid ClientHello is received with
1234 an associated cookie. A defect in the implementation of DTLSv1_listen means
1235 that state is preserved in the SSL object from one invocation to the next
1236 that can lead to a segmentation fault. Errors processing the initial
1237 ClientHello can trigger this scenario. An example of such an error could be
1238 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1241 This issue was reported to OpenSSL by Per Allansson.
1245 *) Segmentation fault in ASN1_TYPE_cmp fix
1247 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1248 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1249 certificate signature algorithm consistency this can be used to crash any
1250 certificate verification operation and exploited in a DoS attack. Any
1251 application which performs certificate verification is vulnerable including
1252 OpenSSL clients and servers which enable client authentication.
1256 *) Segmentation fault for invalid PSS parameters fix
1258 The signature verification routines will crash with a NULL pointer
1259 dereference if presented with an ASN.1 signature using the RSA PSS
1260 algorithm and invalid parameters. Since these routines are used to verify
1261 certificate signature algorithms this can be used to crash any
1262 certificate verification operation and exploited in a DoS attack. Any
1263 application which performs certificate verification is vulnerable including
1264 OpenSSL clients and servers which enable client authentication.
1266 This issue was was reported to OpenSSL by Brian Carpenter.
1270 *) ASN.1 structure reuse memory corruption fix
1272 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1273 memory corruption via an invalid write. Such reuse is and has been
1274 strongly discouraged and is believed to be rare.
1276 Applications that parse structures containing CHOICE or ANY DEFINED BY
1277 components may be affected. Certificate parsing (d2i_X509 and related
1278 functions) are however not affected. OpenSSL clients and servers are
1283 *) PKCS7 NULL pointer dereferences fix
1285 The PKCS#7 parsing code does not handle missing outer ContentInfo
1286 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1287 missing content and trigger a NULL pointer dereference on parsing.
1289 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1290 otherwise parse PKCS#7 structures from untrusted sources are
1291 affected. OpenSSL clients and servers are not affected.
1293 This issue was reported to OpenSSL by Michal Zalewski (Google).
1297 *) DoS via reachable assert in SSLv2 servers fix
1299 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1300 servers that both support SSLv2 and enable export cipher suites by sending
1301 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1303 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1304 (OpenSSL development team).
1308 *) Empty CKE with client auth and DHE fix
1310 If client auth is used then a server can seg fault in the event of a DHE
1311 ciphersuite being selected and a zero length ClientKeyExchange message
1312 being sent by the client. This could be exploited in a DoS attack.
1316 *) Handshake with unseeded PRNG fix
1318 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1319 with an unseeded PRNG. The conditions are:
1320 - The client is on a platform where the PRNG has not been seeded
1321 automatically, and the user has not seeded manually
1322 - A protocol specific client method version has been used (i.e. not
1323 SSL_client_methodv23)
1324 - A ciphersuite is used that does not require additional random data from
1325 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1327 If the handshake succeeds then the client random that has been used will
1328 have been generated from a PRNG with insufficient entropy and therefore the
1329 output may be predictable.
1331 For example using the following command with an unseeded openssl will
1332 succeed on an unpatched platform:
1334 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1338 *) Use After Free following d2i_ECPrivatekey error fix
1340 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1341 could cause a use after free condition. This, in turn, could cause a double
1342 free in several private key parsing functions (such as d2i_PrivateKey
1343 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1344 for applications that receive EC private keys from untrusted
1345 sources. This scenario is considered rare.
1347 This issue was discovered by the BoringSSL project and fixed in their
1352 *) X509_to_X509_REQ NULL pointer deref fix
1354 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1355 the certificate key is invalid. This function is rarely used in practice.
1357 This issue was discovered by Brian Carpenter.
1361 *) Removed the export ciphers from the DEFAULT ciphers
1364 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1366 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1367 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1368 So far those who have to target multiple plaforms would compromise
1369 and argue that binary targeting say ARMv5 would still execute on
1370 ARMv8. "Universal" build resolves this compromise by providing
1371 near-optimal performance even on newer platforms.
1374 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1375 (other platforms pending).
1376 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1378 *) Add support for the SignedCertificateTimestampList certificate and
1379 OCSP response extensions from RFC6962.
1382 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1383 for corner cases. (Certain input points at infinity could lead to
1384 bogus results, with non-infinity inputs mapped to infinity too.)
1387 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1388 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1389 common cases are optimized and there still is room for further
1390 improvements. Vector Permutation AES for Altivec is also added.
1393 *) Add support for little-endian ppc64 Linux target.
1394 [Marcelo Cerri (IBM)]
1396 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1397 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1398 are optimized and there still is room for further improvements.
1399 Both 32- and 64-bit modes are supported.
1400 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1402 *) Improved ARMv7 NEON support.
1405 *) Support for SPARC Architecture 2011 crypto extensions, first
1406 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1407 SHA256/512, MD5, GHASH and modular exponentiation.
1408 [Andy Polyakov, David Miller]
1410 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1412 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1414 *) Support for new and upcoming Intel processors, including AVX2,
1415 BMI and SHA ISA extensions. This includes additional "stitched"
1416 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1419 This work was sponsored by Intel Corp.
1422 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1423 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1424 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1427 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1428 this fixes a limiation in previous versions of OpenSSL.
1431 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1432 MGF1 digest and OAEP label.
1435 *) Add EVP support for key wrapping algorithms, to avoid problems with
1436 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1437 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1438 algorithms and include tests cases.
1441 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1443 [Douglas E. Engert, Steve Henson]
1445 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1446 difference in days and seconds between two tm or ASN1_TIME structures.
1449 *) Add -rev test option to s_server to just reverse order of characters
1450 received by client and send back to server. Also prints an abbreviated
1451 summary of the connection parameters.
1454 *) New option -brief for s_client and s_server to print out a brief summary
1455 of connection parameters.
1458 *) Add callbacks for arbitrary TLS extensions.
1459 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1461 *) New option -crl_download in several openssl utilities to download CRLs
1462 from CRLDP extension in certificates.
1465 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1468 *) New function X509_CRL_diff to generate a delta CRL from the difference
1469 of two full CRLs. Add support to "crl" utility.
1472 *) New functions to set lookup_crls function and to retrieve
1473 X509_STORE from X509_STORE_CTX.
1476 *) Print out deprecated issuer and subject unique ID fields in
1480 *) Extend OCSP I/O functions so they can be used for simple general purpose
1481 HTTP as well as OCSP. New wrapper function which can be used to download
1482 CRLs using the OCSP API.
1485 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1488 *) SSL_CONF* functions. These provide a common framework for application
1489 configuration using configuration files or command lines.
1492 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1493 message callback and prints the results. Needs compile time option
1494 "enable-ssl-trace". New options to s_client and s_server to enable
1498 *) New ctrl and macro to retrieve supported points extensions.
1499 Print out extension in s_server and s_client.
1502 *) New functions to retrieve certificate signature and signature
1506 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1510 *) New Suite B modes for TLS code. These use and enforce the requirements
1511 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1512 only use Suite B curves. The Suite B modes can be set by using the
1513 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1516 *) New chain verification flags for Suite B levels of security. Check
1517 algorithms are acceptable when flags are set in X509_verify_cert.
1520 *) Make tls1_check_chain return a set of flags indicating checks passed
1521 by a certificate chain. Add additional tests to handle client
1522 certificates: checks for matching certificate type and issuer name
1526 *) If an attempt is made to use a signature algorithm not in the peer
1527 preference list abort the handshake. If client has no suitable
1528 signature algorithms in response to a certificate request do not
1529 use the certificate.
1532 *) If server EC tmp key is not in client preference list abort handshake.
1535 *) Add support for certificate stores in CERT structure. This makes it
1536 possible to have different stores per SSL structure or one store in
1537 the parent SSL_CTX. Include distint stores for certificate chain
1538 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1539 to build and store a certificate chain in CERT structure: returing
1540 an error if the chain cannot be built: this will allow applications
1541 to test if a chain is correctly configured.
1543 Note: if the CERT based stores are not set then the parent SSL_CTX
1544 store is used to retain compatibility with existing behaviour.
1548 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1549 mask based on the current session, check mask when sending client
1550 hello and checking the requested ciphersuite.
1553 *) New ctrls to retrieve and set certificate types in a certificate
1554 request message. Print out received values in s_client. If certificate
1555 types is not set with custom values set sensible values based on
1556 supported signature algorithms.
1559 *) Support for distinct client and server supported signature algorithms.
1562 *) Add certificate callback. If set this is called whenever a certificate
1563 is required by client or server. An application can decide which
1564 certificate chain to present based on arbitrary criteria: for example
1565 supported signature algorithms. Add very simple example to s_server.
1566 This fixes many of the problems and restrictions of the existing client
1567 certificate callback: for example you can now clear an existing
1568 certificate and specify the whole chain.
1571 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1572 the certificate can be used for (if anything). Set valid_flags field
1573 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1574 to have similar checks in it.
1576 Add new "cert_flags" field to CERT structure and include a "strict mode".
1577 This enforces some TLS certificate requirements (such as only permitting
1578 certificate signature algorithms contained in the supported algorithms
1579 extension) which some implementations ignore: this option should be used
1580 with caution as it could cause interoperability issues.
1583 *) Update and tidy signature algorithm extension processing. Work out
1584 shared signature algorithms based on preferences and peer algorithms
1585 and print them out in s_client and s_server. Abort handshake if no
1586 shared signature algorithms.
1589 *) Add new functions to allow customised supported signature algorithms
1590 for SSL and SSL_CTX structures. Add options to s_client and s_server
1594 *) New function SSL_certs_clear() to delete all references to certificates
1595 from an SSL structure. Before this once a certificate had been added
1596 it couldn't be removed.
1599 *) Integrate hostname, email address and IP address checking with certificate
1600 verification. New verify options supporting checking in opensl utility.
1603 *) Fixes and wildcard matching support to hostname and email checking
1604 functions. Add manual page.
1605 [Florian Weimer (Red Hat Product Security Team)]
1607 *) New functions to check a hostname email or IP address against a
1608 certificate. Add options x509 utility to print results of checks against
1612 *) Fix OCSP checking.
1613 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1615 *) Initial experimental support for explicitly trusted non-root CAs.
1616 OpenSSL still tries to build a complete chain to a root but if an
1617 intermediate CA has a trust setting included that is used. The first
1618 setting is used: whether to trust (e.g., -addtrust option to the x509
1622 *) Add -trusted_first option which attempts to find certificates in the
1623 trusted store even if an untrusted chain is also supplied.
1626 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1627 platform support for Linux and Android.
1630 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1633 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1634 When in FIPS mode the approved implementations are used as normal,
1635 when not in FIPS mode the internal unapproved versions are used instead.
1636 This means that the FIPS capable OpenSSL isn't forced to use the
1637 (often lower performance) FIPS implementations outside FIPS mode.
1640 *) Transparently support X9.42 DH parameters when calling
1641 PEM_read_bio_DHparameters. This means existing applications can handle
1642 the new parameter format automatically.
1645 *) Initial experimental support for X9.42 DH parameter format: mainly
1646 to support use of 'q' parameter for RFC5114 parameters.
1649 *) Add DH parameters from RFC5114 including test data to dhtest.
1652 *) Support for automatic EC temporary key parameter selection. If enabled
1653 the most preferred EC parameters are automatically used instead of
1654 hardcoded fixed parameters. Now a server just has to call:
1655 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1656 support ECDH and use the most appropriate parameters.
1659 *) Enhance and tidy EC curve and point format TLS extension code. Use
1660 static structures instead of allocation if default values are used.
1661 New ctrls to set curves we wish to support and to retrieve shared curves.
1662 Print out shared curves in s_server. New options to s_server and s_client
1663 to set list of supported curves.
1666 *) New ctrls to retrieve supported signature algorithms and
1667 supported curve values as an array of NIDs. Extend openssl utility
1668 to print out received values.
1671 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1672 between NIDs and the more common NIST names such as "P-256". Enhance
1673 ecparam utility and ECC method to recognise the NIST names for curves.
1676 *) Enhance SSL/TLS certificate chain handling to support different
1677 chains for each certificate instead of one chain in the parent SSL_CTX.
1680 *) Support for fixed DH ciphersuite client authentication: where both
1681 server and client use DH certificates with common parameters.
1684 *) Support for fixed DH ciphersuites: those requiring DH server
1688 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1690 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1691 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1692 X509_CINF_get_signature were reverted post internal team review.
1694 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1696 *) Build fixes for the Windows and OpenVMS platforms
1697 [Matt Caswell and Richard Levitte]
1699 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1701 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1702 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1703 dereference. This could lead to a Denial Of Service attack. Thanks to
1704 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1708 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1709 dtls1_buffer_record function under certain conditions. In particular this
1710 could occur if an attacker sent repeated DTLS records with the same
1711 sequence number but for the next epoch. The memory leak could be exploited
1712 by an attacker in a Denial of Service attack through memory exhaustion.
1713 Thanks to Chris Mueller for reporting this issue.
1717 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1718 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1719 method would be set to NULL which could later result in a NULL pointer
1720 dereference. Thanks to Frank Schmirler for reporting this issue.
1724 *) Abort handshake if server key exchange message is omitted for ephemeral
1727 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1728 reporting this issue.
1732 *) Remove non-export ephemeral RSA code on client and server. This code
1733 violated the TLS standard by allowing the use of temporary RSA keys in
1734 non-export ciphersuites and could be used by a server to effectively
1735 downgrade the RSA key length used to a value smaller than the server
1736 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1737 INRIA or reporting this issue.
1741 *) Fixed issue where DH client certificates are accepted without verification.
1742 An OpenSSL server will accept a DH certificate for client authentication
1743 without the certificate verify message. This effectively allows a client to
1744 authenticate without the use of a private key. This only affects servers
1745 which trust a client certificate authority which issues certificates
1746 containing DH keys: these are extremely rare and hardly ever encountered.
1747 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1752 *) Ensure that the session ID context of an SSL is updated when its
1753 SSL_CTX is updated via SSL_set_SSL_CTX.
1755 The session ID context is typically set from the parent SSL_CTX,
1756 and can vary with the CTX.
1759 *) Fix various certificate fingerprint issues.
1761 By using non-DER or invalid encodings outside the signed portion of a
1762 certificate the fingerprint can be changed without breaking the signature.
1763 Although no details of the signed portion of the certificate can be changed
1764 this can cause problems with some applications: e.g. those using the
1765 certificate fingerprint for blacklists.
1767 1. Reject signatures with non zero unused bits.
1769 If the BIT STRING containing the signature has non zero unused bits reject
1770 the signature. All current signature algorithms require zero unused bits.
1772 2. Check certificate algorithm consistency.
1774 Check the AlgorithmIdentifier inside TBS matches the one in the
1775 certificate signature. NB: this will result in signature failure
1776 errors for some broken certificates.
1778 Thanks to Konrad Kraszewski from Google for reporting this issue.
1780 3. Check DSA/ECDSA signatures use DER.
1782 Reencode DSA/ECDSA signatures and compare with the original received
1783 signature. Return an error if there is a mismatch.
1785 This will reject various cases including garbage after signature
1786 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1787 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1788 (negative or with leading zeroes).
1790 Further analysis was conducted and fixes were developed by Stephen Henson
1791 of the OpenSSL core team.
1796 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1797 results on some platforms, including x86_64. This bug occurs at random
1798 with a very low probability, and is not known to be exploitable in any
1799 way, though its exact impact is difficult to determine. Thanks to Pieter
1800 Wuille (Blockstream) who reported this issue and also suggested an initial
1801 fix. Further analysis was conducted by the OpenSSL development team and
1802 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1803 the OpenSSL core team.
1807 *) Do not resume sessions on the server if the negotiated protocol
1808 version does not match the session's version. Resuming with a different
1809 version, while not strictly forbidden by the RFC, is of questionable
1810 sanity and breaks all known clients.
1811 [David Benjamin, Emilia Käsper]
1813 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1814 early CCS messages during renegotiation. (Note that because
1815 renegotiation is encrypted, this early CCS was not exploitable.)
1818 *) Tighten client-side session ticket handling during renegotiation:
1819 ensure that the client only accepts a session ticket if the server sends
1820 the extension anew in the ServerHello. Previously, a TLS client would
1821 reuse the old extension state and thus accept a session ticket if one was
1822 announced in the initial ServerHello.
1824 Similarly, ensure that the client requires a session ticket if one
1825 was advertised in the ServerHello. Previously, a TLS client would
1826 ignore a missing NewSessionTicket message.
1829 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1831 *) SRTP Memory Leak.
1833 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1834 sends a carefully crafted handshake message, to cause OpenSSL to fail
1835 to free up to 64k of memory causing a memory leak. This could be
1836 exploited in a Denial Of Service attack. This issue affects OpenSSL
1837 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1838 whether SRTP is used or configured. Implementations of OpenSSL that
1839 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1841 The fix was developed by the OpenSSL team.
1845 *) Session Ticket Memory Leak.
1847 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1848 integrity of that ticket is first verified. In the event of a session
1849 ticket integrity check failing, OpenSSL will fail to free memory
1850 causing a memory leak. By sending a large number of invalid session
1851 tickets an attacker could exploit this issue in a Denial Of Service
1856 *) Build option no-ssl3 is incomplete.
1858 When OpenSSL is configured with "no-ssl3" as a build option, servers
1859 could accept and complete a SSL 3.0 handshake, and clients could be
1860 configured to send them.
1862 [Akamai and the OpenSSL team]
1864 *) Add support for TLS_FALLBACK_SCSV.
1865 Client applications doing fallback retries should call
1866 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1868 [Adam Langley, Bodo Moeller]
1870 *) Add additional DigestInfo checks.
1872 Reencode DigestInto in DER and check against the original when
1873 verifying RSA signature: this will reject any improperly encoded
1874 DigestInfo structures.
1876 Note: this is a precautionary measure and no attacks are currently known.
1880 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1882 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1883 SRP code can be overrun an internal buffer. Add sanity check that
1884 g, A, B < N to SRP code.
1886 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1887 Group for discovering this issue.
1891 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1892 TLS 1.0 instead of higher protocol versions when the ClientHello message
1893 is badly fragmented. This allows a man-in-the-middle attacker to force a
1894 downgrade to TLS 1.0 even if both the server and the client support a
1895 higher protocol version, by modifying the client's TLS records.
1897 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1898 researching this issue.
1902 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1903 to a denial of service attack. A malicious server can crash the client
1904 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1905 ciphersuite and sending carefully crafted handshake messages.
1907 Thanks to Felix Gröbert (Google) for discovering and researching this
1912 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1913 to leak memory. This can be exploited through a Denial of Service attack.
1914 Thanks to Adam Langley for discovering and researching this issue.
1918 *) An attacker can force openssl to consume large amounts of memory whilst
1919 processing DTLS handshake messages. This can be exploited through a
1920 Denial of Service attack.
1921 Thanks to Adam Langley for discovering and researching this issue.
1925 *) An attacker can force an error condition which causes openssl to crash
1926 whilst processing DTLS packets due to memory being freed twice. This
1927 can be exploited through a Denial of Service attack.
1928 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1933 *) If a multithreaded client connects to a malicious server using a resumed
1934 session and the server sends an ec point format extension it could write
1935 up to 255 bytes to freed memory.
1937 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1942 *) A malicious server can crash an OpenSSL client with a null pointer
1943 dereference (read) by specifying an SRP ciphersuite even though it was not
1944 properly negotiated with the client. This can be exploited through a
1945 Denial of Service attack.
1947 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1948 discovering and researching this issue.
1952 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1953 X509_name_oneline, X509_name_print_ex et al. to leak some information
1954 from the stack. Applications may be affected if they echo pretty printing
1955 output to the attacker.
1957 Thanks to Ivan Fratric (Google) for discovering this issue.
1959 [Emilia Käsper, and Steve Henson]
1961 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1962 for corner cases. (Certain input points at infinity could lead to
1963 bogus results, with non-infinity inputs mapped to infinity too.)
1966 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1968 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1969 handshake can force the use of weak keying material in OpenSSL
1970 SSL/TLS clients and servers.
1972 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1973 researching this issue. (CVE-2014-0224)
1974 [KIKUCHI Masashi, Steve Henson]
1976 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1977 OpenSSL DTLS client the code can be made to recurse eventually crashing
1980 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1982 [Imre Rad, Steve Henson]
1984 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1985 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1986 client or server. This is potentially exploitable to run arbitrary
1987 code on a vulnerable client or server.
1989 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1990 [Jüri Aedla, Steve Henson]
1992 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1993 are subject to a denial of service attack.
1995 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1996 this issue. (CVE-2014-3470)
1997 [Felix Gröbert, Ivan Fratric, Steve Henson]
1999 *) Harmonize version and its documentation. -f flag is used to display
2001 [mancha <mancha1@zoho.com>]
2003 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2004 in i2d_ECPrivateKey.
2005 [mancha <mancha1@zoho.com>]
2007 *) Fix some double frees. These are not thought to be exploitable.
2008 [mancha <mancha1@zoho.com>]
2010 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2012 *) A missing bounds check in the handling of the TLS heartbeat extension
2013 can be used to reveal up to 64k of memory to a connected client or
2016 Thanks for Neel Mehta of Google Security for discovering this bug and to
2017 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2018 preparing the fix (CVE-2014-0160)
2019 [Adam Langley, Bodo Moeller]
2021 *) Fix for the attack described in the paper "Recovering OpenSSL
2022 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2023 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2024 http://eprint.iacr.org/2014/140
2026 Thanks to Yuval Yarom and Naomi Benger for discovering this
2027 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2028 [Yuval Yarom and Naomi Benger]
2030 *) TLS pad extension: draft-agl-tls-padding-03
2032 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2033 TLS client Hello record length value would otherwise be > 255 and
2034 less that 512 pad with a dummy extension containing zeroes so it
2035 is at least 512 bytes long.
2037 [Adam Langley, Steve Henson]
2039 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2041 *) Fix for TLS record tampering bug. A carefully crafted invalid
2042 handshake could crash OpenSSL with a NULL pointer exception.
2043 Thanks to Anton Johansson for reporting this issues.
2046 *) Keep original DTLS digest and encryption contexts in retransmission
2047 structures so we can use the previous session parameters if they need
2048 to be resent. (CVE-2013-6450)
2051 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2052 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2053 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2054 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2055 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2056 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2057 [Rob Stradling, Adam Langley]
2059 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2061 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2062 supporting platforms or when small records were transferred.
2063 [Andy Polyakov, Steve Henson]
2065 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2067 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2069 This addresses the flaw in CBC record processing discovered by
2070 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2071 at: http://www.isg.rhul.ac.uk/tls/
2073 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2074 Security Group at Royal Holloway, University of London
2075 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2076 Emilia Käsper for the initial patch.
2078 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2080 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2081 ciphersuites which can be exploited in a denial of service attack.
2082 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2083 and detecting this bug and to Wolfgang Ettlinger
2084 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2088 *) Return an error when checking OCSP signatures when key is NULL.
2089 This fixes a DoS attack. (CVE-2013-0166)
2092 *) Make openssl verify return errors.
2093 [Chris Palmer <palmer@google.com> and Ben Laurie]
2095 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2096 the right response is stapled. Also change SSL_get_certificate()
2097 so it returns the certificate actually sent.
2098 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2099 [Rob Stradling <rob.stradling@comodo.com>]
2101 *) Fix possible deadlock when decoding public keys.
2104 *) Don't use TLS 1.0 record version number in initial client hello
2108 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2110 *) Sanity check record length before skipping explicit IV in TLS
2111 1.2, 1.1 and DTLS to fix DoS attack.
2113 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2114 fuzzing as a service testing platform.
2118 *) Initialise tkeylen properly when encrypting CMS messages.
2119 Thanks to Solar Designer of Openwall for reporting this issue.
2122 *) In FIPS mode don't try to use composite ciphers as they are not
2126 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2128 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2129 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2130 mean any application compiled against OpenSSL 1.0.0 headers setting
2131 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2132 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2133 0x10000000L Any application which was previously compiled against
2134 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2135 will need to be recompiled as a result. Letting be results in
2136 inability to disable specifically TLS 1.1 and in client context,
2137 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2140 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2141 disable just protocol X, but all protocols above X *if* there are
2142 protocols *below* X still enabled. In more practical terms it means
2143 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2144 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2145 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2149 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2151 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2152 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2153 in CRYPTO_realloc_clean.
2155 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2156 issue and to Adam Langley <agl@chromium.org> for fixing it.
2158 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2160 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2163 *) Workarounds for some broken servers that "hang" if a client hello
2164 record length exceeds 255 bytes.
2166 1. Do not use record version number > TLS 1.0 in initial client
2167 hello: some (but not all) hanging servers will now work.
2168 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2169 the number of ciphers sent in the client hello. This should be
2170 set to an even number, such as 50, for example by passing:
2171 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2172 Most broken servers should now work.
2173 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2174 TLS 1.2 client support entirely.
2177 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2180 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2182 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2183 STRING form instead of a DigestInfo.
2186 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2187 and the RSA_sign/RSA_verify functions. This was made more apparent when
2188 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2189 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2190 the correct format in RSA_verify so both forms transparently work.
2193 *) Some servers which support TLS 1.0 can choke if we initially indicate
2194 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2195 encrypted premaster secret. As a workaround use the maximum permitted
2196 client version in client hello, this should keep such servers happy
2197 and still work with previous versions of OpenSSL.
2200 *) Add support for TLS/DTLS heartbeats.
2201 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2203 *) Add support for SCTP.
2204 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2206 *) Improved PRNG seeding for VOS.
2207 [Paul Green <Paul.Green@stratus.com>]
2209 *) Extensive assembler packs updates, most notably:
2211 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2212 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2213 - x86_64: bit-sliced AES implementation;
2214 - ARM: NEON support, contemporary platforms optimizations;
2215 - s390x: z196 support;
2216 - *: GHASH and GF(2^m) multiplication implementations;
2220 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2221 (removal of unnecessary code)
2222 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2224 *) Add TLS key material exporter from RFC 5705.
2227 *) Add DTLS-SRTP negotiation from RFC 5764.
2230 *) Add Next Protocol Negotiation,
2231 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2232 disabled with a no-npn flag to config or Configure. Code donated
2234 [Adam Langley <agl@google.com> and Ben Laurie]
2236 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2237 NIST-P256, NIST-P521, with constant-time single point multiplication on
2238 typical inputs. Compiler support for the nonstandard type __uint128_t is
2239 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2240 Code made available under Apache License version 2.0.
2242 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2243 line to include this in your build of OpenSSL, and run "make depend" (or
2244 "make update"). This enables the following EC_METHODs:
2246 EC_GFp_nistp224_method()
2247 EC_GFp_nistp256_method()
2248 EC_GFp_nistp521_method()
2250 EC_GROUP_new_by_curve_name() will automatically use these (while
2251 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2253 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2255 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2256 all platforms. Move ssize_t definition from e_os.h to the public
2257 header file e_os2.h as it now appears in public header file cms.h
2260 *) New -sigopt option to the ca, req and x509 utilities. Additional
2261 signature parameters can be passed using this option and in
2265 *) Add RSA PSS signing function. This will generate and set the
2266 appropriate AlgorithmIdentifiers for PSS based on those in the
2267 corresponding EVP_MD_CTX structure. No application support yet.
2270 *) Support for companion algorithm specific ASN1 signing routines.
2271 New function ASN1_item_sign_ctx() signs a pre-initialised
2272 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2273 the appropriate parameters.
2276 *) Add new algorithm specific ASN1 verification initialisation function
2277 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2278 handling will be the same no matter what EVP_PKEY_METHOD is used.
2279 Add a PSS handler to support verification of PSS signatures: checked
2280 against a number of sample certificates.
2283 *) Add signature printing for PSS. Add PSS OIDs.
2284 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2286 *) Add algorithm specific signature printing. An individual ASN1 method
2287 can now print out signatures instead of the standard hex dump.
2289 More complex signatures (e.g. PSS) can print out more meaningful
2290 information. Include DSA version that prints out the signature
2294 *) Password based recipient info support for CMS library: implementing
2298 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2299 neatly separates the code into cipher and PBE sections and is required
2300 for some algorithms that split PBES2 into separate pieces (such as
2301 password based CMS).
2304 *) Session-handling fixes:
2305 - Fix handling of connections that are resuming with a session ID,
2306 but also support Session Tickets.
2307 - Fix a bug that suppressed issuing of a new ticket if the client
2308 presented a ticket with an expired session.
2309 - Try to set the ticket lifetime hint to something reasonable.
2310 - Make tickets shorter by excluding irrelevant information.
2311 - On the client side, don't ignore renewed tickets.
2312 [Adam Langley, Bodo Moeller (Google)]
2314 *) Fix PSK session representation.
2317 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2319 This work was sponsored by Intel.
2322 *) Add GCM support to TLS library. Some custom code is needed to split
2323 the IV between the fixed (from PRF) and explicit (from TLS record)
2324 portions. This adds all GCM ciphersuites supported by RFC5288 and
2325 RFC5289. Generalise some AES* cipherstrings to include GCM and
2326 add a special AESGCM string for GCM only.
2329 *) Expand range of ctrls for AES GCM. Permit setting invocation
2330 field on decrypt and retrieval of invocation field only on encrypt.
2333 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2334 As required by RFC5289 these ciphersuites cannot be used if for
2335 versions of TLS earlier than 1.2.
2338 *) For FIPS capable OpenSSL interpret a NULL default public key method
2339 as unset and return the appropriate default but do *not* set the default.
2340 This means we can return the appropriate method in applications that
2341 switch between FIPS and non-FIPS modes.
2344 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2345 ENGINE is used then we cannot handle that in the FIPS module so we
2346 keep original code iff non-FIPS operations are allowed.
2349 *) Add -attime option to openssl utilities.
2350 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2352 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2355 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2356 FIPS EC methods unconditionally for now.
2359 *) New build option no-ec2m to disable characteristic 2 code.
2362 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2363 all cases can be covered as some introduce binary incompatibilities.
2366 *) Redirect RSA operations to FIPS module including keygen,
2367 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2370 *) Add similar low level API blocking to ciphers.
2373 *) Low level digest APIs are not approved in FIPS mode: any attempt
2374 to use these will cause a fatal error. Applications that *really* want
2375 to use them can use the private_* version instead.
2378 *) Redirect cipher operations to FIPS module for FIPS builds.
2381 *) Redirect digest operations to FIPS module for FIPS builds.
2384 *) Update build system to add "fips" flag which will link in fipscanister.o
2385 for static and shared library builds embedding a signature if needed.
2388 *) Output TLS supported curves in preference order instead of numerical
2389 order. This is currently hardcoded for the highest order curves first.
2390 This should be configurable so applications can judge speed vs strength.
2393 *) Add TLS v1.2 server support for client authentication.
2396 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2400 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2401 FIPS modules versions.
2404 *) Add TLS v1.2 client side support for client authentication. Keep cache
2405 of handshake records longer as we don't know the hash algorithm to use
2406 until after the certificate request message is received.
2409 *) Initial TLS v1.2 client support. Add a default signature algorithms
2410 extension including all the algorithms we support. Parse new signature
2411 format in client key exchange. Relax some ECC signing restrictions for
2412 TLS v1.2 as indicated in RFC5246.
2415 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2416 to new signature format when needed using client digest preference.
2417 All server ciphersuites should now work correctly in TLS v1.2. No client
2418 support yet and no support for client certificates.
2421 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2422 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2423 ciphersuites. At present only RSA key exchange ciphersuites work with
2424 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2425 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2426 and version checking.
2429 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2430 with this defined it will not be affected by any changes to ssl internal
2431 structures. Add several utility functions to allow openssl application
2432 to work with OPENSSL_NO_SSL_INTERN defined.
2436 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2438 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2441 *) Permit abbreviated handshakes when renegotiating using the function
2442 SSL_renegotiate_abbreviated().
2443 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2445 *) Add call to ENGINE_register_all_complete() to
2446 ENGINE_load_builtin_engines(), so some implementations get used
2447 automatically instead of needing explicit application support.
2450 *) Add support for TLS key exporter as described in RFC5705.
2451 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2453 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2454 a few changes are required:
2456 Add SSL_OP_NO_TLSv1_1 flag.
2457 Add TLSv1_1 methods.
2458 Update version checking logic to handle version 1.1.
2459 Add explicit IV handling (ported from DTLS code).
2460 Add command line options to s_client/s_server.
2463 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2465 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2466 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2467 content decryption and always return the same error. Note: this attack
2468 needs on average 2^20 messages so it only affects automated senders. The
2469 old behaviour can be reenabled in the CMS code by setting the
2470 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2471 an MMA defence is not necessary.
2472 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2473 this issue. (CVE-2012-0884)
2476 *) Fix CVE-2011-4619: make sure we really are receiving a
2477 client hello before rejecting multiple SGC restarts. Thanks to
2478 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2481 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2483 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2484 Thanks to Antonio Martin, Enterprise Secure Access Research and
2485 Development, Cisco Systems, Inc. for discovering this bug and
2486 preparing a fix. (CVE-2012-0050)
2489 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2491 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2492 of the Vaudenay padding oracle attack on CBC mode encryption
2493 which enables an efficient plaintext recovery attack against
2494 the OpenSSL implementation of DTLS. Their attack exploits timing
2495 differences arising during decryption processing. A research
2496 paper describing this attack can be found at:
2497 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2498 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2499 Security Group at Royal Holloway, University of London
2500 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2501 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2502 for preparing the fix. (CVE-2011-4108)
2503 [Robin Seggelmann, Michael Tuexen]
2505 *) Clear bytes used for block padding of SSL 3.0 records.
2507 [Adam Langley (Google)]
2509 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2510 Kadianakis <desnacked@gmail.com> for discovering this issue and
2511 Adam Langley for preparing the fix. (CVE-2011-4619)
2512 [Adam Langley (Google)]
2514 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2515 [Andrey Kulikov <amdeich@gmail.com>]
2517 *) Prevent malformed RFC3779 data triggering an assertion failure.
2518 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2519 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2520 [Rob Austein <sra@hactrn.net>]
2522 *) Improved PRNG seeding for VOS.
2523 [Paul Green <Paul.Green@stratus.com>]
2525 *) Fix ssl_ciph.c set-up race.
2526 [Adam Langley (Google)]
2528 *) Fix spurious failures in ecdsatest.c.
2529 [Emilia Käsper (Google)]
2531 *) Fix the BIO_f_buffer() implementation (which was mixing different
2532 interpretations of the '..._len' fields).
2533 [Adam Langley (Google)]
2535 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2536 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2537 threads won't reuse the same blinding coefficients.
2539 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2540 lock to call BN_BLINDING_invert_ex, and avoids one use of
2541 BN_BLINDING_update for each BN_BLINDING structure (previously,
2542 the last update always remained unused).
2543 [Emilia Käsper (Google)]
2545 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2546 [Bob Buckholz (Google)]
2548 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2550 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2551 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2552 [Kaspar Brand <ossl@velox.ch>]
2554 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2555 for multi-threaded use of ECDH. (CVE-2011-3210)
2556 [Adam Langley (Google)]
2558 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2561 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2562 signature public key algorithm by using OID xref utilities instead.
2563 Before this you could only use some ECC ciphersuites with SHA1 only.
2566 *) Add protection against ECDSA timing attacks as mentioned in the paper
2567 by Billy Bob Brumley and Nicola Tuveri, see:
2569 http://eprint.iacr.org/2011/232.pdf
2571 [Billy Bob Brumley and Nicola Tuveri]
2573 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2575 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2576 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2578 *) Fix bug in string printing code: if *any* escaping is enabled we must
2579 escape the escape character (backslash) or the resulting string is
2583 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2585 *) Disable code workaround for ancient and obsolete Netscape browsers
2586 and servers: an attacker can use it in a ciphersuite downgrade attack.
2587 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2590 *) Fixed J-PAKE implementation error, originally discovered by
2591 Sebastien Martini, further info and confirmation from Stefan
2592 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2595 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2597 *) Fix extension code to avoid race conditions which can result in a buffer
2598 overrun vulnerability: resumed sessions must not be modified as they can
2599 be shared by multiple threads. CVE-2010-3864
2602 *) Fix WIN32 build system to correctly link an ENGINE directory into
2606 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2608 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2610 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2612 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2614 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2615 context. The operation can be customised via the ctrl mechanism in
2616 case ENGINEs want to include additional functionality.
2619 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2622 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2623 output hashes compatible with older versions of OpenSSL.
2624 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2626 *) Fix compression algorithm handling: if resuming a session use the
2627 compression algorithm of the resumed session instead of determining
2628 it from client hello again. Don't allow server to change algorithm.
2631 *) Add load_crls() function to apps tidying load_certs() too. Add option
2632 to verify utility to allow additional CRLs to be included.
2635 *) Update OCSP request code to permit adding custom headers to the request:
2636 some responders need this.
2639 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2641 [Julia Lawall <julia@diku.dk>]
2643 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2644 needlessly dereferenced structures, used obsolete functions and
2645 didn't handle all updated verify codes correctly.
2648 *) Disable MD2 in the default configuration.
2651 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2652 indicate the initial BIO being pushed or popped. This makes it possible
2653 to determine whether the BIO is the one explicitly called or as a result
2654 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2655 it handles reference counts correctly and doesn't zero out the I/O bio
2656 when it is not being explicitly popped. WARNING: applications which
2657 included workarounds for the old buggy behaviour will need to be modified
2658 or they could free up already freed BIOs.
2661 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2662 renaming to all platforms (within the 0.9.8 branch, this was
2663 done conditionally on Netware platforms to avoid a name clash).
2664 [Guenter <lists@gknw.net>]
2666 *) Add ECDHE and PSK support to DTLS.
2667 [Michael Tuexen <tuexen@fh-muenster.de>]
2669 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2673 *) Add "missing" function EVP_MD_flags() (without this the only way to
2674 retrieve a digest flags is by accessing the structure directly. Update
2675 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2676 or cipher is registered as in the "from" argument. Print out all
2677 registered digests in the dgst usage message instead of manually
2678 attempting to work them out.
2681 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2682 this allows the use of compression and extensions. Change default cipher
2683 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2684 by default unless an application cipher string requests it.
2687 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2688 key ids to find matching certificates and keys but some PKCS#12 files
2689 don't follow the (somewhat unwritten) rules and this strategy fails.
2690 Now just gather all certificates together and the first private key
2691 then look for the first certificate that matches the key.
2694 *) Support use of registered digest and cipher names for dgst and cipher
2695 commands instead of having to add each one as a special case. So now
2702 openssl dgst -sha256 foo
2704 and this works for ENGINE based algorithms too.
2708 *) Update Gost ENGINE to support parameter files.
2709 [Victor B. Wagner <vitus@cryptocom.ru>]
2711 *) Support GeneralizedTime in ca utility.
2712 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2714 *) Enhance the hash format used for certificate directory links. The new
2715 form uses the canonical encoding (meaning equivalent names will work
2716 even if they aren't identical) and uses SHA1 instead of MD5. This form
2717 is incompatible with the older format and as a result c_rehash should
2718 be used to rebuild symbolic links.
2721 *) Make PKCS#8 the default write format for private keys, replacing the
2722 traditional format. This form is standardised, more secure and doesn't
2723 include an implicit MD5 dependency.
2726 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2727 committed to OpenSSL should pass this lot as a minimum.
2730 *) Add session ticket override functionality for use by EAP-FAST.
2731 [Jouni Malinen <j@w1.fi>]
2733 *) Modify HMAC functions to return a value. Since these can be implemented
2734 in an ENGINE errors can occur.
2737 *) Type-checked OBJ_bsearch_ex.
2740 *) Type-checked OBJ_bsearch. Also some constification necessitated
2741 by type-checking. Still to come: TXT_DB, bsearch(?),
2742 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2746 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2747 seconds to a tm structure directly, instead of going through OS
2748 specific date routines. This avoids any issues with OS routines such
2749 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2750 and X509_time_adj_ex() to cover the extended range. The existing
2751 X509_time_adj() is still usable and will no longer have any date issues.
2754 *) Delta CRL support. New use deltas option which will attempt to locate
2755 and search any appropriate delta CRLs available.
2757 This work was sponsored by Google.
2760 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2761 code and add additional score elements. Validate alternate CRL paths
2762 as part of the CRL checking and indicate a new error "CRL path validation
2763 error" in this case. Applications wanting additional details can use
2764 the verify callback and check the new "parent" field. If this is not
2765 NULL CRL path validation is taking place. Existing applications wont
2766 see this because it requires extended CRL support which is off by
2769 This work was sponsored by Google.
2772 *) Support for freshest CRL extension.
2774 This work was sponsored by Google.
2777 *) Initial indirect CRL support. Currently only supported in the CRLs
2778 passed directly and not via lookup. Process certificate issuer
2779 CRL entry extension and lookup CRL entries by bother issuer name
2780 and serial number. Check and process CRL issuer entry in IDP extension.
2782 This work was sponsored by Google.
2785 *) Add support for distinct certificate and CRL paths. The CRL issuer
2786 certificate is validated separately in this case. Only enabled if
2787 an extended CRL support flag is set: this flag will enable additional
2788 CRL functionality in future.
2790 This work was sponsored by Google.
2793 *) Add support for policy mappings extension.
2795 This work was sponsored by Google.
2798 *) Fixes to pathlength constraint, self issued certificate handling,
2799 policy processing to align with RFC3280 and PKITS tests.
2801 This work was sponsored by Google.
2804 *) Support for name constraints certificate extension. DN, email, DNS
2805 and URI types are currently supported.
2807 This work was sponsored by Google.
2810 *) To cater for systems that provide a pointer-based thread ID rather
2811 than numeric, deprecate the current numeric thread ID mechanism and
2812 replace it with a structure and associated callback type. This
2813 mechanism allows a numeric "hash" to be extracted from a thread ID in
2814 either case, and on platforms where pointers are larger than 'long',
2815 mixing is done to help ensure the numeric 'hash' is usable even if it
2816 can't be guaranteed unique. The default mechanism is to use "&errno"
2817 as a pointer-based thread ID to distinguish between threads.
2819 Applications that want to provide their own thread IDs should now use
2820 CRYPTO_THREADID_set_callback() to register a callback that will call
2821 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2823 Note that ERR_remove_state() is now deprecated, because it is tied
2824 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2825 to free the current thread's error state should be replaced by
2826 ERR_remove_thread_state(NULL).
2828 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2829 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2830 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2831 application was previously providing a numeric thread callback that
2832 was inappropriate for distinguishing threads, then uniqueness might
2833 have been obtained with &errno that happened immediately in the
2834 intermediate development versions of OpenSSL; this is no longer the
2835 case, the numeric thread callback will now override the automatic use
2837 [Geoff Thorpe, with help from Bodo Moeller]
2839 *) Initial support for different CRL issuing certificates. This covers a
2840 simple case where the self issued certificates in the chain exist and
2841 the real CRL issuer is higher in the existing chain.
2843 This work was sponsored by Google.
2846 *) Removed effectively defunct crypto/store from the build.
2849 *) Revamp of STACK to provide stronger type-checking. Still to come:
2850 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2851 ASN1_STRING, CONF_VALUE.
2854 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2855 RAM on SSL connections. This option can save about 34k per idle SSL.
2858 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2859 STACK, TXT_DB, bsearch, qsort.
2862 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2863 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2864 support for data, signedData, compressedData, digestedData and
2865 encryptedData, envelopedData types included. Scripts to check against
2866 RFC4134 examples draft and interop and consistency checks of many
2867 content types and variants.
2870 *) Add options to enc utility to support use of zlib compression BIO.
2873 *) Extend mk1mf to support importing of options and assembly language
2874 files from Configure script, currently only included in VC-WIN32.
2875 The assembly language rules can now optionally generate the source
2876 files from the associated perl scripts.
2879 *) Implement remaining functionality needed to support GOST ciphersuites.
2880 Interop testing has been performed using CryptoPro implementations.
2881 [Victor B. Wagner <vitus@cryptocom.ru>]
2883 *) s390x assembler pack.
2886 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2890 *) Implement Opaque PRF Input TLS extension as specified in
2891 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2892 official specification yet and no extension type assignment by
2893 IANA exists, this extension (for now) will have to be explicitly
2894 enabled when building OpenSSL by providing the extension number
2895 to use. For example, specify an option
2897 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2899 to the "config" or "Configure" script to enable the extension,
2900 assuming extension number 0x9527 (which is a completely arbitrary
2901 and unofficial assignment based on the MD5 hash of the Internet
2902 Draft). Note that by doing so, you potentially lose
2903 interoperability with other TLS implementations since these might
2904 be using the same extension number for other purposes.
2906 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2907 opaque PRF input value to use in the handshake. This will create
2908 an interal copy of the length-'len' string at 'src', and will
2909 return non-zero for success.
2911 To get more control and flexibility, provide a callback function
2914 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2915 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2919 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2922 Callback function 'cb' will be called in handshakes, and is
2923 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2924 Argument 'arg' is for application purposes (the value as given to
2925 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2926 be provided to the callback function). The callback function
2927 has to return non-zero to report success: usually 1 to use opaque
2928 PRF input just if possible, or 2 to enforce use of the opaque PRF
2929 input. In the latter case, the library will abort the handshake
2930 if opaque PRF input is not successfully negotiated.
2932 Arguments 'peerinput' and 'len' given to the callback function
2933 will always be NULL and 0 in the case of a client. A server will
2934 see the client's opaque PRF input through these variables if
2935 available (NULL and 0 otherwise). Note that if the server
2936 provides an opaque PRF input, the length must be the same as the
2937 length of the client's opaque PRF input.
2939 Note that the callback function will only be called when creating
2940 a new session (session resumption can resume whatever was
2941 previously negotiated), and will not be called in SSL 2.0
2942 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2943 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2944 for applications that need to enforce opaque PRF input.
2948 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2951 [Victor B. Wagner <vitus@cryptocom.ru>]
2953 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2954 RFC4507bis. The encrypted ticket format is an encrypted encoded
2955 SSL_SESSION structure, that way new session features are automatically
2958 If a client application caches session in an SSL_SESSION structure
2959 support is transparent because tickets are now stored in the encoded
2962 The SSL_CTX structure automatically generates keys for ticket
2963 protection in servers so again support should be possible
2964 with no application modification.
2966 If a client or server wishes to disable RFC4507 support then the option
2967 SSL_OP_NO_TICKET can be set.
2969 Add a TLS extension debugging callback to allow the contents of any client
2970 or server extensions to be examined.
2972 This work was sponsored by Google.
2975 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2976 OpenSSL should now compile cleanly on gcc 4.2
2977 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2979 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2980 support including streaming MAC support: this is required for GOST
2981 ciphersuite support.
2982 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2984 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2985 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2986 to output in BER and PEM format.
2989 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2990 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2991 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2992 ENGINE support for HMAC keys which are unextractable. New -mac and
2993 -macopt options to dgst utility.
2996 *) New option -sigopt to dgst utility. Update dgst to use
2997 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2998 alternative signing parameters such as X9.31 or PSS in the dgst
3002 *) Change ssl_cipher_apply_rule(), the internal function that does
3003 the work each time a ciphersuite string requests enabling
3004 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3005 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3006 the order of disabled ciphersuites such that those ciphersuites
3007 that most recently went from enabled to disabled not only stay
3008 in order with respect to each other, but also have higher priority
3009 than other disabled ciphersuites the next time ciphersuites are
3012 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3013 the same ciphersuites as with "HIGH" alone, but in a specific
3014 order where the PSK ciphersuites come first (since they are the
3015 most recently disabled ciphersuites when "HIGH" is parsed).
3017 Also, change ssl_create_cipher_list() (using this new
3018 funcionality) such that between otherwise identical
3019 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3023 *) Change ssl_create_cipher_list() so that it automatically
3024 arranges the ciphersuites in reasonable order before starting
3025 to process the rule string. Thus, the definition for "DEFAULT"
3026 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3027 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3028 This makes it much easier to arrive at a reasonable default order
3029 in applications for which anonymous ciphers are OK (meaning
3030 that you can't actually use DEFAULT).
3031 [Bodo Moeller; suggested by Victor Duchovni]
3033 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3034 processing) into multiple integers instead of setting
3035 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3036 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3037 (These masks as well as the individual bit definitions are hidden
3038 away into the non-exported interface ssl/ssl_locl.h, so this
3039 change to the definition of the SSL_CIPHER structure shouldn't
3040 affect applications.) This give us more bits for each of these
3041 categories, so there is no longer a need to coagulate AES128 and
3042 AES256 into a single algorithm bit, and to coagulate Camellia128
3043 and Camellia256 into a single algorithm bit, which has led to all
3046 Thus, among other things, the kludge introduced in 0.9.7m and
3047 0.9.8e for masking out AES256 independently of AES128 or masking
3048 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3050 With the change, we also introduce new ciphersuite aliases that
3051 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3055 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3056 Use the leftmost N bytes of the signature input if the input is
3057 larger than the prime q (with N being the size in bytes of q).
3060 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3061 it yet and it is largely untested.
3064 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3067 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3068 some compilers (gcc 4.2 and later) reject their use. Safestack is
3069 reimplemented. Update ASN1 to avoid use of legacy functions.
3072 *) Win32/64 targets are linked with Winsock2.
3075 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3076 to external functions. This can be used to increase CRL handling
3077 efficiency especially when CRLs are very large by (for example) storing
3078 the CRL revoked certificates in a database.
3081 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3082 new CRLs added to a directory can be used. New command line option
3083 -verify_return_error to s_client and s_server. This causes real errors
3084 to be returned by the verify callback instead of carrying on no matter
3085 what. This reflects the way a "real world" verify callback would behave.
3088 *) GOST engine, supporting several GOST algorithms and public key formats.
3089 Kindly donated by Cryptocom.
3092 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3093 partitioned by DP are handled but no indirect CRL or reason partitioning
3094 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3095 selected via a scoring technique which handles IDP and AKID in CRLs.
3098 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3099 will ultimately be used for all verify operations: this will remove the
3100 X509_STORE dependency on certificate verification and allow alternative
3101 lookup methods. X509_STORE based implementations of these two callbacks.
3104 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3105 Modify get_crl() to find a valid (unexpired) CRL if possible.
3108 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3109 this would be called X509_CRL_cmp() but that name is already used by
3110 a function that just compares CRL issuer names. Cache several CRL
3111 extensions in X509_CRL structure and cache CRLDP in X509.
3114 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3115 this maps equivalent X509_NAME structures into a consistent structure.
3116 Name comparison can then be performed rapidly using memcmp().
3119 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3123 *) Allow digests to supply their own micalg string for S/MIME type using
3124 the ctrl EVP_MD_CTRL_MICALG.
3127 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3128 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3129 ctrl. It can then customise the structure before and/or after signing
3133 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3134 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3135 to free up any added signature OIDs.
3138 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3139 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3140 digest and cipher tables. New options added to openssl utility:
3141 list-message-digest-algorithms and list-cipher-algorithms.
3144 *) Change the array representation of binary polynomials: the list
3145 of degrees of non-zero coefficients is now terminated with -1.
3146 Previously it was terminated with 0, which was also part of the
3147 value; thus, the array representation was not applicable to
3148 polynomials where t^0 has coefficient zero. This change makes
3149 the array representation useful in a more general context.
3152 *) Various modifications and fixes to SSL/TLS cipher string
3153 handling. For ECC, the code now distinguishes between fixed ECDH
3154 with RSA certificates on the one hand and with ECDSA certificates
3155 on the other hand, since these are separate ciphersuites. The
3156 unused code for Fortezza ciphersuites has been removed.
3158 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3159 (not "ECDHE"). For consistency with the code for DH
3160 certificates, use of ECDH certificates is now considered ECDH
3161 authentication, not RSA or ECDSA authentication (the latter is
3162 merely the CA's signing algorithm and not actively used in the
3165 The temporary ciphersuite alias "ECCdraft" is no longer
3166 available, and ECC ciphersuites are no longer excluded from "ALL"
3167 and "DEFAULT". The following aliases now exist for RFC 4492
3168 ciphersuites, most of these by analogy with the DH case:
3170 kECDHr - ECDH cert, signed with RSA
3171 kECDHe - ECDH cert, signed with ECDSA
3172 kECDH - ECDH cert (signed with either RSA or ECDSA)
3173 kEECDH - ephemeral ECDH
3174 ECDH - ECDH cert or ephemeral ECDH
3180 AECDH - anonymous ECDH
3181 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3185 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3186 Use correct micalg parameters depending on digest(s) in signed message.
3189 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3190 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3193 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3194 an engine to register a method. Add ENGINE lookups for methods and
3195 functional reference processing.
3198 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3199 EVP_{Sign,Verify}* which allow an application to customise the signature
3203 *) New -resign option to smime utility. This adds one or more signers
3204 to an existing PKCS#7 signedData structure. Also -md option to use an
3205 alternative message digest algorithm for signing.
3208 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3209 create PKCS7 structures containing multiple signers. Update smime
3210 application to support multiple signers.
3213 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3217 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3218 Reorganize PBE internals to lookup from a static table using NIDs,
3219 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3220 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3221 PRF which will be automatically used with PBES2.
3224 *) Replace the algorithm specific calls to generate keys in "req" with the
3228 *) Update PKCS#7 enveloped data routines to use new API. This is now
3229 supported by any public key method supporting the encrypt operation. A
3230 ctrl is added to allow the public key algorithm to examine or modify
3231 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3235 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3236 a default digest type to use. In most cases this will be SHA1 but some
3237 algorithms (such as GOST) need to specify an alternative digest. The
3238 return value indicates how strong the preference is 1 means optional and
3239 2 is mandatory (that is it is the only supported type). Modify
3240 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3241 use the default md. Update openssl utilities to use the default digest
3242 type for signing if it is not explicitly indicated.
3245 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3246 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3247 signing method from the key type. This effectively removes the link
3248 between digests and public key types.
3251 *) Add an OID cross reference table and utility functions. Its purpose is to
3252 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3253 rsaEncryption. This will allow some of the algorithm specific hackery
3254 needed to use the correct OID to be removed.
3257 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3258 structures for PKCS7_sign(). They are now set up by the relevant public
3262 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3265 *) Add support for key derivation (agreement) in the API, DH method and
3269 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3270 public and private key formats. As a side effect these add additional
3271 command line functionality not previously available: DSA signatures can be
3272 generated and verified using pkeyutl and DH key support and generation in
3277 [Oliver Tappe <zooey@hirschkaefer.de>]
3279 *) New make target "install_html_docs" installs HTML renditions of the
3281 [Oliver Tappe <zooey@hirschkaefer.de>]
3283 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3284 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3285 support key and parameter generation and add initial key generation
3286 functionality for RSA.
3289 *) Add functions for main EVP_PKEY_method operations. The undocumented
3290 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3291 EVP_PKEY_{encrypt,decrypt}_old.
3294 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3295 key API, doesn't do much yet.
3298 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3299 public key algorithms. New option to openssl utility:
3300 "list-public-key-algorithms" to print out info.
3303 *) Implement the Supported Elliptic Curves Extension for
3304 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3307 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3308 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3311 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3312 utilities such as rsa, dsa, dsaparam etc except they process any key
3316 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3317 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3318 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3322 *) Initial support for pluggable public key ASN1.
3323 De-spaghettify the public key ASN1 handling. Move public and private
3324 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3325 algorithm specific handling to a single module within the relevant
3326 algorithm directory. Add functions to allow (near) opaque processing
3327 of public and private key structures.
3330 *) Implement the Supported Point Formats Extension for
3331 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3334 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3335 for the psk identity [hint] and the psk callback functions to the
3336 SSL_SESSION, SSL and SSL_CTX structure.
3339 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3343 SSL_CTX_use_psk_identity_hint
3344 SSL_get_psk_identity_hint
3345 SSL_get_psk_identity
3346 SSL_use_psk_identity_hint
3348 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3350 *) Add RFC 3161 compliant time stamp request creation, response generation
3351 and response verification functionality.
3352 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3354 *) Add initial support for TLS extensions, specifically for the server_name
3355 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3356 have new members for a host name. The SSL data structure has an
3357 additional member SSL_CTX *initial_ctx so that new sessions can be
3358 stored in that context to allow for session resumption, even after the
3359 SSL has been switched to a new SSL_CTX in reaction to a client's
3360 server_name extension.
3362 New functions (subject to change):
3364 SSL_get_servername()
3365 SSL_get_servername_type()
3368 New CTRL codes and macros (subject to change):
3370 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3371 - SSL_CTX_set_tlsext_servername_callback()
3372 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3373 - SSL_CTX_set_tlsext_servername_arg()
3374 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3376 openssl s_client has a new '-servername ...' option.
3378 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3379 '-key2 ...', '-servername_fatal' (subject to change). This allows
3380 testing the HostName extension for a specific single host name ('-cert'
3381 and '-key' remain fallbacks for handshakes without HostName
3382 negotiation). If the unrecognized_name alert has to be sent, this by
3383 default is a warning; it becomes fatal with the '-servername_fatal'
3386 [Peter Sylvester, Remy Allais, Christophe Renou]
3388 *) Whirlpool hash implementation is added.
3391 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3392 bn(64,32). Because of instruction set limitations it doesn't have
3393 any negative impact on performance. This was done mostly in order
3394 to make it possible to share assembler modules, such as bn_mul_mont
3395 implementations, between 32- and 64-bit builds without hassle.
3398 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3399 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3403 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3404 dedicated Montgomery multiplication procedure, is introduced.
3405 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3406 "64-bit" performance on certain 32-bit targets.
3409 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3410 in SSL structures. New SSL ctrl to set maximum send fragment size.
3411 Save memory by seeting the I/O buffer sizes dynamically instead of
3412 using the maximum available value.
3415 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3416 in addition to the text details.
3419 *) Very, very preliminary EXPERIMENTAL support for printing of general
3420 ASN1 structures. This currently produces rather ugly output and doesn't
3421 handle several customised structures at all.
3424 *) Integrated support for PVK file format and some related formats such
3425 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3426 these in the 'rsa' and 'dsa' utilities.
3429 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3432 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3433 place for the (very old) "NETSCAPE" format certificates which are now
3434 handled using new ASN1 code equivalents.
3437 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3438 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3439 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3442 *) Modify CRL distribution points extension code to print out previously
3443 unsupported fields. Enhance extension setting code to allow setting of
3447 *) Add print and set support for Issuing Distribution Point CRL extension.
3450 *) Change 'Configure' script to enable Camellia by default.
3453 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3455 *) When rejecting SSL/TLS records due to an incorrect version number, never
3456 update s->server with a new major version number. As of
3457 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3458 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3459 the previous behavior could result in a read attempt at NULL when
3460 receiving specific incorrect SSL/TLS records once record payload
3461 protection is active. (CVE-2010-0740)
3462 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3464 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3465 could be crashed if the relevant tables were not present (e.g. chrooted).
3466 [Tomas Hoger <thoger@redhat.com>]
3468 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3470 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3471 [Martin Olsson, Neel Mehta]
3473 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3474 accommodate for stack sorting, always a write lock!).
3477 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3478 excessive delays in the RAND_poll(): over a minute. As a workaround
3479 include a time check in the inner Heap32Next loop too.
3482 *) The code that handled flushing of data in SSL/TLS originally used the
3483 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3484 the problem outlined in PR#1949. The fix suggested there however can
3485 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3486 of Apache). So instead simplify the code to flush unconditionally.
3487 This should be fine since flushing with no data to flush is a no op.
3490 *) Handle TLS versions 2.0 and later properly and correctly use the
3491 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3492 off ancient servers have a habit of sticking around for a while...
3495 *) Modify compression code so it frees up structures without using the
3496 ex_data callbacks. This works around a problem where some applications
3497 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3498 restarting) then use compression (e.g. SSL with compression) later.
3499 This results in significant per-connection memory leaks and
3500 has caused some security issues including CVE-2008-1678 and
3504 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3505 change when encrypting or decrypting.
3508 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3509 connect and renegotiate with servers which do not support RI.
3510 Until RI is more widely deployed this option is enabled by default.
3513 *) Add "missing" ssl ctrls to clear options and mode.
3516 *) If client attempts to renegotiate and doesn't support RI respond with
3517 a no_renegotiation alert as required by RFC5746. Some renegotiating
3518 TLS clients will continue a connection gracefully when they receive
3519 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3520 waiting for a server hello which it will never receive. Now we treat a
3521 received no_renegotiation alert as a fatal error. This is because
3522 applications requesting a renegotiation might well expect it to succeed
3523 and would have no code in place to handle the server denying it so the
3524 only safe thing to do is to terminate the connection.
3527 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3528 peer supports secure renegotiation and 0 otherwise. Print out peer
3529 renegotiation support in s_client/s_server.
3532 *) Replace the highly broken and deprecated SPKAC certification method with
3533 the updated NID creation version. This should correctly handle UTF8.
3536 *) Implement RFC5746. Re-enable renegotiation but require the extension
3537 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3538 turns out to be a bad idea. It has been replaced by
3539 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3540 SSL_CTX_set_options(). This is really not recommended unless you
3541 know what you are doing.
3542 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3544 *) Fixes to stateless session resumption handling. Use initial_ctx when
3545 issuing and attempting to decrypt tickets in case it has changed during
3546 servername handling. Use a non-zero length session ID when attempting
3547 stateless session resumption: this makes it possible to determine if
3548 a resumption has occurred immediately after receiving server hello
3549 (several places in OpenSSL subtly assume this) instead of later in
3553 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3554 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3555 fixes for a few places where the return code is not checked
3557 [Julia Lawall <julia@diku.dk>]
3559 *) Add --strict-warnings option to Configure script to include devteam
3560 warnings in other configurations.
3563 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3564 makes it possible to install openssl libraries in locations which
3565 have names other than "lib", for example "/usr/lib64" which some
3567 [Steve Henson, based on patch from Jeremy Utley]
3569 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3570 X690 8.9.12 and can produce some misleading textual output of OIDs.
3571 [Steve Henson, reported by Dan Kaminsky]
3573 *) Delete MD2 from algorithm tables. This follows the recommendation in
3574 several standards that it is not used in new applications due to
3575 several cryptographic weaknesses. For binary compatibility reasons
3576 the MD2 API is still compiled in by default.
3579 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3583 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3584 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3586 [Guenter <lists@gknw.net>]
3588 *) Fix the server certificate chain building code to use X509_verify_cert(),
3589 it used to have an ad-hoc builder which was unable to cope with anything
3590 other than a simple chain.
3591 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3593 *) Don't check self signed certificate signatures in X509_verify_cert()
3594 by default (a flag can override this): it just wastes time without
3595 adding any security. As a useful side effect self signed root CAs
3596 with non-FIPS digests are now usable in FIPS mode.
3599 *) In dtls1_process_out_of_seq_message() the check if the current message
3600 is already buffered was missing. For every new message was memory
3601 allocated, allowing an attacker to perform an denial of service attack
3602 with sending out of seq handshake messages until there is no memory
3603 left. Additionally every future messege was buffered, even if the
3604 sequence number made no sense and would be part of another handshake.
3605 So only messages with sequence numbers less than 10 in advance will be
3606 buffered. (CVE-2009-1378)
3607 [Robin Seggelmann, discovered by Daniel Mentz]
3609 *) Records are buffered if they arrive with a future epoch to be
3610 processed after finishing the corresponding handshake. There is
3611 currently no limitation to this buffer allowing an attacker to perform
3612 a DOS attack with sending records with future epochs until there is no
3613 memory left. This patch adds the pqueue_size() function to determine
3614 the size of a buffer and limits the record buffer to 100 entries.
3616 [Robin Seggelmann, discovered by Daniel Mentz]
3618 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3619 parent structure is freed. (CVE-2009-1379)
3622 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3623 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3626 [Ilya O. <vrghost@gmail.com>]
3628 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3630 *) Disable renegotiation completely - this fixes a severe security
3631 problem (CVE-2009-3555) at the cost of breaking all
3632 renegotiation. Renegotiation can be re-enabled by setting
3633 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3634 run-time. This is really not recommended unless you know what
3638 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3640 *) Don't set val to NULL when freeing up structures, it is freed up by
3641 underlying code. If sizeof(void *) > sizeof(long) this can result in
3642 zeroing past the valid field. (CVE-2009-0789)
3643 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3645 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3646 checked correctly. This would allow some invalid signed attributes to
3647 appear to verify correctly. (CVE-2009-0591)
3648 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3650 *) Reject UniversalString and BMPString types with invalid lengths. This
3651 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3652 a legal length. (CVE-2009-0590)
3655 *) Set S/MIME signing as the default purpose rather than setting it
3656 unconditionally. This allows applications to override it at the store
3660 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3661 to handle some structures.
3664 *) Improve efficiency of mem_gets: don't search whole buffer each time
3666 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3668 *) New -hex option for openssl rand.
3671 *) Print out UTF8String and NumericString when parsing ASN1.
3674 *) Support NumericString type for name components.
3677 *) Allow CC in the environment to override the automatically chosen
3678 compiler. Note that nothing is done to ensure flags work with the
3682 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3684 *) Properly check EVP_VerifyFinal() and similar return values
3686 [Ben Laurie, Bodo Moeller, Google Security Team]
3688 *) Enable TLS extensions by default.
3691 *) Allow the CHIL engine to be loaded, whether the application is
3692 multithreaded or not. (This does not release the developer from the
3693 obligation to set up the dynamic locking callbacks.)
3694 [Sander Temme <sander@temme.net>]
3696 *) Use correct exit code if there is an error in dgst command.
3697 [Steve Henson; problem pointed out by Roland Dirlewanger]
3699 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3700 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3703 *) Add experimental JPAKE support, including demo authentication in
3704 s_client and s_server.
3707 *) Set the comparison function in v3_addr_canonize().
3708 [Rob Austein <sra@hactrn.net>]
3710 *) Add support for XMPP STARTTLS in s_client.
3711 [Philip Paeps <philip@freebsd.org>]
3713 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3714 to ensure that even with this option, only ciphersuites in the
3715 server's preference list will be accepted. (Note that the option
3716 applies only when resuming a session, so the earlier behavior was
3717 just about the algorithm choice for symmetric cryptography.)
3720 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3722 *) Fix NULL pointer dereference if a DTLS server received
3723 ChangeCipherSpec as first record (CVE-2009-1386).
3726 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3727 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3730 *) The fix in 0.9.8c that supposedly got rid of unsafe
3731 double-checked locking was incomplete for RSA blinding,
3732 addressing just one layer of what turns out to have been
3733 doubly unsafe triple-checked locking.
3735 So now fix this for real by retiring the MONT_HELPER macro
3736 in crypto/rsa/rsa_eay.c.
3738 [Bodo Moeller; problem pointed out by Marius Schilder]
3740 *) Various precautionary measures:
3742 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3744 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3745 (NB: This would require knowledge of the secret session ticket key
3746 to exploit, in which case you'd be SOL either way.)
3748 - Change bn_nist.c so that it will properly handle input BIGNUMs
3749 outside the expected range.
3751 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3754 [Neel Mehta, Bodo Moeller]
3756 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3757 the load fails. Useful for distros.
3758 [Ben Laurie and the FreeBSD team]
3760 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3763 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3766 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3768 This work was sponsored by Logica.
3771 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3772 keystores. Support for SSL/TLS client authentication too.
3773 Not compiled unless enable-capieng specified to Configure.
3775 This work was sponsored by Logica.
3778 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3779 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3780 attribute creation routines such as certifcate requests and PKCS#12
3784 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3786 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3787 handshake which could lead to a cilent crash as found using the
3788 Codenomicon TLS test suite (CVE-2008-1672)
3789 [Steve Henson, Mark Cox]
3791 *) Fix double free in TLS server name extensions which could lead to
3792 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3795 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3797 Clear the error queue to ensure that error entries left from
3798 older function calls do not interfere with the correct operation.
3799 [Lutz Jaenicke, Erik de Castro Lopo]
3801 *) Remove root CA certificates of commercial CAs:
3803 The OpenSSL project does not recommend any specific CA and does not
3804 have any policy with respect to including or excluding any CA.
3805 Therefore it does not make any sense to ship an arbitrary selection
3806 of root CA certificates with the OpenSSL software.
3809 *) RSA OAEP patches to fix two separate invalid memory reads.
3810 The first one involves inputs when 'lzero' is greater than
3811 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3812 before the beginning of from). The second one involves inputs where
3813 the 'db' section contains nothing but zeroes (there is a one-byte
3814 invalid read after the end of 'db').
3815 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3817 *) Partial backport from 0.9.9-dev:
3819 Introduce bn_mul_mont (dedicated Montgomery multiplication
3820 procedure) as a candidate for BIGNUM assembler implementation.
3821 While 0.9.9-dev uses assembler for various architectures, only
3822 x86_64 is available by default here in the 0.9.8 branch, and
3823 32-bit x86 is available through a compile-time setting.
3825 To try the 32-bit x86 assembler implementation, use Configure
3826 option "enable-montasm" (which exists only for this backport).
3828 As "enable-montasm" for 32-bit x86 disclaims code stability
3829 anyway, in this constellation we activate additional code
3830 backported from 0.9.9-dev for further performance improvements,
3831 namely BN_from_montgomery_word. (To enable this otherwise,
3832 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3834 [Andy Polyakov (backport partially by Bodo Moeller)]
3836 *) Add TLS session ticket callback. This allows an application to set
3837 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3838 values. This is useful for key rollover for example where several key
3839 sets may exist with different names.
3842 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3843 This was broken until now in 0.9.8 releases, such that the only way
3844 a registered ENGINE could be used (assuming it initialises
3845 successfully on the host) was to explicitly set it as the default
3846 for the relevant algorithms. This is in contradiction with 0.9.7
3847 behaviour and the documentation. With this fix, when an ENGINE is
3848 registered into a given algorithm's table of implementations, the
3849 'uptodate' flag is reset so that auto-discovery will be used next
3850 time a new context for that algorithm attempts to select an
3852 [Ian Lister (tweaked by Geoff Thorpe)]
3854 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3855 implemention in the following ways:
3857 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3860 Lack of BER streaming support means one pass streaming processing is
3861 only supported if data is detached: setting the streaming flag is
3862 ignored for embedded content.
3864 CMS support is disabled by default and must be explicitly enabled
3865 with the enable-cms configuration option.
3868 *) Update the GMP engine glue to do direct copies between BIGNUM and
3869 mpz_t when openssl and GMP use the same limb size. Otherwise the
3870 existing "conversion via a text string export" trick is still used.
3871 [Paul Sheer <paulsheer@gmail.com>]
3873 *) Zlib compression BIO. This is a filter BIO which compressed and
3874 uncompresses any data passed through it.
3877 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3878 RFC3394 compatible AES key wrapping.
3881 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3882 sets string data without copying. X509_ALGOR_set0() and
3883 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3884 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3885 from an X509_ATTRIBUTE structure optionally checking it occurs only
3886 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3890 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3891 to get the expected BN_FLG_CONSTTIME behavior.
3892 [Bodo Moeller (Google)]
3896 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3897 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3898 - added some more tests to do_tests.pl
3899 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3900 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3901 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3902 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3903 - various changes to netware.pl to enable gcc-cross builds on Win32
3905 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3906 - various changes to fix missing prototype warnings
3907 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3908 - added AES, WHIRLPOOL and CPUID assembler code to build files
3909 - added missing AES assembler make rules to mk1mf.pl
3910 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3911 [Guenter Knauf <eflash@gmx.net>]
3913 *) Implement certificate status request TLS extension defined in RFC3546.
3914 A client can set the appropriate parameters and receive the encoded
3915 OCSP response via a callback. A server can query the supplied parameters
3916 and set the encoded OCSP response in the callback. Add simplified examples
3917 to s_client and s_server.
3920 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3922 *) Fix various bugs:
3923 + Binary incompatibility of ssl_ctx_st structure
3924 + DTLS interoperation with non-compliant servers
3925 + Don't call get_session_cb() without proposed session
3926 + Fix ia64 assembler code
3927 [Andy Polyakov, Steve Henson]
3929 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3931 *) DTLS Handshake overhaul. There were longstanding issues with
3932 OpenSSL DTLS implementation, which were making it impossible for
3933 RFC 4347 compliant client to communicate with OpenSSL server.
3934 Unfortunately just fixing these incompatibilities would "cut off"
3935 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3936 server keeps tolerating non RFC compliant syntax. The opposite is
3937 not true, 0.9.8f client can not communicate with earlier server.
3938 This update even addresses CVE-2007-4995.
3941 *) Changes to avoid need for function casts in OpenSSL: some compilers
3942 (gcc 4.2 and later) reject their use.
3943 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3946 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3947 RFC4507bis. The encrypted ticket format is an encrypted encoded
3948 SSL_SESSION structure, that way new session features are automatically
3951 If a client application caches session in an SSL_SESSION structure
3952 support is transparent because tickets are now stored in the encoded
3955 The SSL_CTX structure automatically generates keys for ticket
3956 protection in servers so again support should be possible
3957 with no application modification.
3959 If a client or server wishes to disable RFC4507 support then the option
3960 SSL_OP_NO_TICKET can be set.
3962 Add a TLS extension debugging callback to allow the contents of any client
3963 or server extensions to be examined.
3965 This work was sponsored by Google.
3968 *) Add initial support for TLS extensions, specifically for the server_name
3969 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3970 have new members for a host name. The SSL data structure has an
3971 additional member SSL_CTX *initial_ctx so that new sessions can be
3972 stored in that context to allow for session resumption, even after the
3973 SSL has been switched to a new SSL_CTX in reaction to a client's
3974 server_name extension.
3976 New functions (subject to change):
3978 SSL_get_servername()
3979 SSL_get_servername_type()
3982 New CTRL codes and macros (subject to change):
3984 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3985 - SSL_CTX_set_tlsext_servername_callback()
3986 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3987 - SSL_CTX_set_tlsext_servername_arg()
3988 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3990 openssl s_client has a new '-servername ...' option.
3992 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3993 '-key2 ...', '-servername_fatal' (subject to change). This allows
3994 testing the HostName extension for a specific single host name ('-cert'
3995 and '-key' remain fallbacks for handshakes without HostName
3996 negotiation). If the unrecognized_name alert has to be sent, this by
3997 default is a warning; it becomes fatal with the '-servername_fatal'
4000 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4002 *) Add AES and SSE2 assembly language support to VC++ build.
4005 *) Mitigate attack on final subtraction in Montgomery reduction.
4008 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4009 (which previously caused an internal error).
4012 *) Squeeze another 10% out of IGE mode when in != out.
4015 *) AES IGE mode speedup.
4016 [Dean Gaudet (Google)]
4018 *) Add the Korean symmetric 128-bit cipher SEED (see
4019 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4020 add SEED ciphersuites from RFC 4162:
4022 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4023 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4024 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4025 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4027 To minimize changes between patchlevels in the OpenSSL 0.9.8
4028 series, SEED remains excluded from compilation unless OpenSSL
4029 is configured with 'enable-seed'.
4030 [KISA, Bodo Moeller]
4032 *) Mitigate branch prediction attacks, which can be practical if a
4033 single processor is shared, allowing a spy process to extract
4034 information. For detailed background information, see
4035 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4036 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4037 and Necessary Software Countermeasures"). The core of the change
4038 are new versions BN_div_no_branch() and
4039 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4040 respectively, which are slower, but avoid the security-relevant
4041 conditional branches. These are automatically called by BN_div()
4042 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4043 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4044 remove a conditional branch.
4046 BN_FLG_CONSTTIME is the new name for the previous
4047 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4048 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4049 in the exponent causes BN_mod_exp_mont() to use the alternative
4050 implementation in BN_mod_exp_mont_consttime().) The old name
4051 remains as a deprecated alias.
4053 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4054 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4055 constant-time implementations for more than just exponentiation.
4056 Here too the old name is kept as a deprecated alias.
4058 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4059 the BN_BLINDING structure gets an independent copy of the
4060 modulus. This means that the previous "BIGNUM *m" argument to
4061 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4062 essentially becomes "const BIGNUM *m", although we can't actually
4063 change this in the header file before 0.9.9. It allows
4064 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4065 enable BN_FLG_CONSTTIME.
4067 [Matthew D Wood (Intel Corp)]
4069 *) In the SSL/TLS server implementation, be strict about session ID
4070 context matching (which matters if an application uses a single
4071 external cache for different purposes). Previously,
4072 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4073 set. This did ensure strict client verification, but meant that,
4074 with applications using a single external cache for quite
4075 different requirements, clients could circumvent ciphersuite
4076 restrictions for a given session ID context by starting a session
4077 in a different context.
4080 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4081 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4082 authentication-only ciphersuites.
4085 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4086 not complete and could lead to a possible single byte overflow
4087 (CVE-2007-5135) [Ben Laurie]
4089 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4091 *) Since AES128 and AES256 (and similarly Camellia128 and
4092 Camellia256) share a single mask bit in the logic of
4093 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4094 kludge to work properly if AES128 is available and AES256 isn't
4095 (or if Camellia128 is available and Camellia256 isn't).
4098 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4099 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4100 When a point or a seed is encoded in a BIT STRING, we need to
4101 prevent the removal of trailing zero bits to get the proper DER
4102 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4103 of a NamedBitList, for which trailing 0 bits need to be removed.)
4106 *) Have SSL/TLS server implementation tolerate "mismatched" record
4107 protocol version while receiving ClientHello even if the
4108 ClientHello is fragmented. (The server can't insist on the
4109 particular protocol version it has chosen before the ServerHello
4110 message has informed the client about his choice.)
4113 *) Add RFC 3779 support.
4114 [Rob Austein for ARIN, Ben Laurie]
4116 *) Load error codes if they are not already present instead of using a
4117 static variable. This allows them to be cleanly unloaded and reloaded.
4118 Improve header file function name parsing.
4121 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4122 or CAPABILITY handshake as required by RFCs.
4125 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4127 *) Introduce limits to prevent malicious keys being able to
4128 cause a denial of service. (CVE-2006-2940)
4129 [Steve Henson, Bodo Moeller]
4131 *) Fix ASN.1 parsing of certain invalid structures that can result
4132 in a denial of service. (CVE-2006-2937) [Steve Henson]
4134 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4135 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4137 *) Fix SSL client code which could crash if connecting to a
4138 malicious SSLv2 server. (CVE-2006-4343)
4139 [Tavis Ormandy and Will Drewry, Google Security Team]
4141 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4142 match only those. Before that, "AES256-SHA" would be interpreted
4143 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4144 the same strength classification in 0.9.7h) as we currently only
4145 have a single AES bit in the ciphersuite description bitmap.
4146 That change, however, also applied to ciphersuite strings such as
4147 "RC4-MD5" that intentionally matched multiple ciphersuites --
4148 namely, SSL 2.0 ciphersuites in addition to the more common ones
4149 from SSL 3.0/TLS 1.0.
4151 So we change the selection algorithm again: Naming an explicit
4152 ciphersuite selects this one ciphersuite, and any other similar
4153 ciphersuite (same bitmap) from *other* protocol versions.
4154 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4155 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4157 Since SSL 2.0 does not have any ciphersuites for which the
4158 128/256 bit distinction would be relevant, this works for now.
4159 The proper fix will be to use different bits for AES128 and
4160 AES256, which would have avoided the problems from the beginning;
4161 however, bits are scarce, so we can only do this in a new release
4162 (not just a patchlevel) when we can change the SSL_CIPHER
4163 definition to split the single 'unsigned long mask' bitmap into
4164 multiple values to extend the available space.
4168 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4170 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4171 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4173 *) Add AES IGE and biIGE modes.
4176 *) Change the Unix randomness entropy gathering to use poll() when
4177 possible instead of select(), since the latter has some
4178 undesirable limitations.
4179 [Darryl Miles via Richard Levitte and Bodo Moeller]
4181 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4182 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4183 cannot be implicitly activated as part of, e.g., the "AES" alias.
4184 However, please upgrade to OpenSSL 0.9.9[-dev] for
4185 non-experimental use of the ECC ciphersuites to get TLS extension
4186 support, which is required for curve and point format negotiation
4187 to avoid potential handshake problems.
4190 *) Disable rogue ciphersuites:
4192 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4193 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4194 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4196 The latter two were purportedly from
4197 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4200 Also deactivate the remaining ciphersuites from
4201 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4202 unofficial, and the ID has long expired.
4205 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4206 dual-core machines) and other potential thread-safety issues.
4209 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4210 versions), which is now available for royalty-free use
4211 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4212 Also, add Camellia TLS ciphersuites from RFC 4132.
4214 To minimize changes between patchlevels in the OpenSSL 0.9.8
4215 series, Camellia remains excluded from compilation unless OpenSSL
4216 is configured with 'enable-camellia'.
4219 *) Disable the padding bug check when compression is in use. The padding
4220 bug check assumes the first packet is of even length, this is not
4221 necessarily true if compresssion is enabled and can result in false
4222 positives causing handshake failure. The actual bug test is ancient
4223 code so it is hoped that implementations will either have fixed it by
4224 now or any which still have the bug do not support compression.
4227 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4229 *) When applying a cipher rule check to see if string match is an explicit
4230 cipher suite and only match that one cipher suite if it is.
4233 *) Link in manifests for VC++ if needed.
4234 [Austin Ziegler <halostatue@gmail.com>]
4236 *) Update support for ECC-based TLS ciphersuites according to
4237 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4238 TLS extensions, which are supported starting with the 0.9.9
4239 branch, not in the OpenSSL 0.9.8 branch).
4242 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4243 opaque EVP_CIPHER_CTX handling.
4246 *) Fixes and enhancements to zlib compression code. We now only use
4247 "zlib1.dll" and use the default __cdecl calling convention on Win32
4248 to conform with the standards mentioned here:
4249 http://www.zlib.net/DLL_FAQ.txt
4250 Static zlib linking now works on Windows and the new --with-zlib-include
4251 --with-zlib-lib options to Configure can be used to supply the location
4252 of the headers and library. Gracefully handle case where zlib library
4256 *) Several fixes and enhancements to the OID generation code. The old code
4257 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4258 handle numbers larger than ULONG_MAX, truncated printing and had a
4259 non standard OBJ_obj2txt() behaviour.
4262 *) Add support for building of engines under engine/ as shared libraries
4263 under VC++ build system.
4266 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4267 Hopefully, we will not see any false combination of paths any more.
4270 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4272 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4273 (part of SSL_OP_ALL). This option used to disable the
4274 countermeasure against man-in-the-middle protocol-version
4275 rollback in the SSL 2.0 server implementation, which is a bad
4276 idea. (CVE-2005-2969)
4278 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4279 for Information Security, National Institute of Advanced Industrial
4280 Science and Technology [AIST], Japan)]
4282 *) Add two function to clear and return the verify parameter flags.
4285 *) Keep cipherlists sorted in the source instead of sorting them at
4286 runtime, thus removing the need for a lock.
4289 *) Avoid some small subgroup attacks in Diffie-Hellman.
4290 [Nick Mathewson and Ben Laurie]
4292 *) Add functions for well-known primes.
4295 *) Extended Windows CE support.
4296 [Satoshi Nakamura and Andy Polyakov]
4298 *) Initialize SSL_METHOD structures at compile time instead of during
4299 runtime, thus removing the need for a lock.
4302 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4303 attempting to decrypt each encrypted key in turn. Add support to
4307 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4309 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4312 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4315 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4316 key into the same file any more.
4319 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4322 *) Add -utf8 command line and config file option to 'ca'.
4323 [Stefan <stf@udoma.org]
4325 *) Removed the macro des_crypt(), as it seems to conflict with some
4326 libraries. Use DES_crypt().
4329 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4330 involves renaming the source and generated shared-libs for
4331 both. The engines will accept the corrected or legacy ids
4332 ('ncipher' and '4758_cca' respectively) when binding. NB,
4333 this only applies when building 'shared'.
4334 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4336 *) Add attribute functions to EVP_PKEY structure. Modify
4337 PKCS12_create() to recognize a CSP name attribute and
4338 use it. Make -CSP option work again in pkcs12 utility.
4341 *) Add new functionality to the bn blinding code:
4342 - automatic re-creation of the BN_BLINDING parameters after
4343 a fixed number of uses (currently 32)
4344 - add new function for parameter creation
4345 - introduce flags to control the update behaviour of the
4346 BN_BLINDING parameters
4347 - hide BN_BLINDING structure
4348 Add a second BN_BLINDING slot to the RSA structure to improve
4349 performance when a single RSA object is shared among several
4353 *) Add support for DTLS.
4354 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4356 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4357 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4360 *) Remove buggy and incomplete DH cert support from
4361 ssl/ssl_rsa.c and ssl/s3_both.c
4364 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4365 the apps/openssl applications.
4368 *) Compile clean with "-Wall -Wmissing-prototypes
4369 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4370 DEBUG_SAFESTACK must also be set.
4373 *) Change ./Configure so that certain algorithms can be disabled by default.
4374 The new counterpiece to "no-xxx" is "enable-xxx".
4376 The patented RC5 and MDC2 algorithms will now be disabled unless
4377 "enable-rc5" and "enable-mdc2", respectively, are specified.
4379 (IDEA remains enabled despite being patented. This is because IDEA
4380 is frequently required for interoperability, and there is no license
4381 fee for non-commercial use. As before, "no-idea" can be used to
4382 avoid this algorithm.)
4386 *) Add processing of proxy certificates (see RFC 3820). This work was
4387 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4388 EGEE (Enabling Grids for E-science in Europe).
4391 *) RC4 performance overhaul on modern architectures/implementations, such
4392 as Intel P4, IA-64 and AMD64.
4395 *) New utility extract-section.pl. This can be used specify an alternative
4396 section number in a pod file instead of having to treat each file as
4397 a separate case in Makefile. This can be done by adding two lines to the
4400 =for comment openssl_section:XXX
4402 The blank line is mandatory.
4406 *) New arguments -certform, -keyform and -pass for s_client and s_server
4407 to allow alternative format key and certificate files and passphrase
4411 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4412 update associated structures and add various utility functions.
4414 Add new policy related verify parameters, include policy checking in
4415 standard verify code. Enhance 'smime' application with extra parameters
4416 to support policy checking and print out.
4419 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4420 Nehemiah processors. These extensions support AES encryption in hardware
4421 as well as RNG (though RNG support is currently disabled).
4422 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4424 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4427 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4428 [Andy Polyakov and a number of other people]
4430 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4431 implementation contributed by IBM.
4432 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4434 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4435 exponent rather than 'unsigned long'. There is a corresponding change to
4436 the new 'rsa_keygen' element of the RSA_METHOD structure.
4437 [Jelte Jansen, Geoff Thorpe]
4439 *) Functionality for creating the initial serial number file is now
4440 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4442 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4443 number file to 1, which is bound to cause problems. To avoid
4444 the problems while respecting compatibility between different 0.9.7
4445 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4446 CA.pl for serial number initialization. With the new release 0.9.8,
4447 we can fix the problem directly in the 'ca' utility.)
4450 *) Reduced header interdepencies by declaring more opaque objects in
4451 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4452 give fewer recursive includes, which could break lazy source code - so
4453 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4454 developers should define this symbol when building and using openssl to
4455 ensure they track the recommended behaviour, interfaces, [etc], but
4456 backwards-compatible behaviour prevails when this isn't defined.
4459 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4462 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4463 This will generate a random key of the appropriate length based on the
4464 cipher context. The EVP_CIPHER can provide its own random key generation
4465 routine to support keys of a specific form. This is used in the des and
4466 3des routines to generate a key of the correct parity. Update S/MIME
4467 code to use new functions and hence generate correct parity DES keys.
4468 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4469 valid (weak or incorrect parity).
4472 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4473 as looking them up. This is useful when the verified structure may contain
4474 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4475 present unless the new PKCS7_NO_CRL flag is asserted.
4478 *) Extend ASN1 oid configuration module. It now additionally accepts the
4481 shortName = some long name, 1.2.3.4
4484 *) Reimplemented the BN_CTX implementation. There is now no more static
4485 limitation on the number of variables it can handle nor the depth of the
4486 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4487 information can now expand as required, and rather than having a single
4488 static array of bignums, BN_CTX now uses a linked-list of such arrays
4489 allowing it to expand on demand whilst maintaining the usefulness of
4490 BN_CTX's "bundling".
4493 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4494 to allow all RSA operations to function using a single BN_CTX.
4497 *) Preliminary support for certificate policy evaluation and checking. This
4498 is initially intended to pass the tests outlined in "Conformance Testing
4499 of Relying Party Client Certificate Path Processing Logic" v1.07.
4502 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4503 remained unused and not that useful. A variety of other little bignum
4504 tweaks and fixes have also been made continuing on from the audit (see
4508 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4509 associated ASN1, EVP and SSL functions and old ASN1 macros.
4512 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4513 and this should never fail. So the return value from the use of
4514 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4515 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4518 *) BN_CTX_get() should return zero-valued bignums, providing the same
4519 initialised value as BN_new().
4520 [Geoff Thorpe, suggested by Ulf Möller]
4522 *) Support for inhibitAnyPolicy certificate extension.
4525 *) An audit of the BIGNUM code is underway, for which debugging code is
4526 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4527 is considered valid when processing BIGNUMs, and causes execution to
4528 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4529 further steps are taken to deliberately pollute unused data in BIGNUM
4530 structures to try and expose faulty code further on. For now, openssl will
4531 (in its default mode of operation) continue to tolerate the inconsistent
4532 forms that it has tolerated in the past, but authors and packagers should
4533 consider trying openssl and their own applications when compiled with
4534 these debugging symbols defined. It will help highlight potential bugs in
4535 their own code, and will improve the test coverage for OpenSSL itself. At
4536 some point, these tighter rules will become openssl's default to improve
4537 maintainability, though the assert()s and other overheads will remain only
4538 in debugging configurations. See bn.h for more details.
4539 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4541 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4542 that can only be obtained through BN_CTX_new() (which implicitly
4543 initialises it). The presence of this function only made it possible
4544 to overwrite an existing structure (and cause memory leaks).
4547 *) Because of the callback-based approach for implementing LHASH as a
4548 template type, lh_insert() adds opaque objects to hash-tables and
4549 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4550 to clean up those corresponding objects before destroying the hash table
4551 (and losing the object pointers). So some over-zealous constifications in
4552 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4553 objects as "const" and the lh_doall[_arg] callback wrappers are not
4554 prototyped to have "const" restrictions on the object pointers they are
4555 given (and so aren't required to cast them away any more).
4558 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4559 (speed) prefers to use its own implementation. The two implementations
4560 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4561 its object type properly exposed (MS_TM) instead of casting to/from "char
4562 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4563 aren't necessarily the greatest nomenclatures - but this is what was used
4564 internally to the implementation so I've used that for now.
4567 *) Ensure that deprecated functions do not get compiled when
4568 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4569 the self-tests were still using deprecated key-generation functions so
4570 these have been updated also.
4573 *) Reorganise PKCS#7 code to separate the digest location functionality
4574 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4575 New function PKCS7_set_digest() to set the digest type for PKCS#7
4576 digestedData type. Add additional code to correctly generate the
4577 digestedData type and add support for this type in PKCS7 initialization
4581 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4582 structure of type "other".
4585 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4586 sure the loop does correctly stop and breaking ("division by zero")
4587 modulus operations are not performed. The (pre-generated) prime
4588 table crypto/bn/bn_prime.h was already correct, but it could not be
4589 re-generated on some platforms because of the "division by zero"
4590 situation in the script.
4591 [Ralf S. Engelschall]
4593 *) Update support for ECC-based TLS ciphersuites according to
4594 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4595 SHA-1 now is only used for "small" curves (where the
4596 representation of a field element takes up to 24 bytes); for
4597 larger curves, the field element resulting from ECDH is directly
4598 used as premaster secret.
4599 [Douglas Stebila (Sun Microsystems Laboratories)]
4601 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4602 curve secp160r1 to the tests.
4603 [Douglas Stebila (Sun Microsystems Laboratories)]
4605 *) Add the possibility to load symbols globally with DSO.
4606 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4608 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4609 control of the error stack.
4612 *) Add support for STORE in ENGINE.
4615 *) Add the STORE type. The intention is to provide a common interface
4616 to certificate and key stores, be they simple file-based stores, or
4617 HSM-type store, or LDAP stores, or...
4618 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4621 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4622 pass a list of arguments to any function as well as provide a way
4623 for a function to pass data back to the caller.
4626 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4627 works like BUF_strdup() but can be used to duplicate a portion of
4628 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4632 *) Add the function sk_find_ex() which works like sk_find(), but will
4633 return an index to an element even if an exact match couldn't be
4634 found. The index is guaranteed to point at the element where the
4635 searched-for key would be inserted to preserve sorting order.
4638 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4639 takes an extra flags argument for optional functionality. Currently,
4640 the following flags are defined:
4642 OBJ_BSEARCH_VALUE_ON_NOMATCH
4643 This one gets OBJ_bsearch_ex() to return a pointer to the first
4644 element where the comparing function returns a negative or zero
4647 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4648 This one gets OBJ_bsearch_ex() to return a pointer to the first
4649 element where the comparing function returns zero. This is useful
4650 if there are more than one element where the comparing function
4654 *) Make it possible to create self-signed certificates with 'openssl ca'
4655 in such a way that the self-signed certificate becomes part of the
4656 CA database and uses the same mechanisms for serial number generation
4657 as all other certificate signing. The new flag '-selfsign' enables
4658 this functionality. Adapt CA.sh and CA.pl.in.
4661 *) Add functionality to check the public key of a certificate request
4662 against a given private. This is useful to check that a certificate
4663 request can be signed by that key (self-signing).
4666 *) Make it possible to have multiple active certificates with the same
4667 subject in the CA index file. This is done only if the keyword
4668 'unique_subject' is set to 'no' in the main CA section (default
4669 if 'CA_default') of the configuration file. The value is saved
4670 with the database itself in a separate index attribute file,
4671 named like the index file with '.attr' appended to the name.
4674 *) Generate muti valued AVAs using '+' notation in config files for
4678 *) Support for nameConstraints certificate extension.
4681 *) Support for policyConstraints certificate extension.
4684 *) Support for policyMappings certificate extension.
4687 *) Make sure the default DSA_METHOD implementation only uses its
4688 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4689 and change its own handlers to be NULL so as to remove unnecessary
4690 indirection. This lets alternative implementations fallback to the
4691 default implementation more easily.
4694 *) Support for directoryName in GeneralName related extensions
4698 *) Make it possible to link applications using Makefile.shared.
4699 Make that possible even when linking against static libraries!
4702 *) Support for single pass processing for S/MIME signing. This now
4703 means that S/MIME signing can be done from a pipe, in addition
4704 cleartext signing (multipart/signed type) is effectively streaming
4705 and the signed data does not need to be all held in memory.
4707 This is done with a new flag PKCS7_STREAM. When this flag is set
4708 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4709 is done after the data is output (and digests calculated) in
4710 SMIME_write_PKCS7().
4713 *) Add full support for -rpath/-R, both in shared libraries and
4714 applications, at least on the platforms where it's known how
4718 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4719 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4720 will now compute a table of multiples of the generator that
4721 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4722 faster (notably in the case of a single point multiplication,
4723 scalar * generator).
4724 [Nils Larsch, Bodo Moeller]
4726 *) IPv6 support for certificate extensions. The various extensions
4727 which use the IP:a.b.c.d can now take IPv6 addresses using the
4728 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4732 *) Added an ENGINE that implements RSA by performing private key
4733 exponentiations with the GMP library. The conversions to and from
4734 GMP's mpz_t format aren't optimised nor are any montgomery forms
4735 cached, and on x86 it appears OpenSSL's own performance has caught up.
4736 However there are likely to be other architectures where GMP could
4737 provide a boost. This ENGINE is not built in by default, but it can be
4738 specified at Configure time and should be accompanied by the necessary
4739 linker additions, eg;
4740 ./config -DOPENSSL_USE_GMP -lgmp
4743 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4744 testing availability of engines with "-t" - the old behaviour is
4745 produced by increasing the feature's verbosity with "-tt".
4748 *) ECDSA routines: under certain error conditions uninitialized BN objects
4749 could be freed. Solution: make sure initialization is performed early
4750 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4754 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4755 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4756 software implementations. For DSA and DH, parameter generation can
4757 also be overriden by providing the appropriate method callbacks.
4760 *) Change the "progress" mechanism used in key-generation and
4761 primality testing to functions that take a new BN_GENCB pointer in
4762 place of callback/argument pairs. The new API functions have "_ex"
4763 postfixes and the older functions are reimplemented as wrappers for
4764 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4765 declarations of the old functions to help (graceful) attempts to
4766 migrate to the new functions. Also, the new key-generation API
4767 functions operate on a caller-supplied key-structure and return
4768 success/failure rather than returning a key or NULL - this is to
4769 help make "keygen" another member function of RSA_METHOD etc.
4771 Example for using the new callback interface:
4773 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4777 BN_GENCB_set(&my_cb, my_callback, my_arg);
4779 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4780 /* For the meaning of a, b in calls to my_callback(), see the
4781 * documentation of the function that calls the callback.
4782 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4783 * my_callback should return 1 if it wants BN_is_prime_ex()
4784 * to continue, or 0 to stop.
4789 *) Change the ZLIB compression method to be stateful, and make it
4790 available to TLS with the number defined in
4791 draft-ietf-tls-compression-04.txt.
4794 *) Add the ASN.1 structures and functions for CertificatePair, which
4795 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4797 CertificatePair ::= SEQUENCE {
4798 forward [0] Certificate OPTIONAL,
4799 reverse [1] Certificate OPTIONAL,
4800 -- at least one of the pair shall be present -- }
4802 Also implement the PEM functions to read and write certificate
4803 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4805 This needed to be defined, mostly for the sake of the LDAP
4806 attribute crossCertificatePair, but may prove useful elsewhere as
4810 *) Make it possible to inhibit symlinking of shared libraries in
4811 Makefile.shared, for Cygwin's sake.
4814 *) Extend the BIGNUM API by creating a function
4815 void BN_set_negative(BIGNUM *a, int neg);
4816 and a macro that behave like
4817 int BN_is_negative(const BIGNUM *a);
4819 to avoid the need to access 'a->neg' directly in applications.
4822 *) Implement fast modular reduction for pseudo-Mersenne primes
4823 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4824 EC_GROUP_new_curve_GFp() will now automatically use this
4826 [Nils Larsch <nla@trustcenter.de>]
4828 *) Add new lock type (CRYPTO_LOCK_BN).
4831 *) Change the ENGINE framework to automatically load engines
4832 dynamically from specific directories unless they could be
4833 found to already be built in or loaded. Move all the
4834 current engines except for the cryptodev one to a new
4836 The engines in engines/ are built as shared libraries if
4837 the "shared" options was given to ./Configure or ./config.
4838 Otherwise, they are inserted in libcrypto.a.
4839 /usr/local/ssl/engines is the default directory for dynamic
4840 engines, but that can be overriden at configure time through
4841 the usual use of --prefix and/or --openssldir, and at run
4842 time with the environment variable OPENSSL_ENGINES.
4843 [Geoff Thorpe and Richard Levitte]
4845 *) Add Makefile.shared, a helper makefile to build shared
4846 libraries. Adapt Makefile.org.
4849 *) Add version info to Win32 DLLs.
4850 [Peter 'Luna' Runestig" <peter@runestig.com>]
4852 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4853 can be added using this API to created arbitrary PKCS#12
4854 files while avoiding the low level API.
4856 New options to PKCS12_create(), key or cert can be NULL and
4857 will then be omitted from the output file. The encryption
4858 algorithm NIDs can be set to -1 for no encryption, the mac
4859 iteration count can be set to 0 to omit the mac.
4861 Enhance pkcs12 utility by making the -nokeys and -nocerts
4862 options work when creating a PKCS#12 file. New option -nomac
4863 to omit the mac, NONE can be set for an encryption algorithm.
4864 New code is modified to use the enhanced PKCS12_create()
4865 instead of the low level API.
4868 *) Extend ASN1 encoder to support indefinite length constructed
4869 encoding. This can output sequences tags and octet strings in
4870 this form. Modify pk7_asn1.c to support indefinite length
4871 encoding. This is experimental and needs additional code to
4872 be useful, such as an ASN1 bio and some enhanced streaming
4875 Extend template encode functionality so that tagging is passed
4876 down to the template encoder.
4879 *) Let 'openssl req' fail if an argument to '-newkey' is not
4880 recognized instead of using RSA as a default.
4883 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4884 As these are not official, they are not included in "ALL";
4885 the "ECCdraft" ciphersuite group alias can be used to select them.
4886 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4888 *) Add ECDH engine support.
4889 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4891 *) Add ECDH in new directory crypto/ecdh/.
4892 [Douglas Stebila (Sun Microsystems Laboratories)]
4894 *) Let BN_rand_range() abort with an error after 100 iterations
4895 without success (which indicates a broken PRNG).
4898 *) Change BN_mod_sqrt() so that it verifies that the input value
4899 is really the square of the return value. (Previously,
4900 BN_mod_sqrt would show GIGO behaviour.)
4903 *) Add named elliptic curves over binary fields from X9.62, SECG,
4904 and WAP/WTLS; add OIDs that were still missing.
4906 [Sheueling Chang Shantz and Douglas Stebila
4907 (Sun Microsystems Laboratories)]
4909 *) Extend the EC library for elliptic curves over binary fields
4910 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4913 EC_GF2m_simple_method
4917 EC_GROUP_new_curve_GF2m
4918 EC_GROUP_set_curve_GF2m
4919 EC_GROUP_get_curve_GF2m
4920 EC_POINT_set_affine_coordinates_GF2m
4921 EC_POINT_get_affine_coordinates_GF2m
4922 EC_POINT_set_compressed_coordinates_GF2m
4924 Point compression for binary fields is disabled by default for
4925 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4928 As binary polynomials are represented as BIGNUMs, various members
4929 of the EC_GROUP and EC_POINT data structures can be shared
4930 between the implementations for prime fields and binary fields;
4931 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4932 are essentially identical to their ..._GFp counterparts.
4933 (For simplicity, the '..._GFp' prefix has been dropped from
4934 various internal method names.)
4936 An internal 'field_div' method (similar to 'field_mul' and
4937 'field_sqr') has been added; this is used only for binary fields.
4939 [Sheueling Chang Shantz and Douglas Stebila
4940 (Sun Microsystems Laboratories)]
4942 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4943 through methods ('mul', 'precompute_mult').
4945 The generic implementations (now internally called 'ec_wNAF_mul'
4946 and 'ec_wNAF_precomputed_mult') remain the default if these
4947 methods are undefined.
4949 [Sheueling Chang Shantz and Douglas Stebila
4950 (Sun Microsystems Laboratories)]
4952 *) New function EC_GROUP_get_degree, which is defined through
4953 EC_METHOD. For curves over prime fields, this returns the bit
4954 length of the modulus.
4956 [Sheueling Chang Shantz and Douglas Stebila
4957 (Sun Microsystems Laboratories)]
4959 *) New functions EC_GROUP_dup, EC_POINT_dup.
4960 (These simply call ..._new and ..._copy).
4962 [Sheueling Chang Shantz and Douglas Stebila
4963 (Sun Microsystems Laboratories)]
4965 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4966 Polynomials are represented as BIGNUMs (where the sign bit is not
4967 used) in the following functions [macros]:
4970 BN_GF2m_sub [= BN_GF2m_add]
4971 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4972 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4973 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4975 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4976 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4977 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4978 BN_GF2m_cmp [= BN_ucmp]
4980 (Note that only the 'mod' functions are actually for fields GF(2^m).
4981 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4983 For some functions, an the irreducible polynomial defining a
4984 field can be given as an 'unsigned int[]' with strictly
4985 decreasing elements giving the indices of those bits that are set;
4986 i.e., p[] represents the polynomial
4987 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4989 p[0] > p[1] > ... > p[k] = 0.
4990 This applies to the following functions:
4995 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4996 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4998 BN_GF2m_mod_sqrt_arr
4999 BN_GF2m_mod_solve_quad_arr
5003 Conversion can be performed by the following functions:
5008 bntest.c has additional tests for binary polynomial arithmetic.
5010 Two implementations for BN_GF2m_mod_div() are available.
5011 The default algorithm simply uses BN_GF2m_mod_inv() and
5012 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5013 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5014 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5016 [Sheueling Chang Shantz and Douglas Stebila
5017 (Sun Microsystems Laboratories)]
5019 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5020 functionality is disabled at compile-time.
5021 [Douglas Stebila <douglas.stebila@sun.com>]
5023 *) Change default behaviour of 'openssl asn1parse' so that more
5024 information is visible when viewing, e.g., a certificate:
5026 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5027 mode the content of non-printable OCTET STRINGs is output in a
5028 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5029 avoid the appearance of a printable string.
5030 [Nils Larsch <nla@trustcenter.de>]
5032 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5034 EC_GROUP_set_asn1_flag()
5035 EC_GROUP_get_asn1_flag()
5036 EC_GROUP_set_point_conversion_form()
5037 EC_GROUP_get_point_conversion_form()
5038 These control ASN1 encoding details:
5039 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5040 has been set to OPENSSL_EC_NAMED_CURVE.
5041 - Points are encoded in uncompressed form by default; options for
5042 asn1_for are as for point2oct, namely
5043 POINT_CONVERSION_COMPRESSED
5044 POINT_CONVERSION_UNCOMPRESSED
5045 POINT_CONVERSION_HYBRID
5047 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5050 EC_GROUP_get0_seed()
5051 EC_GROUP_get_seed_len()
5052 This is used only for ASN1 purposes (so far).
5053 [Nils Larsch <nla@trustcenter.de>]
5055 *) Add 'field_type' member to EC_METHOD, which holds the NID
5056 of the appropriate field type OID. The new function
5057 EC_METHOD_get_field_type() returns this value.
5058 [Nils Larsch <nla@trustcenter.de>]
5063 EC_POINT_point2hex()
5064 EC_POINT_hex2point()
5065 providing useful interfaces to EC_POINT_point2oct() and
5066 EC_POINT_oct2point().
5067 [Nils Larsch <nla@trustcenter.de>]
5069 *) Change internals of the EC library so that the functions
5070 EC_GROUP_set_generator()
5071 EC_GROUP_get_generator()
5072 EC_GROUP_get_order()
5073 EC_GROUP_get_cofactor()
5074 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5075 to methods, which would lead to unnecessary code duplication when
5076 adding different types of curves.
5077 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5079 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5080 arithmetic, and such that modified wNAFs are generated
5081 (which avoid length expansion in many cases).
5084 *) Add a function EC_GROUP_check_discriminant() (defined via
5085 EC_METHOD) that verifies that the curve discriminant is non-zero.
5087 Add a function EC_GROUP_check() that makes some sanity tests
5088 on a EC_GROUP, its generator and order. This includes
5089 EC_GROUP_check_discriminant().
5090 [Nils Larsch <nla@trustcenter.de>]
5092 *) Add ECDSA in new directory crypto/ecdsa/.
5094 Add applications 'openssl ecparam' and 'openssl ecdsa'
5095 (these are based on 'openssl dsaparam' and 'openssl dsa').
5097 ECDSA support is also included in various other files across the
5098 library. Most notably,
5099 - 'openssl req' now has a '-newkey ecdsa:file' option;
5100 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5101 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5102 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5103 them suitable for ECDSA where domain parameters must be
5104 extracted before the specific public key;
5105 - ECDSA engine support has been added.
5106 [Nils Larsch <nla@trustcenter.de>]
5108 *) Include some named elliptic curves, and add OIDs from X9.62,
5109 SECG, and WAP/WTLS. Each curve can be obtained from the new
5111 EC_GROUP_new_by_curve_name(),
5112 and the list of available named curves can be obtained with
5113 EC_get_builtin_curves().
5114 Also add a 'curve_name' member to EC_GROUP objects, which can be
5116 EC_GROUP_set_curve_name()
5117 EC_GROUP_get_curve_name()
5118 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5120 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5121 was actually never needed) and in BN_mul(). The removal in BN_mul()
5122 required a small change in bn_mul_part_recursive() and the addition
5123 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5124 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5125 bn_sub_words() and bn_add_words() except they take arrays with
5129 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5131 *) Cleanse PEM buffers before freeing them since they may contain
5133 [Benjamin Bennett <ben@psc.edu>]
5135 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5136 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5137 authentication-only ciphersuites.
5140 *) Since AES128 and AES256 share a single mask bit in the logic of
5141 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5142 kludge to work properly if AES128 is available and AES256 isn't.
5145 *) Expand security boundary to match 1.1.1 module.
5148 *) Remove redundant features: hash file source, editing of test vectors
5149 modify fipsld to use external fips_premain.c signature.
5152 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5153 run algorithm test programs.
5156 *) Make algorithm test programs more tolerant of whitespace.
5159 *) Have SSL/TLS server implementation tolerate "mismatched" record
5160 protocol version while receiving ClientHello even if the
5161 ClientHello is fragmented. (The server can't insist on the
5162 particular protocol version it has chosen before the ServerHello
5163 message has informed the client about his choice.)
5166 *) Load error codes if they are not already present instead of using a
5167 static variable. This allows them to be cleanly unloaded and reloaded.
5170 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5172 *) Introduce limits to prevent malicious keys being able to
5173 cause a denial of service. (CVE-2006-2940)
5174 [Steve Henson, Bodo Moeller]
5176 *) Fix ASN.1 parsing of certain invalid structures that can result
5177 in a denial of service. (CVE-2006-2937) [Steve Henson]
5179 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5180 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5182 *) Fix SSL client code which could crash if connecting to a
5183 malicious SSLv2 server. (CVE-2006-4343)
5184 [Tavis Ormandy and Will Drewry, Google Security Team]
5186 *) Change ciphersuite string processing so that an explicit
5187 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5188 will no longer include "AES128-SHA"), and any other similar
5189 ciphersuite (same bitmap) from *other* protocol versions (so that
5190 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5191 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5192 changes from 0.9.8b and 0.9.8d.
5195 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5197 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5198 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5200 *) Change the Unix randomness entropy gathering to use poll() when
5201 possible instead of select(), since the latter has some
5202 undesirable limitations.
5203 [Darryl Miles via Richard Levitte and Bodo Moeller]
5205 *) Disable rogue ciphersuites:
5207 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5208 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5209 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5211 The latter two were purportedly from
5212 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5215 Also deactive the remaining ciphersuites from
5216 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5217 unofficial, and the ID has long expired.
5220 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5221 dual-core machines) and other potential thread-safety issues.
5224 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5226 *) Adapt fipsld and the build system to link against the validated FIPS
5227 module in FIPS mode.
5230 *) Fixes for VC++ 2005 build under Windows.
5233 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5234 from a Windows bash shell such as MSYS. It is autodetected from the
5235 "config" script when run from a VC++ environment. Modify standard VC++
5236 build to use fipscanister.o from the GNU make build.
5239 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5241 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5242 The value now differs depending on if you build for FIPS or not.
5243 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5244 safely run with a non-FIPSed libcrypto, as it may crash because of
5245 the difference induced by this change.
5248 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5250 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5251 (part of SSL_OP_ALL). This option used to disable the
5252 countermeasure against man-in-the-middle protocol-version
5253 rollback in the SSL 2.0 server implementation, which is a bad
5254 idea. (CVE-2005-2969)
5256 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5257 for Information Security, National Institute of Advanced Industrial
5258 Science and Technology [AIST], Japan)]
5260 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5261 mainly for FIPS compliance and not fully integrated at this stage.
5264 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5265 the exponentiation using a fixed-length exponent. (Otherwise,
5266 the information leaked through timing could expose the secret key
5267 after many signatures; cf. Bleichenbacher's attack on DSA with
5271 *) Make a new fixed-window mod_exp implementation the default for
5272 RSA, DSA, and DH private-key operations so that the sequence of
5273 squares and multiplies and the memory access pattern are
5274 independent of the particular secret key. This will mitigate
5275 cache-timing and potential related attacks.
5277 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5278 and this is automatically used by BN_mod_exp_mont() if the new flag
5279 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5280 will use this BN flag for private exponents unless the flag
5281 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5282 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5284 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5286 *) Change the client implementation for SSLv23_method() and
5287 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5288 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5289 (Previously, the SSL 2.0 backwards compatible Client Hello
5290 message format would be used even with SSL_OP_NO_SSLv2.)
5293 *) Add support for smime-type MIME parameter in S/MIME messages which some
5297 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5298 a threadsafe manner. Modify rsa code to use new function and add calls
5299 to dsa and dh code (which had race conditions before).
5302 *) Include the fixed error library code in the C error file definitions
5303 instead of fixing them up at runtime. This keeps the error code
5304 structures constant.
5307 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5309 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5312 *) Fixes for newer kerberos headers. NB: the casts are needed because
5313 the 'length' field is signed on one version and unsigned on another
5314 with no (?) obvious way to tell the difference, without these VC++
5315 complains. Also the "definition" of FAR (blank) is no longer included
5316 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5317 some needed definitions.
5320 *) Undo Cygwin change.
5323 *) Added support for proxy certificates according to RFC 3820.
5324 Because they may be a security thread to unaware applications,
5325 they must be explicitly allowed in run-time. See
5326 docs/HOWTO/proxy_certificates.txt for further information.
5329 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5331 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5332 server and client random values. Previously
5333 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5334 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5336 This change has negligible security impact because:
5338 1. Server and client random values still have 24 bytes of pseudo random
5341 2. Server and client random values are sent in the clear in the initial
5344 3. The master secret is derived using the premaster secret (48 bytes in
5345 size for static RSA ciphersuites) as well as client server and random
5348 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5351 [Stephen Henson, reported by UK NISCC]
5353 *) Use Windows randomness collection on Cygwin.
5356 *) Fix hang in EGD/PRNGD query when communication socket is closed
5357 prematurely by EGD/PRNGD.
5358 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5360 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5363 *) Back-port of selected performance improvements from development
5364 branch, as well as improved support for PowerPC platforms.
5367 *) Add lots of checks for memory allocation failure, error codes to indicate
5368 failure and freeing up memory if a failure occurs.
5369 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5371 *) Add new -passin argument to dgst.
5374 *) Perform some character comparisons of different types in X509_NAME_cmp:
5375 this is needed for some certificates that reencode DNs into UTF8Strings
5376 (in violation of RFC3280) and can't or wont issue name rollover
5380 *) Make an explicit check during certificate validation to see that
5381 the CA setting in each certificate on the chain is correct. As a
5382 side effect always do the following basic checks on extensions,
5383 not just when there's an associated purpose to the check:
5385 - if there is an unhandled critical extension (unless the user
5386 has chosen to ignore this fault)
5387 - if the path length has been exceeded (if one is set at all)
5388 - that certain extensions fit the associated purpose (if one has
5392 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5394 *) Avoid a race condition when CRLs are checked in a multi threaded
5395 environment. This would happen due to the reordering of the revoked
5396 entries during signature checking and serial number lookup. Now the
5397 encoding is cached and the serial number sort performed under a lock.
5398 Add new STACK function sk_is_sorted().
5401 *) Add Delta CRL to the extension code.
5404 *) Various fixes to s3_pkt.c so alerts are sent properly.
5405 [David Holmes <d.holmes@f5.com>]
5407 *) Reduce the chances of duplicate issuer name and serial numbers (in
5408 violation of RFC3280) using the OpenSSL certificate creation utilities.
5409 This is done by creating a random 64 bit value for the initial serial
5410 number when a serial number file is created or when a self signed
5411 certificate is created using 'openssl req -x509'. The initial serial
5412 number file is created using 'openssl x509 -next_serial' in CA.pl
5413 rather than being initialized to 1.
5416 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5418 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5419 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5420 [Joe Orton, Steve Henson]
5422 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5424 [Joe Orton, Steve Henson]
5426 *) Make it possible to have multiple active certificates with the same
5427 subject in the CA index file. This is done only if the keyword
5428 'unique_subject' is set to 'no' in the main CA section (default
5429 if 'CA_default') of the configuration file. The value is saved
5430 with the database itself in a separate index attribute file,
5431 named like the index file with '.attr' appended to the name.
5434 *) X509 verify fixes. Disable broken certificate workarounds when
5435 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5436 keyUsage extension present. Don't accept CRLs with unhandled critical
5437 extensions: since verify currently doesn't process CRL extensions this
5438 rejects a CRL with *any* critical extensions. Add new verify error codes
5442 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5443 A clarification of RFC2560 will require the use of OCTET STRINGs and
5444 some implementations cannot handle the current raw format. Since OpenSSL
5445 copies and compares OCSP nonces as opaque blobs without any attempt at
5446 parsing them this should not create any compatibility issues.
5449 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5450 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5451 this HMAC (and other) operations are several times slower than OpenSSL
5455 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5456 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5458 *) Use the correct content when signing type "other".
5461 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5463 *) Fix various bugs revealed by running the NISCC test suite:
5465 Stop out of bounds reads in the ASN1 code when presented with
5466 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5468 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5470 If verify callback ignores invalid public key errors don't try to check
5471 certificate signature with the NULL public key.
5475 *) New -ignore_err option in ocsp application to stop the server
5476 exiting on the first error in a request.
5479 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5480 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5484 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5485 extra data after the compression methods not only for TLS 1.0
5486 but also for SSL 3.0 (as required by the specification).
5487 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5489 *) Change X509_certificate_type() to mark the key as exported/exportable
5490 when it's 512 *bits* long, not 512 bytes.
5493 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5494 blocks during encryption.
5497 *) Various fixes to base64 BIO and non blocking I/O. On write
5498 flushes were not handled properly if the BIO retried. On read
5499 data was not being buffered properly and had various logic bugs.
5500 This also affects blocking I/O when the data being decoded is a
5504 *) Various S/MIME bugfixes and compatibility changes:
5505 output correct application/pkcs7 MIME type if
5506 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5507 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5508 of files as .eml work). Correctly handle very long lines in MIME
5512 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5514 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5515 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5516 a protocol version number mismatch like a decryption error
5517 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5520 *) Turn on RSA blinding by default in the default implementation
5521 to avoid a timing attack. Applications that don't want it can call
5522 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5523 They would be ill-advised to do so in most cases.
5524 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5526 *) Change RSA blinding code so that it works when the PRNG is not
5527 seeded (in this case, the secret RSA exponent is abused as
5528 an unpredictable seed -- if it is not unpredictable, there
5529 is no point in blinding anyway). Make RSA blinding thread-safe
5530 by remembering the creator's thread ID in rsa->blinding and
5531 having all other threads use local one-time blinding factors
5532 (this requires more computation than sharing rsa->blinding, but
5533 avoids excessive locking; and if an RSA object is not shared
5534 between threads, blinding will still be very fast).
5537 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5538 ENGINE as defaults for all supported algorithms irrespective of
5539 the 'flags' parameter. 'flags' is now honoured, so applications
5540 should make sure they are passing it correctly.
5543 *) Target "mingw" now allows native Windows code to be generated in
5544 the Cygwin environment as well as with the MinGW compiler.
5547 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5549 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5550 via timing by performing a MAC computation even if incorrect
5551 block cipher padding has been found. This is a countermeasure
5552 against active attacks where the attacker has to distinguish
5553 between bad padding and a MAC verification error. (CVE-2003-0078)
5555 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5556 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5557 Martin Vuagnoux (EPFL, Ilion)]
5559 *) Make the no-err option work as intended. The intention with no-err
5560 is not to have the whole error stack handling routines removed from
5561 libcrypto, it's only intended to remove all the function name and
5562 reason texts, thereby removing some of the footprint that may not
5563 be interesting if those errors aren't displayed anyway.
5565 NOTE: it's still possible for any application or module to have it's
5566 own set of error texts inserted. The routines are there, just not
5567 used by default when no-err is given.
5570 *) Add support for FreeBSD on IA64.
5571 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5573 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5574 Kerberos function mit_des_cbc_cksum(). Before this change,
5575 the value returned by DES_cbc_cksum() was like the one from
5576 mit_des_cbc_cksum(), except the bytes were swapped.
5577 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5579 *) Allow an application to disable the automatic SSL chain building.
5580 Before this a rather primitive chain build was always performed in
5581 ssl3_output_cert_chain(): an application had no way to send the
5582 correct chain if the automatic operation produced an incorrect result.
5584 Now the chain builder is disabled if either:
5586 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5588 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5590 The reasoning behind this is that an application would not want the
5591 auto chain building to take place if extra chain certificates are
5592 present and it might also want a means of sending no additional
5593 certificates (for example the chain has two certificates and the
5597 *) Add the possibility to build without the ENGINE framework.
5598 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5600 *) Under Win32 gmtime() can return NULL: check return value in
5601 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5604 *) DSA routines: under certain error conditions uninitialized BN objects
5605 could be freed. Solution: make sure initialization is performed early
5606 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5607 Nils Larsch <nla@trustcenter.de> via PR#459)
5610 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5611 checked on reconnect on the client side, therefore session resumption
5612 could still fail with a "ssl session id is different" error. This
5613 behaviour is masked when SSL_OP_ALL is used due to
5614 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5615 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5616 followup to PR #377.
5619 *) IA-32 assembler support enhancements: unified ELF targets, support
5620 for SCO/Caldera platforms, fix for Cygwin shared build.
5623 *) Add support for FreeBSD on sparc64. As a consequence, support for
5624 FreeBSD on non-x86 processors is separate from x86 processors on
5625 the config script, much like the NetBSD support.
5626 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5628 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5630 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5633 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5634 code (06) was taken as the first octet of the session ID and the last
5635 octet was ignored consequently. As a result SSLv2 client side session
5636 caching could not have worked due to the session ID mismatch between
5638 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5642 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5643 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5647 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5648 seems that in spite of existing for more than a year, many application
5649 author have done nothing to provide the necessary callbacks, which
5650 means that this particular engine will not work properly anywhere.
5651 This is a very unfortunate situation which forces us, in the name
5652 of usability, to give the hw_ncipher.c a static lock, which is part
5654 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5655 appear in 0.9.8 or later. We EXPECT application authors to have
5656 dealt properly with this when 0.9.8 is released (unless we actually
5657 make such changes in the libcrypto locking code that changes will
5658 have to be made anyway).
5661 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5662 octets have been read, EOF or an error occurs. Without this change
5663 some truncated ASN1 structures will not produce an error.
5666 *) Disable Heimdal support, since it hasn't been fully implemented.
5667 Still give the possibility to force the use of Heimdal, but with
5668 warnings and a request that patches get sent to openssl-dev.
5671 *) Add the VC-CE target, introduce the WINCE sysname, and add
5672 INSTALL.WCE and appropriate conditionals to make it build.
5673 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5675 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5676 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5677 edit numbers of the version.
5678 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5680 *) Introduce safe string copy and catenation functions
5681 (BUF_strlcpy() and BUF_strlcat()).
5682 [Ben Laurie (CHATS) and Richard Levitte]
5684 *) Avoid using fixed-size buffers for one-line DNs.
5685 [Ben Laurie (CHATS)]
5687 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5688 resizing buffers containing secrets, and use where appropriate.
5689 [Ben Laurie (CHATS)]
5691 *) Avoid using fixed size buffers for configuration file location.
5692 [Ben Laurie (CHATS)]
5694 *) Avoid filename truncation for various CA files.
5695 [Ben Laurie (CHATS)]
5697 *) Use sizeof in preference to magic numbers.
5698 [Ben Laurie (CHATS)]
5700 *) Avoid filename truncation in cert requests.
5701 [Ben Laurie (CHATS)]
5703 *) Add assertions to check for (supposedly impossible) buffer
5705 [Ben Laurie (CHATS)]
5707 *) Don't cache truncated DNS entries in the local cache (this could
5708 potentially lead to a spoofing attack).
5709 [Ben Laurie (CHATS)]
5711 *) Fix various buffers to be large enough for hex/decimal
5712 representations in a platform independent manner.
5713 [Ben Laurie (CHATS)]
5715 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5716 resizing buffers containing secrets, and use where appropriate.
5717 [Ben Laurie (CHATS)]
5719 *) Add BIO_indent() to avoid much slightly worrying code to do
5721 [Ben Laurie (CHATS)]
5723 *) Convert sprintf()/BIO_puts() to BIO_printf().
5724 [Ben Laurie (CHATS)]
5726 *) buffer_gets() could terminate with the buffer only half
5728 [Ben Laurie (CHATS)]
5730 *) Add assertions to prevent user-supplied crypto functions from
5731 overflowing internal buffers by having large block sizes, etc.
5732 [Ben Laurie (CHATS)]
5734 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5736 [Ben Laurie (CHATS)]
5738 *) Eliminate unused copy of key in RC4.
5739 [Ben Laurie (CHATS)]
5741 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5742 [Ben Laurie (CHATS)]
5744 *) Fix off-by-one error in EGD path.
5745 [Ben Laurie (CHATS)]
5747 *) If RANDFILE path is too long, ignore instead of truncating.
5748 [Ben Laurie (CHATS)]
5750 *) Eliminate unused and incorrectly sized X.509 structure
5752 [Ben Laurie (CHATS)]
5754 *) Eliminate unused and dangerous function knumber().
5755 [Ben Laurie (CHATS)]
5757 *) Eliminate unused and dangerous structure, KSSL_ERR.
5758 [Ben Laurie (CHATS)]
5760 *) Protect against overlong session ID context length in an encoded
5761 session object. Since these are local, this does not appear to be
5763 [Ben Laurie (CHATS)]
5765 *) Change from security patch (see 0.9.6e below) that did not affect
5766 the 0.9.6 release series:
5768 Remote buffer overflow in SSL3 protocol - an attacker could
5769 supply an oversized master key in Kerberos-enabled versions.
5771 [Ben Laurie (CHATS)]
5773 *) Change the SSL kerb5 codes to match RFC 2712.
5776 *) Make -nameopt work fully for req and add -reqopt switch.
5777 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5779 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5780 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5782 *) Make sure tests can be performed even if the corresponding algorithms
5783 have been removed entirely. This was also the last step to make
5784 OpenSSL compilable with DJGPP under all reasonable conditions.
5785 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5787 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5788 to allow version independent disabling of normally unselected ciphers,
5789 which may be activated as a side-effect of selecting a single cipher.
5791 (E.g., cipher list string "RSA" enables ciphersuites that are left
5792 out of "ALL" because they do not provide symmetric encryption.
5793 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5794 [Lutz Jaenicke, Bodo Moeller]
5796 *) Add appropriate support for separate platform-dependent build
5797 directories. The recommended way to make a platform-dependent
5798 build directory is the following (tested on Linux), maybe with
5801 # Place yourself outside of the OpenSSL source tree. In
5802 # this example, the environment variable OPENSSL_SOURCE
5803 # is assumed to contain the absolute OpenSSL source directory.
5804 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5805 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5806 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5807 mkdir -p `dirname $F`
5808 ln -s $OPENSSL_SOURCE/$F $F
5811 To be absolutely sure not to disturb the source tree, a "make clean"
5812 is a good thing. If it isn't successful, don't worry about it,
5813 it probably means the source directory is very clean.
5816 *) Make sure any ENGINE control commands make local copies of string
5817 pointers passed to them whenever necessary. Otherwise it is possible
5818 the caller may have overwritten (or deallocated) the original string
5819 data when a later ENGINE operation tries to use the stored values.
5820 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5822 *) Improve diagnostics in file reading and command-line digests.
5823 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5825 *) Add AES modes CFB and OFB to the object database. Correct an
5826 error in AES-CFB decryption.
5829 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5830 allows existing EVP_CIPHER_CTX structures to be reused after
5831 calling EVP_*Final(). This behaviour is used by encryption
5832 BIOs and some applications. This has the side effect that
5833 applications must explicitly clean up cipher contexts with
5834 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5837 *) Check the values of dna and dnb in bn_mul_recursive before calling
5838 bn_mul_comba (a non zero value means the a or b arrays do not contain
5839 n2 elements) and fallback to bn_mul_normal if either is not zero.
5842 *) Fix escaping of non-ASCII characters when using the -subj option
5843 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5846 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5847 form for "surname", serialNumber has no short form.
5848 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5849 therefore remove "mail" short name for "internet 7".
5850 The OID for unique identifiers in X509 certificates is
5851 x500UniqueIdentifier, not uniqueIdentifier.
5852 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5855 *) Add an "init" command to the ENGINE config module and auto initialize
5856 ENGINEs. Without any "init" command the ENGINE will be initialized
5857 after all ctrl commands have been executed on it. If init=1 the
5858 ENGINE is initailized at that point (ctrls before that point are run
5859 on the uninitialized ENGINE and after on the initialized one). If
5860 init=0 then the ENGINE will not be iniatialized at all.
5863 *) Fix the 'app_verify_callback' interface so that the user-defined
5864 argument is actually passed to the callback: In the
5865 SSL_CTX_set_cert_verify_callback() prototype, the callback
5866 declaration has been changed from
5869 int (*cb)(X509_STORE_CTX *,void *);
5870 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5871 i=s->ctx->app_verify_callback(&ctx)
5872 has been changed into
5873 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5875 To update applications using SSL_CTX_set_cert_verify_callback(),
5876 a dummy argument can be added to their callback functions.
5877 [D. K. Smetters <smetters@parc.xerox.com>]
5879 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5880 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5882 *) Add and OPENSSL_LOAD_CONF define which will cause
5883 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5884 This allows older applications to transparently support certain
5885 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5886 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5887 load the config file and OPENSSL_add_all_algorithms_conf() which will
5888 always load it have also been added.
5891 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5892 Adjust NIDs and EVP layer.
5893 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5895 *) Config modules support in openssl utility.
5897 Most commands now load modules from the config file,
5898 though in a few (such as version) this isn't done
5899 because it couldn't be used for anything.
5901 In the case of ca and req the config file used is
5902 the same as the utility itself: that is the -config
5903 command line option can be used to specify an
5907 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5908 use "openssl_conf" if filename is NULL use default openssl config file.
5911 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5912 config section name. Add a new flag to tolerate a missing config file
5913 and move code to CONF_modules_load_file().
5916 *) Support for crypto accelerator cards from Accelerated Encryption
5917 Processing, www.aep.ie. (Use engine 'aep')
5918 The support was copied from 0.9.6c [engine] and adapted/corrected
5919 to work with the new engine framework.
5920 [AEP Inc. and Richard Levitte]
5922 *) Support for SureWare crypto accelerator cards from Baltimore
5923 Technologies. (Use engine 'sureware')
5924 The support was copied from 0.9.6c [engine] and adapted
5925 to work with the new engine framework.
5928 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5929 make the newer ENGINE framework commands for the CHIL engine work.
5930 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5932 *) Make it possible to produce shared libraries on ReliantUNIX.
5933 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5935 *) Add the configuration target debug-linux-ppro.
5936 Make 'openssl rsa' use the general key loading routines
5937 implemented in apps.c, and make those routines able to
5938 handle the key format FORMAT_NETSCAPE and the variant
5940 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5942 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5943 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5945 *) Add -keyform to rsautl, and document -engine.
5946 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5948 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5949 BIO_R_NO_SUCH_FILE error code rather than the generic
5950 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5953 *) Add new functions
5955 ERR_peek_last_error_line
5956 ERR_peek_last_error_line_data.
5957 These are similar to
5960 ERR_peek_error_line_data,
5961 but report on the latest error recorded rather than the first one
5962 still in the error queue.
5963 [Ben Laurie, Bodo Moeller]
5965 *) default_algorithms option in ENGINE config module. This allows things
5967 default_algorithms = ALL
5968 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5971 *) Preliminary ENGINE config module.
5974 *) New experimental application configuration code.
5977 *) Change the AES code to follow the same name structure as all other
5978 symmetric ciphers, and behave the same way. Move everything to
5979 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5980 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5982 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5983 [Ben Laurie and Theo de Raadt]
5985 *) Add option to output public keys in req command.
5986 [Massimiliano Pala madwolf@openca.org]
5988 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5989 (up to about 10% better than before for P-192 and P-224).
5992 *) New functions/macros
5994 SSL_CTX_set_msg_callback(ctx, cb)
5995 SSL_CTX_set_msg_callback_arg(ctx, arg)
5996 SSL_set_msg_callback(ssl, cb)
5997 SSL_set_msg_callback_arg(ssl, arg)
5999 to request calling a callback function
6001 void cb(int write_p, int version, int content_type,
6002 const void *buf, size_t len, SSL *ssl, void *arg)
6004 whenever a protocol message has been completely received
6005 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6006 protocol version according to which the SSL library interprets
6007 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6008 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6009 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6010 specification (change_cipher_spec(20), alert(21), handshake(22)).
6011 'buf' and 'len' point to the actual message, 'ssl' to the
6012 SSL object, and 'arg' is the application-defined value set by
6013 SSL[_CTX]_set_msg_callback_arg().
6015 'openssl s_client' and 'openssl s_server' have new '-msg' options
6016 to enable a callback that displays all protocol messages.
6019 *) Change the shared library support so shared libraries are built as
6020 soon as the corresponding static library is finished, and thereby get
6021 openssl and the test programs linked against the shared library.
6022 This still only happens when the keyword "shard" has been given to
6023 the configuration scripts.
6025 NOTE: shared library support is still an experimental thing, and
6026 backward binary compatibility is still not guaranteed.
6027 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6029 *) Add support for Subject Information Access extension.
6030 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6032 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6033 additional bytes when new memory had to be allocated, not just
6034 when reusing an existing buffer.
6037 *) New command line and configuration option 'utf8' for the req command.
6038 This allows field values to be specified as UTF8 strings.
6041 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6042 runs for the former and machine-readable output for the latter.
6045 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6046 of the e-mail address in the DN (i.e., it will go into a certificate
6047 extension only). The new configuration file option 'email_in_dn = no'
6048 has the same effect.
6049 [Massimiliano Pala madwolf@openca.org]
6051 *) Change all functions with names starting with des_ to be starting
6052 with DES_ instead. Add wrappers that are compatible with libdes,
6053 but are named _ossl_old_des_*. Finally, add macros that map the
6054 des_* symbols to the corresponding _ossl_old_des_* if libdes
6055 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6056 desired, the des_* symbols will be mapped to DES_*, with one
6059 Since we provide two compatibility mappings, the user needs to
6060 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6061 compatibility is desired. The default (i.e., when that macro
6062 isn't defined) is OpenSSL 0.9.6c compatibility.
6064 There are also macros that enable and disable the support of old
6065 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6066 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6067 are defined, the default will apply: to support the old des routines.
6069 In either case, one must include openssl/des.h to get the correct
6070 definitions. Do not try to just include openssl/des_old.h, that
6073 NOTE: This is a major break of an old API into a new one. Software
6074 authors are encouraged to switch to the DES_ style functions. Some
6075 time in the future, des_old.h and the libdes compatibility functions
6076 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6077 default), and then completely removed.
6080 *) Test for certificates which contain unsupported critical extensions.
6081 If such a certificate is found during a verify operation it is
6082 rejected by default: this behaviour can be overridden by either
6083 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6084 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6085 X509_supported_extension() has also been added which returns 1 if a
6086 particular extension is supported.
6089 *) Modify the behaviour of EVP cipher functions in similar way to digests
6090 to retain compatibility with existing code.
6093 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6094 compatibility with existing code. In particular the 'ctx' parameter does
6095 not have to be to be initialized before the call to EVP_DigestInit() and
6096 it is tidied up after a call to EVP_DigestFinal(). New function
6097 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6098 EVP_MD_CTX_copy() changed to not require the destination to be
6099 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6100 requires the destination to be valid.
6102 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6103 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6106 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6107 so that complete 'Handshake' protocol structures are kept in memory
6108 instead of overwriting 'msg_type' and 'length' with 'body' data.
6111 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6112 [Massimo Santin via Richard Levitte]
6114 *) Major restructuring to the underlying ENGINE code. This includes
6115 reduction of linker bloat, separation of pure "ENGINE" manipulation
6116 (initialisation, etc) from functionality dealing with implementations
6117 of specific crypto iterfaces. This change also introduces integrated
6118 support for symmetric ciphers and digest implementations - so ENGINEs
6119 can now accelerate these by providing EVP_CIPHER and EVP_MD
6120 implementations of their own. This is detailed in crypto/engine/README
6121 as it couldn't be adequately described here. However, there are a few
6122 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6123 were changed in the original introduction of ENGINE code have now
6124 reverted back - the hooking from this code to ENGINE is now a good
6125 deal more passive and at run-time, operations deal directly with
6126 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6127 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6128 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6129 they were not being used by the framework as there is no concept of a
6130 BIGNUM_METHOD and they could not be generalised to the new
6131 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6132 ENGINE_cpy() has been removed as it cannot be consistently defined in
6136 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6139 *) Change mkdef.pl to sort symbols that get the same entry number,
6140 and make sure the automatically generated functions ERR_load_*
6141 become part of libeay.num as well.
6144 *) New function SSL_renegotiate_pending(). This returns true once
6145 renegotiation has been requested (either SSL_renegotiate() call
6146 or HelloRequest/ClientHello received from the peer) and becomes
6147 false once a handshake has been completed.
6148 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6149 sends a HelloRequest, but does not ensure that a handshake takes
6150 place. SSL_renegotiate_pending() is useful for checking if the
6151 client has followed the request.)
6154 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6155 By default, clients may request session resumption even during
6156 renegotiation (if session ID contexts permit); with this option,
6157 session resumption is possible only in the first handshake.
6159 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6160 more bits available for options that should not be part of
6161 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6164 *) Add some demos for certificate and certificate request creation.
6167 *) Make maximum certificate chain size accepted from the peer application
6168 settable (SSL*_get/set_max_cert_list()), as proposed by
6169 "Douglas E. Engert" <deengert@anl.gov>.
6172 *) Add support for shared libraries for Unixware-7
6173 (Boyd Lynn Gerber <gerberb@zenez.com>).
6176 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6177 be done prior to destruction. Use this to unload error strings from
6178 ENGINEs that load their own error strings. NB: This adds two new API
6179 functions to "get" and "set" this destroy handler in an ENGINE.
6182 *) Alter all existing ENGINE implementations (except "openssl" and
6183 "openbsd") to dynamically instantiate their own error strings. This
6184 makes them more flexible to be built both as statically-linked ENGINEs
6185 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6186 Also, add stub code to each that makes building them as self-contained
6187 shared-libraries easier (see README.ENGINE).
6190 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6191 implementations into applications that are completely implemented in
6192 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6193 commands that can be used to configure what shared-library to load and
6194 to control aspects of the way it is handled. Also, made an update to
6195 the README.ENGINE file that brings its information up-to-date and
6196 provides some information and instructions on the "dynamic" ENGINE
6197 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6200 *) Make it possible to unload ranges of ERR strings with a new
6201 "ERR_unload_strings" function.
6204 *) Add a copy() function to EVP_MD.
6207 *) Make EVP_MD routines take a context pointer instead of just the
6208 md_data void pointer.
6211 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6212 that the digest can only process a single chunk of data
6213 (typically because it is provided by a piece of
6214 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6215 is only going to provide a single chunk of data, and hence the
6216 framework needn't accumulate the data for oneshot drivers.
6219 *) As with "ERR", make it possible to replace the underlying "ex_data"
6220 functions. This change also alters the storage and management of global
6221 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6222 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6223 index counters. The API functions that use this state have been changed
6224 to take a "class_index" rather than pointers to the class's local STACK
6225 and counter, and there is now an API function to dynamically create new
6226 classes. This centralisation allows us to (a) plug a lot of the
6227 thread-safety problems that existed, and (b) makes it possible to clean
6228 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6229 such data would previously have always leaked in application code and
6230 workarounds were in place to make the memory debugging turn a blind eye
6231 to it. Application code that doesn't use this new function will still
6232 leak as before, but their memory debugging output will announce it now
6233 rather than letting it slide.
6235 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6236 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6237 has a return value to indicate success or failure.
6240 *) Make it possible to replace the underlying "ERR" functions such that the
6241 global state (2 LHASH tables and 2 locks) is only used by the "default"
6242 implementation. This change also adds two functions to "get" and "set"
6243 the implementation prior to it being automatically set the first time
6244 any other ERR function takes place. Ie. an application can call "get",
6245 pass the return value to a module it has just loaded, and that module
6246 can call its own "set" function using that value. This means the
6247 module's "ERR" operations will use (and modify) the error state in the
6248 application and not in its own statically linked copy of OpenSSL code.
6251 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6252 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6253 the operation, and provides a more encapsulated way for external code
6254 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6255 to use these functions rather than manually incrementing the counts.
6257 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6260 *) Add EVP test program.
6263 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6266 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6267 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6268 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6269 These allow a CRL to be built without having to access X509_CRL fields
6270 directly. Modify 'ca' application to use new functions.
6273 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6274 bug workarounds. Rollback attack detection is a security feature.
6275 The problem will only arise on OpenSSL servers when TLSv1 is not
6276 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6277 Software authors not wanting to support TLSv1 will have special reasons
6278 for their choice and can explicitly enable this option.
6279 [Bodo Moeller, Lutz Jaenicke]
6281 *) Rationalise EVP so it can be extended: don't include a union of
6282 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6283 (similar to those existing for EVP_CIPHER_CTX).
6288 EVP_MD_CTX_init(&md); /* new function call */
6289 EVP_DigestInit(&md, EVP_sha1());
6290 EVP_DigestUpdate(&md, in, len);
6291 EVP_DigestFinal(&md, out, NULL);
6292 EVP_MD_CTX_cleanup(&md); /* new function call */
6296 *) Make DES key schedule conform to the usual scheme, as well as
6297 correcting its structure. This means that calls to DES functions
6298 now have to pass a pointer to a des_key_schedule instead of a
6299 plain des_key_schedule (which was actually always a pointer
6302 des_key_schedule ks;
6304 des_set_key_checked(..., &ks);
6305 des_ncbc_encrypt(..., &ks, ...);
6307 (Note that a later change renames 'des_...' into 'DES_...'.)
6310 *) Initial reduction of linker bloat: the use of some functions, such as
6311 PEM causes large amounts of unused functions to be linked in due to
6312 poor organisation. For example pem_all.c contains every PEM function
6313 which has a knock on effect of linking in large amounts of (unused)
6314 ASN1 code. Grouping together similar functions and splitting unrelated
6315 functions prevents this.
6318 *) Cleanup of EVP macros.
6321 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6322 correct _ecb suffix.
6325 *) Add initial OCSP responder support to ocsp application. The
6326 revocation information is handled using the text based index
6327 use by the ca application. The responder can either handle
6328 requests generated internally, supplied in files (for example
6329 via a CGI script) or using an internal minimal server.
6332 *) Add configuration choices to get zlib compression for TLS.
6335 *) Changes to Kerberos SSL for RFC 2712 compliance:
6336 1. Implemented real KerberosWrapper, instead of just using
6337 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6338 2. Implemented optional authenticator field of KerberosWrapper.
6340 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6341 and authenticator structs; see crypto/krb5/.
6343 Generalized Kerberos calls to support multiple Kerberos libraries.
6344 [Vern Staats <staatsvr@asc.hpc.mil>,
6345 Jeffrey Altman <jaltman@columbia.edu>
6346 via Richard Levitte]
6348 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6349 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6350 values for each of the key sizes rather than having just
6351 parameters (and 'speed' generating keys each time).
6354 *) Speed up EVP routines.
6357 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6358 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6359 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6360 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6362 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6363 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6364 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6367 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6369 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6372 *) Added the OS2-EMX target.
6373 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6375 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6376 to support NCONF routines in extension code. New function CONF_set_nconf()
6377 to allow functions which take an NCONF to also handle the old LHASH
6378 structure: this means that the old CONF compatible routines can be
6379 retained (in particular wrt extensions) without having to duplicate the
6380 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6383 *) Enhance the general user interface with mechanisms for inner control
6384 and with possibilities to have yes/no kind of prompts.
6387 *) Change all calls to low level digest routines in the library and
6388 applications to use EVP. Add missing calls to HMAC_cleanup() and
6389 don't assume HMAC_CTX can be copied using memcpy().
6390 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6392 *) Add the possibility to control engines through control names but with
6393 arbitrary arguments instead of just a string.
6394 Change the key loaders to take a UI_METHOD instead of a callback
6395 function pointer. NOTE: this breaks binary compatibility with earlier
6396 versions of OpenSSL [engine].
6397 Adapt the nCipher code for these new conditions and add a card insertion
6401 *) Enhance the general user interface with mechanisms to better support
6402 dialog box interfaces, application-defined prompts, the possibility
6403 to use defaults (for example default passwords from somewhere else)
6404 and interrupts/cancellations.
6407 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6408 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6411 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6412 tidy up some unnecessarily weird code in 'sk_new()').
6413 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6415 *) Change the key loading routines for ENGINEs to use the same kind
6416 callback (pem_password_cb) as all other routines that need this
6420 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6421 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6422 than this minimum value is recommended.
6425 *) New random seeder for OpenVMS, using the system process statistics
6426 that are easily reachable.
6429 *) Windows apparently can't transparently handle global
6430 variables defined in DLLs. Initialisations such as:
6432 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6434 wont compile. This is used by the any applications that need to
6435 declare their own ASN1 modules. This was fixed by adding the option
6436 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6437 needed for static libraries under Win32.
6440 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6441 setting of purpose and trust fields. New X509_STORE trust and
6442 purpose functions and tidy up setting in other SSL functions.
6445 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6446 structure. These are inherited by X509_STORE_CTX when it is
6447 initialised. This allows various defaults to be set in the
6448 X509_STORE structure (such as flags for CRL checking and custom
6449 purpose or trust settings) for functions which only use X509_STORE_CTX
6450 internally such as S/MIME.
6452 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6453 trust settings if they are not set in X509_STORE. This allows X509_STORE
6454 purposes and trust (in S/MIME for example) to override any set by default.
6456 Add command line options for CRL checking to smime, s_client and s_server
6460 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6461 are set then the CRL is looked up in the X509_STORE structure and
6462 its validity and signature checked, then if the certificate is found
6463 in the CRL the verify fails with a revoked error.
6465 Various new CRL related callbacks added to X509_STORE_CTX structure.
6467 Command line options added to 'verify' application to support this.
6469 This needs some additional work, such as being able to handle multiple
6470 CRLs with different times, extension based lookup (rather than just
6471 by subject name) and ultimately more complete V2 CRL extension
6475 *) Add a general user interface API (crypto/ui/). This is designed
6476 to replace things like des_read_password and friends (backward
6477 compatibility functions using this new API are provided).
6478 The purpose is to remove prompting functions from the DES code
6479 section as well as provide for prompting through dialog boxes in
6480 a window system and the like.
6483 *) Add "ex_data" support to ENGINE so implementations can add state at a
6484 per-structure level rather than having to store it globally.
6487 *) Make it possible for ENGINE structures to be copied when retrieved by
6488 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6489 This causes the "original" ENGINE structure to act like a template,
6490 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6491 operational state can be localised to each ENGINE structure, despite the
6492 fact they all share the same "methods". New ENGINE structures returned in
6493 this case have no functional references and the return value is the single
6494 structural reference. This matches the single structural reference returned
6495 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6499 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6500 needs to match any other type at all we need to manually clear the
6504 *) Changes to the "openssl engine" utility to include;
6505 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6506 about an ENGINE's available control commands.
6507 - executing control commands from command line arguments using the
6508 '-pre' and '-post' switches. '-post' is only used if '-t' is
6509 specified and the ENGINE is successfully initialised. The syntax for
6510 the individual commands are colon-separated, for example;
6511 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6514 *) New dynamic control command support for ENGINEs. ENGINEs can now
6515 declare their own commands (numbers), names (strings), descriptions,
6516 and input types for run-time discovery by calling applications. A
6517 subset of these commands are implicitly classed as "executable"
6518 depending on their input type, and only these can be invoked through
6519 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6520 can be based on user input, config files, etc). The distinction is
6521 that "executable" commands cannot return anything other than a boolean
6522 result and can only support numeric or string input, whereas some
6523 discoverable commands may only be for direct use through
6524 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6525 pointers, or other custom uses. The "executable" commands are to
6526 support parameterisations of ENGINE behaviour that can be
6527 unambiguously defined by ENGINEs and used consistently across any
6528 OpenSSL-based application. Commands have been added to all the
6529 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6530 control over shared-library paths without source code alterations.
6533 *) Changed all ENGINE implementations to dynamically allocate their
6534 ENGINEs rather than declaring them statically. Apart from this being
6535 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6536 this also allows the implementations to compile without using the
6537 internal engine_int.h header.
6540 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6541 'const' value. Any code that should be able to modify a RAND_METHOD
6542 should already have non-const pointers to it (ie. they should only
6543 modify their own ones).
6546 *) Made a variety of little tweaks to the ENGINE code.
6547 - "atalla" and "ubsec" string definitions were moved from header files
6548 to C code. "nuron" string definitions were placed in variables
6549 rather than hard-coded - allowing parameterisation of these values
6550 later on via ctrl() commands.
6551 - Removed unused "#if 0"'d code.
6552 - Fixed engine list iteration code so it uses ENGINE_free() to release
6553 structural references.
6554 - Constified the RAND_METHOD element of ENGINE structures.
6555 - Constified various get/set functions as appropriate and added
6556 missing functions (including a catch-all ENGINE_cpy that duplicates
6557 all ENGINE values onto a new ENGINE except reference counts/state).
6558 - Removed NULL parameter checks in get/set functions. Setting a method
6559 or function to NULL is a way of cancelling out a previously set
6560 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6561 and doesn't justify the extra error symbols and code.
6562 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6563 flags from engine_int.h to engine.h.
6564 - Changed prototypes for ENGINE handler functions (init(), finish(),
6565 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6568 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6569 to the algorithm using long division. The binary algorithm can be
6570 used only if the modulus is odd. On 32-bit systems, it is faster
6571 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6572 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6573 up to 450 bits. In 64-bit environments, the binary algorithm
6574 appears to be advantageous for much longer moduli; here we use it
6575 for moduli up to 2048 bits.
6578 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6579 could not support the combine flag in choice fields.
6582 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6583 extensions from a certificate request to the certificate.
6586 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6587 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6588 file: this allows the display of the certificate about to be
6589 signed to be customised, to allow certain fields to be included
6590 or excluded and extension details. The old system didn't display
6591 multicharacter strings properly, omitted fields not in the policy
6592 and couldn't display additional details such as extensions.
6595 *) Function EC_POINTs_mul for multiple scalar multiplication
6596 of an arbitrary number of elliptic curve points
6597 \sum scalars[i]*points[i],
6598 optionally including the generator defined for the EC_GROUP:
6599 scalar*generator + \sum scalars[i]*points[i].
6601 EC_POINT_mul is a simple wrapper function for the typical case
6602 that the point list has just one item (besides the optional
6606 *) First EC_METHODs for curves over GF(p):
6608 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6609 operations and provides various method functions that can also
6610 operate with faster implementations of modular arithmetic.
6612 EC_GFp_mont_method() reuses most functions that are part of
6613 EC_GFp_simple_method, but uses Montgomery arithmetic.
6615 [Bodo Moeller; point addition and point doubling
6616 implementation directly derived from source code provided by
6617 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6619 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6620 crypto/ec/ec_lib.c):
6622 Curves are EC_GROUP objects (with an optional group generator)
6623 based on EC_METHODs that are built into the library.
6625 Points are EC_POINT objects based on EC_GROUP objects.
6627 Most of the framework would be able to handle curves over arbitrary
6628 finite fields, but as there are no obvious types for fields other
6629 than GF(p), some functions are limited to that for now.
6632 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6633 that the file contains a complete HTTP response.
6636 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6637 change the def and num file printf format specifier from "%-40sXXX"
6638 to "%-39s XXX". The latter will always guarantee a space after the
6639 field while the former will cause them to run together if the field
6640 is 40 of more characters long.
6643 *) Constify the cipher and digest 'method' functions and structures
6644 and modify related functions to take constant EVP_MD and EVP_CIPHER
6648 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6649 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6652 *) Modify EVP_Digest*() routines so they now return values. Although the
6653 internal software routines can never fail additional hardware versions
6657 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6659 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6660 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6663 ERR_R_NESTED_ASN1_ERROR
6665 ERR_R_MISSING_ASN1_EOS
6666 were 4 .. 9, conflicting with
6667 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6669 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6670 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6672 Add new error code 'ERR_R_INTERNAL_ERROR'.
6675 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6679 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6680 sets the subject name for a new request or supersedes the
6681 subject name in a given request. Formats that can be parsed are
6682 'CN=Some Name, OU=myOU, C=IT'
6684 'CN=Some Name/OU=myOU/C=IT'.
6686 Add options '-batch' and '-verbose' to 'openssl req'.
6687 [Massimiliano Pala <madwolf@hackmasters.net>]
6689 *) Introduce the possibility to access global variables through
6690 functions on platform were that's the best way to handle exporting
6691 global variables in shared libraries. To enable this functionality,
6692 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6693 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6694 is normally done by Configure or something similar).
6696 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6697 in the source file (foo.c) like this:
6699 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6700 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6702 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6703 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6705 OPENSSL_DECLARE_GLOBAL(int,foo);
6706 #define foo OPENSSL_GLOBAL_REF(foo)
6707 OPENSSL_DECLARE_GLOBAL(double,bar);
6708 #define bar OPENSSL_GLOBAL_REF(bar)
6710 The #defines are very important, and therefore so is including the
6711 header file everywhere where the defined globals are used.
6713 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6714 of ASN.1 items, but that structure is a bit different.
6716 The largest change is in util/mkdef.pl which has been enhanced with
6717 better and easier to understand logic to choose which symbols should
6718 go into the Windows .def files as well as a number of fixes and code
6719 cleanup (among others, algorithm keywords are now sorted
6720 lexicographically to avoid constant rewrites).
6723 *) In BN_div() keep a copy of the sign of 'num' before writing the
6724 result to 'rm' because if rm==num the value will be overwritten
6725 and produce the wrong result if 'num' is negative: this caused
6726 problems with BN_mod() and BN_nnmod().
6729 *) Function OCSP_request_verify(). This checks the signature on an
6730 OCSP request and verifies the signer certificate. The signer
6731 certificate is just checked for a generic purpose and OCSP request
6735 *) Add OCSP_check_validity() function to check the validity of OCSP
6736 responses. OCSP responses are prepared in real time and may only
6737 be a few seconds old. Simply checking that the current time lies
6738 between thisUpdate and nextUpdate max reject otherwise valid responses
6739 caused by either OCSP responder or client clock inaccuracy. Instead
6740 we allow thisUpdate and nextUpdate to fall within a certain period of
6741 the current time. The age of the response can also optionally be
6742 checked. Two new options -validity_period and -status_age added to
6746 *) If signature or public key algorithm is unrecognized print out its
6747 OID rather that just UNKNOWN.
6750 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6751 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6752 ID to be generated from the issuer certificate alone which can then be
6753 passed to OCSP_id_issuer_cmp().
6756 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6757 ASN1 modules to export functions returning ASN1_ITEM pointers
6758 instead of the ASN1_ITEM structures themselves. This adds several
6759 new macros which allow the underlying ASN1 function/structure to
6760 be accessed transparently. As a result code should not use ASN1_ITEM
6761 references directly (such as &X509_it) but instead use the relevant
6762 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6763 use of the new ASN1 code on platforms where exporting structures
6764 is problematical (for example in shared libraries) but exporting
6765 functions returning pointers to structures is not.
6768 *) Add support for overriding the generation of SSL/TLS session IDs.
6769 These callbacks can be registered either in an SSL_CTX or per SSL.
6770 The purpose of this is to allow applications to control, if they wish,
6771 the arbitrary values chosen for use as session IDs, particularly as it
6772 can be useful for session caching in multiple-server environments. A
6773 command-line switch for testing this (and any client code that wishes
6774 to use such a feature) has been added to "s_server".
6775 [Geoff Thorpe, Lutz Jaenicke]
6777 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6778 of the form '#if defined(...) || defined(...) || ...' and
6779 '#if !defined(...) && !defined(...) && ...'. This also avoids
6780 the growing number of special cases it was previously handling.
6783 *) Make all configuration macros available for application by making
6784 sure they are available in opensslconf.h, by giving them names starting
6785 with "OPENSSL_" to avoid conflicts with other packages and by making
6786 sure e_os2.h will cover all platform-specific cases together with
6788 Additionally, it is now possible to define configuration/platform-
6789 specific names (called "system identities"). In the C code, these
6790 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6791 macro with the name beginning with "OPENSSL_SYS_", which is determined
6792 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6796 *) New option -set_serial to 'req' and 'x509' this allows the serial
6797 number to use to be specified on the command line. Previously self
6798 signed certificates were hard coded with serial number 0 and the
6799 CA options of 'x509' had to use a serial number in a file which was
6803 *) New options to 'ca' utility to support V2 CRL entry extensions.
6804 Currently CRL reason, invalidity date and hold instruction are
6805 supported. Add new CRL extensions to V3 code and some new objects.
6808 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6809 disable standard block padding (aka PKCS#5 padding) in the EVP
6810 API, which was previously mandatory. This means that the data is
6811 not padded in any way and so the total length much be a multiple
6812 of the block size, otherwise an error occurs.
6815 *) Initial (incomplete) OCSP SSL support.
6818 *) New function OCSP_parse_url(). This splits up a URL into its host,
6819 port and path components: primarily to parse OCSP URLs. New -url
6820 option to ocsp utility.
6823 *) New nonce behavior. The return value of OCSP_check_nonce() now
6824 reflects the various checks performed. Applications can decide
6825 whether to tolerate certain situations such as an absent nonce
6826 in a response when one was present in a request: the ocsp application
6827 just prints out a warning. New function OCSP_add1_basic_nonce()
6828 this is to allow responders to include a nonce in a response even if
6829 the request is nonce-less.
6832 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6833 skipped when using openssl x509 multiple times on a single input file,
6834 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6837 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6838 set string type: to handle setting ASN1_TIME structures. Fix ca
6839 utility to correctly initialize revocation date of CRLs.
6842 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6843 the clients preferred ciphersuites and rather use its own preferences.
6844 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6845 Internet Explorer by ensuring unchanged hash method during stepup.
6846 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6849 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6850 to aes and add a new 'exist' option to print out symbols that don't
6854 *) Additional options to ocsp utility to allow flags to be set and
6855 additional certificates supplied.
6858 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6859 OCSP client a number of certificate to only verify the response
6863 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6864 handle the new API. Currently only ECB, CBC modes supported. Add new
6867 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6868 Encryption Standard (AES) Ciphersuites for Transport Layer
6869 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6870 not enabled by default and were not part of the "ALL" ciphersuite
6871 alias because they were not yet official; they could be
6872 explicitly requested by specifying the "AESdraft" ciphersuite
6873 group alias. In the final release of OpenSSL 0.9.7, the group
6874 alias is called "AES" and is part of "ALL".)
6875 [Ben Laurie, Steve Henson, Bodo Moeller]
6877 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6878 request to response.
6881 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6882 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6883 extract information from a certificate request. OCSP_response_create()
6884 creates a response and optionally adds a basic response structure.
6885 OCSP_basic_add1_status() adds a complete single response to a basic
6886 response and returns the OCSP_SINGLERESP structure just added (to allow
6887 extensions to be included for example). OCSP_basic_add1_cert() adds a
6888 certificate to a basic response and OCSP_basic_sign() signs a basic
6889 response with various flags. New helper functions ASN1_TIME_check()
6890 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6891 (converts ASN1_TIME to GeneralizedTime).
6894 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6895 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6896 structure from a certificate. X509_pubkey_digest() digests the public_key
6897 contents: this is used in various key identifiers.
6900 *) Make sk_sort() tolerate a NULL argument.
6901 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6903 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6904 passed by the function are trusted implicitly. If any of them signed the
6905 response then it is assumed to be valid and is not verified.
6908 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6909 to data. This was previously part of the PKCS7 ASN1 code. This
6910 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6911 [Steve Henson, reported by Kenneth R. Robinette
6912 <support@securenetterm.com>]
6914 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6915 routines: without these tracing memory leaks is very painful.
6916 Fix leaks in PKCS12 and PKCS7 routines.
6919 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6920 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6921 effectively meant GeneralizedTime would never be used. Now it
6922 is initialised to -1 but X509_time_adj() now has to check the value
6923 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6924 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6925 [Steve Henson, reported by Kenneth R. Robinette
6926 <support@securenetterm.com>]
6928 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6929 result in a zero length in the ASN1_INTEGER structure which was
6930 not consistent with the structure when d2i_ASN1_INTEGER() was used
6931 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6932 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6933 where it did not print out a minus for negative ASN1_INTEGER.
6936 *) Add summary printout to ocsp utility. The various functions which
6937 convert status values to strings have been renamed to:
6938 OCSP_response_status_str(), OCSP_cert_status_str() and
6939 OCSP_crl_reason_str() and are no longer static. New options
6940 to verify nonce values and to disable verification. OCSP response
6941 printout format cleaned up.
6944 *) Add additional OCSP certificate checks. These are those specified
6945 in RFC2560. This consists of two separate checks: the CA of the
6946 certificate being checked must either be the OCSP signer certificate
6947 or the issuer of the OCSP signer certificate. In the latter case the
6948 OCSP signer certificate must contain the OCSP signing extended key
6949 usage. This check is performed by attempting to match the OCSP
6950 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6951 in the OCSP_CERTID structures of the response.
6954 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6955 and related routines. This uses the standard OpenSSL certificate
6956 verify routines to perform initial checks (just CA validity) and
6957 to obtain the certificate chain. Then additional checks will be
6958 performed on the chain. Currently the root CA is checked to see
6959 if it is explicitly trusted for OCSP signing. This is used to set
6960 a root CA as a global signing root: that is any certificate that
6961 chains to that CA is an acceptable OCSP signing certificate.
6964 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6965 extensions from a separate configuration file.
6966 As when reading extensions from the main configuration file,
6967 the '-extensions ...' option may be used for specifying the
6969 [Massimiliano Pala <madwolf@comune.modena.it>]
6971 *) New OCSP utility. Allows OCSP requests to be generated or
6972 read. The request can be sent to a responder and the output
6973 parsed, outputed or printed in text form. Not complete yet:
6974 still needs to check the OCSP response validity.
6977 *) New subcommands for 'openssl ca':
6978 'openssl ca -status <serial>' prints the status of the cert with
6979 the given serial number (according to the index file).
6980 'openssl ca -updatedb' updates the expiry status of certificates
6982 [Massimiliano Pala <madwolf@comune.modena.it>]
6984 *) New '-newreq-nodes' command option to CA.pl. This is like
6985 '-newreq', but calls 'openssl req' with the '-nodes' option
6986 so that the resulting key is not encrypted.
6987 [Damien Miller <djm@mindrot.org>]
6989 *) New configuration for the GNU Hurd.
6990 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6992 *) Initial code to implement OCSP basic response verify. This
6993 is currently incomplete. Currently just finds the signer's
6994 certificate and verifies the signature on the response.
6997 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6998 value of OPENSSLDIR. This is available via the new '-d' option
6999 to 'openssl version', and is also included in 'openssl version -a'.
7002 *) Allowing defining memory allocation callbacks that will be given
7003 file name and line number information in additional arguments
7004 (a const char* and an int). The basic functionality remains, as
7005 well as the original possibility to just replace malloc(),
7006 realloc() and free() by functions that do not know about these
7007 additional arguments. To register and find out the current
7008 settings for extended allocation functions, the following
7009 functions are provided:
7011 CRYPTO_set_mem_ex_functions
7012 CRYPTO_set_locked_mem_ex_functions
7013 CRYPTO_get_mem_ex_functions
7014 CRYPTO_get_locked_mem_ex_functions
7016 These work the same way as CRYPTO_set_mem_functions and friends.
7017 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7018 extended allocation function is enabled.
7019 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7020 a conventional allocation function is enabled.
7021 [Richard Levitte, Bodo Moeller]
7023 *) Finish off removing the remaining LHASH function pointer casts.
7024 There should no longer be any prototype-casting required when using
7025 the LHASH abstraction, and any casts that remain are "bugs". See
7026 the callback types and macros at the head of lhash.h for details
7027 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7030 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7031 If /dev/[u]random devices are not available or do not return enough
7032 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7034 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7035 /etc/entropy will be queried once each in this sequence, quering stops
7036 when enough entropy was collected without querying more sockets.
7039 *) Change the Unix RAND_poll() variant to be able to poll several
7040 random devices, as specified by DEVRANDOM, until a sufficient amount
7041 of data has been collected. We spend at most 10 ms on each file
7042 (select timeout) and read in non-blocking mode. DEVRANDOM now
7043 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7044 (previously it was just the string "/dev/urandom"), so on typical
7045 platforms the 10 ms delay will never occur.
7046 Also separate out the Unix variant to its own file, rand_unix.c.
7047 For VMS, there's a currently-empty rand_vms.c.
7050 *) Move OCSP client related routines to ocsp_cl.c. These
7051 provide utility functions which an application needing
7052 to issue a request to an OCSP responder and analyse the
7053 response will typically need: as opposed to those which an
7054 OCSP responder itself would need which will be added later.
7056 OCSP_request_sign() signs an OCSP request with an API similar
7057 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7058 response. OCSP_response_get1_basic() extracts basic response
7059 from response. OCSP_resp_find_status(): finds and extracts status
7060 information from an OCSP_CERTID structure (which will be created
7061 when the request structure is built). These are built from lower
7062 level functions which work on OCSP_SINGLERESP structures but
7063 wont normally be used unless the application wishes to examine
7064 extensions in the OCSP response for example.
7066 Replace nonce routines with a pair of functions.
7067 OCSP_request_add1_nonce() adds a nonce value and optionally
7068 generates a random value. OCSP_check_nonce() checks the
7069 validity of the nonce in an OCSP response.
7072 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7073 This doesn't copy the supplied OCSP_CERTID and avoids the
7074 need to free up the newly created id. Change return type
7075 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7076 This can then be used to add extensions to the request.
7077 Deleted OCSP_request_new(), since most of its functionality
7078 is now in OCSP_REQUEST_new() (and the case insensitive name
7079 clash) apart from the ability to set the request name which
7080 will be added elsewhere.
7083 *) Update OCSP API. Remove obsolete extensions argument from
7084 various functions. Extensions are now handled using the new
7085 OCSP extension code. New simple OCSP HTTP function which
7086 can be used to send requests and parse the response.
7089 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7090 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7091 uses the special reorder version of SET OF to sort the attributes
7092 and reorder them to match the encoded order. This resolves a long
7093 standing problem: a verify on a PKCS7 structure just after signing
7094 it used to fail because the attribute order did not match the
7095 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7096 it uses the received order. This is necessary to tolerate some broken
7097 software that does not order SET OF. This is handled by encoding
7098 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7099 to produce the required SET OF.
7102 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7103 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7104 files to get correct declarations of the ASN.1 item variables.
7107 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7108 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7109 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7110 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7111 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7112 ASN1_ITEM and no wrapper functions.
7115 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7116 replace the old function pointer based I/O routines. Change most of
7117 the *_d2i_bio() and *_d2i_fp() functions to use these.
7120 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7121 lines, recognice more "algorithms" that can be deselected, and make
7122 it complain about algorithm deselection that isn't recognised.
7125 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7126 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7127 to use new functions. Add NO_ASN1_OLD which can be set to remove
7128 some old style ASN1 functions: this can be used to determine if old
7129 code will still work when these eventually go away.
7132 *) New extension functions for OCSP structures, these follow the
7133 same conventions as certificates and CRLs.
7136 *) New function X509V3_add1_i2d(). This automatically encodes and
7137 adds an extension. Its behaviour can be customised with various
7138 flags to append, replace or delete. Various wrappers added for
7139 certifcates and CRLs.
7142 *) Fix to avoid calling the underlying ASN1 print routine when
7143 an extension cannot be parsed. Correct a typo in the
7144 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7147 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7148 entries for variables.
7151 *) Add functionality to apps/openssl.c for detecting locking
7152 problems: As the program is single-threaded, all we have
7153 to do is register a locking callback using an array for
7154 storing which locks are currently held by the program.
7157 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7158 SSL_get_ex_data_X509_STORE_idx(), which is used in
7159 ssl_verify_cert_chain() and thus can be called at any time
7160 during TLS/SSL handshakes so that thread-safety is essential.
7161 Unfortunately, the ex_data design is not at all suited
7162 for multi-threaded use, so it probably should be abolished.
7165 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7166 [Broadcom, tweaked and integrated by Geoff Thorpe]
7168 *) Move common extension printing code to new function
7169 X509V3_print_extensions(). Reorganise OCSP print routines and
7170 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7173 *) New function X509_signature_print() to remove duplication in some
7177 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7178 set (this was treated exactly the same as SET OF previously). This
7179 is used to reorder the STACK representing the structure to match the
7180 encoding. This will be used to get round a problem where a PKCS7
7181 structure which was signed could not be verified because the STACK
7182 order did not reflect the encoded order.
7185 *) Reimplement the OCSP ASN1 module using the new code.
7188 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7189 for its ASN1 operations. The old style function pointers still exist
7190 for now but they will eventually go away.
7193 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7194 completely replaces the old ASN1 functionality with a table driven
7195 encoder and decoder which interprets an ASN1_ITEM structure describing
7196 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7197 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7198 has also been converted to the new form.
7201 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7202 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7203 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7204 for negative moduli.
7207 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7208 of not touching the result's sign bit.
7211 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7215 *) Changed the LHASH code to use prototypes for callbacks, and created
7216 macros to declare and implement thin (optionally static) functions
7217 that provide type-safety and avoid function pointer casting for the
7218 type-specific callbacks.
7221 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7223 [Veers Staats <staatsvr@asc.hpc.mil>,
7224 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7226 *) Reformat the FAQ so the different questions and answers can be divided
7227 in sections depending on the subject.
7230 *) Have the zlib compression code load ZLIB.DLL dynamically under
7234 *) New function BN_mod_sqrt for computing square roots modulo a prime
7235 (using the probabilistic Tonelli-Shanks algorithm unless
7236 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7237 be handled deterministically).
7238 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7240 *) Make BN_mod_inverse faster by explicitly handling small quotients
7241 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7242 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7245 *) New function BN_kronecker.
7248 *) Fix BN_gcd so that it works on negative inputs; the result is
7249 positive unless both parameters are zero.
7250 Previously something reasonably close to an infinite loop was
7251 possible because numbers could be growing instead of shrinking
7252 in the implementation of Euclid's algorithm.
7255 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7256 sign of the number in question.
7258 Fix BN_is_word(a,w) to work correctly for w == 0.
7260 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7261 because its test if the absolute value of 'a' equals 'w'.
7262 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7263 it exists mostly for use in the implementations of BN_is_zero(),
7264 BN_is_one(), and BN_is_word().
7267 *) New function BN_swap.
7270 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7271 the exponentiation functions are more likely to produce reasonable
7272 results on negative inputs.
7275 *) Change BN_mod_mul so that the result is always non-negative.
7276 Previously, it could be negative if one of the factors was negative;
7277 I don't think anyone really wanted that behaviour.
7280 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7281 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7282 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7283 and add new functions:
7292 BN_mod_lshift1_quick
7296 These functions always generate non-negative results.
7298 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7299 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7301 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7302 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7303 be reduced modulo m.
7304 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7307 The following entry accidentally appeared in the CHANGES file
7308 distributed with OpenSSL 0.9.7. The modifications described in
7309 it do *not* apply to OpenSSL 0.9.7.
7311 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7312 was actually never needed) and in BN_mul(). The removal in BN_mul()
7313 required a small change in bn_mul_part_recursive() and the addition
7314 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7315 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7316 bn_sub_words() and bn_add_words() except they take arrays with
7321 *) In 'openssl passwd', verify passwords read from the terminal
7322 unless the '-salt' option is used (which usually means that
7323 verification would just waste user's time since the resulting
7324 hash is going to be compared with some given password hash)
7325 or the new '-noverify' option is used.
7327 This is an incompatible change, but it does not affect
7328 non-interactive use of 'openssl passwd' (passwords on the command
7329 line, '-stdin' option, '-in ...' option) and thus should not
7333 *) Remove all references to RSAref, since there's no more need for it.
7336 *) Make DSO load along a path given through an environment variable
7337 (SHLIB_PATH) with shl_load().
7340 *) Constify the ENGINE code as a result of BIGNUM constification.
7341 Also constify the RSA code and most things related to it. In a
7342 few places, most notable in the depth of the ASN.1 code, ugly
7343 casts back to non-const were required (to be solved at a later
7347 *) Make it so the openssl application has all engines loaded by default.
7350 *) Constify the BIGNUM routines a little more.
7353 *) Add the following functions:
7355 ENGINE_load_cswift()
7357 ENGINE_load_atalla()
7359 ENGINE_load_builtin_engines()
7361 That way, an application can itself choose if external engines that
7362 are built-in in OpenSSL shall ever be used or not. The benefit is
7363 that applications won't have to be linked with libdl or other dso
7364 libraries unless it's really needed.
7366 Changed 'openssl engine' to load all engines on demand.
7367 Changed the engine header files to avoid the duplication of some
7368 declarations (they differed!).
7371 *) 'openssl engine' can now list capabilities.
7374 *) Better error reporting in 'openssl engine'.
7377 *) Never call load_dh_param(NULL) in s_server.
7380 *) Add engine application. It can currently list engines by name and
7381 identity, and test if they are actually available.
7384 *) Improve RPM specification file by forcing symbolic linking and making
7385 sure the installed documentation is also owned by root.root.
7386 [Damien Miller <djm@mindrot.org>]
7388 *) Give the OpenSSL applications more possibilities to make use of
7389 keys (public as well as private) handled by engines.
7392 *) Add OCSP code that comes from CertCo.
7395 *) Add VMS support for the Rijndael code.
7398 *) Added untested support for Nuron crypto accelerator.
7401 *) Add support for external cryptographic devices. This code was
7402 previously distributed separately as the "engine" branch.
7403 [Geoff Thorpe, Richard Levitte]
7405 *) Rework the filename-translation in the DSO code. It is now possible to
7406 have far greater control over how a "name" is turned into a filename
7407 depending on the operating environment and any oddities about the
7408 different shared library filenames on each system.
7411 *) Support threads on FreeBSD-elf in Configure.
7414 *) Fix for SHA1 assembly problem with MASM: it produces
7415 warnings about corrupt line number information when assembling
7416 with debugging information. This is caused by the overlapping
7418 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7421 NCONF_get_number() has no error checking at all. As a replacement,
7422 NCONF_get_number_e() is defined (_e for "error checking") and is
7423 promoted strongly. The old NCONF_get_number is kept around for
7424 binary backward compatibility.
7425 Make it possible for methods to load from something other than a BIO,
7426 by providing a function pointer that is given a name instead of a BIO.
7427 For example, this could be used to load configuration data from an
7431 *) Fix for non blocking accept BIOs. Added new I/O special reason
7432 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7433 with non blocking I/O was not possible because no retry code was
7434 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7438 *) Added the beginnings of Rijndael support.
7441 *) Fix for bug in DirectoryString mask setting. Add support for
7442 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7443 to allow certificate printing to more controllable, additional
7444 'certopt' option to 'x509' to allow new printing options to be
7448 *) Clean old EAY MD5 hack from e_os.h.
7451 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7453 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7454 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7455 [Joe Orton, Steve Henson]
7457 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7459 *) Fix additional bug revealed by the NISCC test suite:
7461 Stop bug triggering large recursion when presented with
7462 certain ASN.1 tags (CVE-2003-0851)
7465 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7467 *) Fix various bugs revealed by running the NISCC test suite:
7469 Stop out of bounds reads in the ASN1 code when presented with
7470 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7472 If verify callback ignores invalid public key errors don't try to check
7473 certificate signature with the NULL public key.
7477 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7478 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7482 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7483 extra data after the compression methods not only for TLS 1.0
7484 but also for SSL 3.0 (as required by the specification).
7485 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7487 *) Change X509_certificate_type() to mark the key as exported/exportable
7488 when it's 512 *bits* long, not 512 bytes.
7491 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7493 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7494 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7495 a protocol version number mismatch like a decryption error
7496 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7499 *) Turn on RSA blinding by default in the default implementation
7500 to avoid a timing attack. Applications that don't want it can call
7501 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7502 They would be ill-advised to do so in most cases.
7503 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7505 *) Change RSA blinding code so that it works when the PRNG is not
7506 seeded (in this case, the secret RSA exponent is abused as
7507 an unpredictable seed -- if it is not unpredictable, there
7508 is no point in blinding anyway). Make RSA blinding thread-safe
7509 by remembering the creator's thread ID in rsa->blinding and
7510 having all other threads use local one-time blinding factors
7511 (this requires more computation than sharing rsa->blinding, but
7512 avoids excessive locking; and if an RSA object is not shared
7513 between threads, blinding will still be very fast).
7516 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7518 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7519 via timing by performing a MAC computation even if incorrrect
7520 block cipher padding has been found. This is a countermeasure
7521 against active attacks where the attacker has to distinguish
7522 between bad padding and a MAC verification error. (CVE-2003-0078)
7524 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7525 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7526 Martin Vuagnoux (EPFL, Ilion)]
7528 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7530 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7531 memory from it's contents. This is done with a counter that will
7532 place alternating values in each byte. This can be used to solve
7533 two issues: 1) the removal of calls to memset() by highly optimizing
7534 compilers, and 2) cleansing with other values than 0, since those can
7535 be read through on certain media, for example a swap space on disk.
7538 *) Bugfix: client side session caching did not work with external caching,
7539 because the session->cipher setting was not restored when reloading
7540 from the external cache. This problem was masked, when
7541 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7542 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7545 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7546 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7547 [Zeev Lieber <zeev-l@yahoo.com>]
7549 *) Undo an undocumented change introduced in 0.9.6e which caused
7550 repeated calls to OpenSSL_add_all_ciphers() and
7551 OpenSSL_add_all_digests() to be ignored, even after calling
7555 *) Change the default configuration reader to deal with last line not
7556 being properly terminated.
7559 *) Change X509_NAME_cmp() so it applies the special rules on handling
7560 DN values that are of type PrintableString, as well as RDNs of type
7561 emailAddress where the value has the type ia5String.
7562 [stefank@valicert.com via Richard Levitte]
7564 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7565 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7566 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7567 the bitwise-OR of the two for use by the majority of applications
7568 wanting this behaviour, and update the docs. The documented
7569 behaviour and actual behaviour were inconsistent and had been
7570 changing anyway, so this is more a bug-fix than a behavioural
7572 [Geoff Thorpe, diagnosed by Nadav Har'El]
7574 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7575 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7578 *) Fix initialization code race conditions in
7579 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7580 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7581 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7582 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7583 ssl2_get_cipher_by_char(),
7584 ssl3_get_cipher_by_char().
7585 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7587 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7588 the cached sessions are flushed, as the remove_cb() might use ex_data
7589 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7590 (see [openssl.org #212]).
7591 [Geoff Thorpe, Lutz Jaenicke]
7593 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7594 length, instead of the encoding length to d2i_ASN1_OBJECT.
7597 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7599 *) [In 0.9.6g-engine release:]
7600 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7601 [Lynn Gazis <lgazis@rainbow.com>]
7603 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7605 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7606 and get fix the header length calculation.
7607 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7608 Alon Kantor <alonk@checkpoint.com> (and others),
7611 *) Use proper error handling instead of 'assertions' in buffer
7612 overflow checks added in 0.9.6e. This prevents DoS (the
7613 assertions could call abort()).
7614 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7616 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7618 *) Add various sanity checks to asn1_get_length() to reject
7619 the ASN1 length bytes if they exceed sizeof(long), will appear
7620 negative or the content length exceeds the length of the
7622 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7624 *) Fix cipher selection routines: ciphers without encryption had no flags
7625 for the cipher strength set and where therefore not handled correctly
7626 by the selection routines (PR #130).
7629 *) Fix EVP_dsa_sha macro.
7633 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7634 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7635 that was added in OpenSSL 0.9.6d.
7637 As the countermeasure turned out to be incompatible with some
7638 broken SSL implementations, the new option is part of SSL_OP_ALL.
7639 SSL_OP_ALL is usually employed when compatibility with weird SSL
7640 implementations is desired (e.g. '-bugs' option to 's_client' and
7641 's_server'), so the new option is automatically set in many
7645 *) Changes in security patch:
7647 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7648 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7649 Air Force Materiel Command, USAF, under agreement number
7652 *) Add various sanity checks to asn1_get_length() to reject
7653 the ASN1 length bytes if they exceed sizeof(long), will appear
7654 negative or the content length exceeds the length of the
7655 supplied buffer. (CVE-2002-0659)
7656 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7658 *) Assertions for various potential buffer overflows, not known to
7660 [Ben Laurie (CHATS)]
7662 *) Various temporary buffers to hold ASCII versions of integers were
7663 too small for 64 bit platforms. (CVE-2002-0655)
7664 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7666 *) Remote buffer overflow in SSL3 protocol - an attacker could
7667 supply an oversized session ID to a client. (CVE-2002-0656)
7668 [Ben Laurie (CHATS)]
7670 *) Remote buffer overflow in SSL2 protocol - an attacker could
7671 supply an oversized client master key. (CVE-2002-0656)
7672 [Ben Laurie (CHATS)]
7674 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7676 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7677 encoded as NULL) with id-dsa-with-sha1.
7678 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7680 *) Check various X509_...() return values in apps/req.c.
7681 [Nils Larsch <nla@trustcenter.de>]
7683 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7684 an end-of-file condition would erronously be flagged, when the CRLF
7685 was just at the end of a processed block. The bug was discovered when
7686 processing data through a buffering memory BIO handing the data to a
7687 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7688 <ptsekov@syntrex.com> and Nedelcho Stanev.
7691 *) Implement a countermeasure against a vulnerability recently found
7692 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7693 before application data chunks to avoid the use of known IVs
7694 with data potentially chosen by the attacker.
7697 *) Fix length checks in ssl3_get_client_hello().
7700 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7701 to prevent ssl3_read_internal() from incorrectly assuming that
7702 ssl3_read_bytes() found application data while handshake
7703 processing was enabled when in fact s->s3->in_read_app_data was
7704 merely automatically cleared during the initial handshake.
7705 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7707 *) Fix object definitions for Private and Enterprise: they were not
7708 recognized in their shortname (=lowercase) representation. Extend
7709 obj_dat.pl to issue an error when using undefined keywords instead
7710 of silently ignoring the problem (Svenning Sorensen
7711 <sss@sss.dnsalias.net>).
7714 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7715 generators, i.e. generators other than 2 and 5. (Previously, the
7716 code did not properly initialise the 'add' and 'rem' values to
7717 BN_generate_prime().)
7719 In the new general case, we do not insist that 'generator' is
7720 actually a primitive root: This requirement is rather pointless;
7721 a generator of the order-q subgroup is just as good, if not
7725 *) Map new X509 verification errors to alerts. Discovered and submitted by
7726 Tom Wu <tom@arcot.com>.
7729 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7730 returning non-zero before the data has been completely received
7731 when using non-blocking I/O.
7732 [Bodo Moeller; problem pointed out by John Hughes]
7734 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7735 [Ben Laurie, Lutz Jaenicke]
7737 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7738 Yoram Zahavi <YoramZ@gilian.com>).
7741 *) Add information about CygWin 1.3 and on, and preserve proper
7742 configuration for the versions before that.
7743 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7745 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7746 check whether we deal with a copy of a session and do not delete from
7747 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7748 <izhar@checkpoint.com>.
7751 *) Do not store session data into the internal session cache, if it
7752 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7753 flag is set). Proposed by Aslam <aslam@funk.com>.
7756 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7760 *) [In 0.9.6d-engine release:]
7761 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7762 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7764 *) Add the configuration target linux-s390x.
7765 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7767 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7768 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7769 variable as an indication that a ClientHello message has been
7770 received. As the flag value will be lost between multiple
7771 invocations of ssl3_accept when using non-blocking I/O, the
7772 function may not be aware that a handshake has actually taken
7773 place, thus preventing a new session from being added to the
7776 To avoid this problem, we now set s->new_session to 2 instead of
7777 using a local variable.
7778 [Lutz Jaenicke, Bodo Moeller]
7780 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7781 if the SSL_R_LENGTH_MISMATCH error is detected.
7782 [Geoff Thorpe, Bodo Moeller]
7784 *) New 'shared_ldflag' column in Configure platform table.
7787 *) Fix EVP_CIPHER_mode macro.
7788 ["Dan S. Camper" <dan@bti.net>]
7790 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7791 type, we must throw them away by setting rr->length to 0.
7792 [D P Chang <dpc@qualys.com>]
7794 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7796 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7797 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7798 worked incorrectly for those cases where range = 10..._2 and
7799 3*range is two bits longer than range.)
7802 *) Only add signing time to PKCS7 structures if it is not already
7806 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7807 OBJ_ld_ce should be OBJ_id_ce.
7808 Also some ip-pda OIDs in crypto/objects/objects.txt were
7809 incorrect (cf. RFC 3039).
7810 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7812 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7813 returns early because it has nothing to do.
7814 [Andy Schneider <andy.schneider@bjss.co.uk>]
7816 *) [In 0.9.6c-engine release:]
7817 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7818 [Andy Schneider <andy.schneider@bjss.co.uk>]
7820 *) [In 0.9.6c-engine release:]
7821 Add support for Cryptographic Appliance's keyserver technology.
7822 (Use engine 'keyclient')
7823 [Cryptographic Appliances and Geoff Thorpe]
7825 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7826 is called via tools/c89.sh because arguments have to be
7827 rearranged (all '-L' options must appear before the first object
7829 [Richard Shapiro <rshapiro@abinitio.com>]
7831 *) [In 0.9.6c-engine release:]
7832 Add support for Broadcom crypto accelerator cards, backported
7834 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7836 *) [In 0.9.6c-engine release:]
7837 Add support for SureWare crypto accelerator cards from
7838 Baltimore Technologies. (Use engine 'sureware')
7839 [Baltimore Technologies and Mark Cox]
7841 *) [In 0.9.6c-engine release:]
7842 Add support for crypto accelerator cards from Accelerated
7843 Encryption Processing, www.aep.ie. (Use engine 'aep')
7844 [AEP Inc. and Mark Cox]
7846 *) Add a configuration entry for gcc on UnixWare.
7847 [Gary Benson <gbenson@redhat.com>]
7849 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7850 messages are stored in a single piece (fixed-length part and
7851 variable-length part combined) and fix various bugs found on the way.
7854 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7855 instead. BIO_gethostbyname() does not know what timeouts are
7856 appropriate, so entries would stay in cache even when they have
7858 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7860 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7861 faced with a pathologically small ClientHello fragment that does
7862 not contain client_version: Instead of aborting with an error,
7863 simply choose the highest available protocol version (i.e.,
7864 TLS 1.0 unless it is disabled). In practice, ClientHello
7865 messages are never sent like this, but this change gives us
7866 strictly correct behaviour at least for TLS.
7869 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7870 never resets s->method to s->ctx->method when called from within
7871 one of the SSL handshake functions.
7872 [Bodo Moeller; problem pointed out by Niko Baric]
7874 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7875 (sent using the client's version number) if client_version is
7876 smaller than the protocol version in use. Also change
7877 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7878 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7879 the client will at least see that alert.
7882 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7886 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7887 client receives HelloRequest while in a handshake.
7888 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7890 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7891 should end in 'break', not 'goto end' which circumvents various
7892 cleanups done in state SSL_ST_OK. But session related stuff
7893 must be disabled for SSL_ST_OK in the case that we just sent a
7896 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7897 before just sending a HelloRequest.
7898 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7900 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7901 reveal whether illegal block cipher padding was found or a MAC
7902 verification error occurred. (Neither SSLerr() codes nor alerts
7903 are directly visible to potential attackers, but the information
7904 may leak via logfiles.)
7906 Similar changes are not required for the SSL 2.0 implementation
7907 because the number of padding bytes is sent in clear for SSL 2.0,
7908 and the extra bytes are just ignored. However ssl/s2_pkt.c
7909 failed to verify that the purported number of padding bytes is in
7913 *) Add OpenUNIX-8 support including shared libraries
7914 (Boyd Lynn Gerber <gerberb@zenez.com>).
7917 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7918 'wristwatch attack' using huge encoding parameters (cf.
7919 James H. Manger's CRYPTO 2001 paper). Note that the
7920 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7921 encoding parameters and hence was not vulnerable.
7924 *) BN_sqr() bug fix.
7925 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7927 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7928 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7929 followed by modular reduction.
7930 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7932 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7933 equivalent based on BN_pseudo_rand() instead of BN_rand().
7936 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7937 This function was broken, as the check for a new client hello message
7938 to handle SGC did not allow these large messages.
7939 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7942 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7945 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7946 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7949 *) Rework the configuration and shared library support for Tru64 Unix.
7950 The configuration part makes use of modern compiler features and
7951 still retains old compiler behavior for those that run older versions
7952 of the OS. The shared library support part includes a variant that
7953 uses the RPATH feature, and is available through the special
7954 configuration target "alpha-cc-rpath", which will never be selected
7956 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7958 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7959 with the same message size as in ssl3_get_certificate_request().
7960 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7961 messages might inadvertently be reject as too long.
7962 [Petr Lampa <lampa@fee.vutbr.cz>]
7964 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7967 *) Modified SSL library such that the verify_callback that has been set
7968 specificly for an SSL object with SSL_set_verify() is actually being
7969 used. Before the change, a verify_callback set with this function was
7970 ignored and the verify_callback() set in the SSL_CTX at the time of
7971 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7972 to allow the necessary settings.
7975 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7976 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7977 done automatically (in contradiction to the requirements of the C
7978 standard). This made problems when used from OpenSSH.
7981 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7982 dh->length and always used
7984 BN_rand_range(priv_key, dh->p).
7986 BN_rand_range() is not necessary for Diffie-Hellman, and this
7987 specific range makes Diffie-Hellman unnecessarily inefficient if
7988 dh->length (recommended exponent length) is much smaller than the
7989 length of dh->p. We could use BN_rand_range() if the order of
7990 the subgroup was stored in the DH structure, but we only have
7995 BN_rand(priv_key, l, ...)
7997 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8003 RSA_eay_public_encrypt
8004 RSA_eay_private_decrypt
8005 RSA_eay_private_encrypt (signing)
8006 RSA_eay_public_decrypt (signature verification)
8008 (default implementations for RSA_public_encrypt,
8009 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8010 always reject numbers >= n.
8013 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8014 to synchronize access to 'locking_thread'. This is necessary on
8015 systems where access to 'locking_thread' (an 'unsigned long'
8016 variable) is not atomic.
8019 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8020 *before* setting the 'crypto_lock_rand' flag. The previous code had
8021 a race condition if 0 is a valid thread ID.
8022 [Travis Vitek <vitek@roguewave.com>]
8024 *) Add support for shared libraries under Irix.
8025 [Albert Chin-A-Young <china@thewrittenword.com>]
8027 *) Add configuration option to build on Linux on both big-endian and
8029 [Ralf Baechle <ralf@uni-koblenz.de>]
8031 *) Add the possibility to create shared libraries on HP-UX.
8034 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8036 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8037 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8038 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8039 PRNG state recovery was possible based on the output of
8040 one PRNG request appropriately sized to gain knowledge on
8041 'md' followed by enough consecutive 1-byte PRNG requests
8042 to traverse all of 'state'.
8044 1. When updating 'md_local' (the current thread's copy of 'md')
8045 during PRNG output generation, hash all of the previous
8046 'md_local' value, not just the half used for PRNG output.
8048 2. Make the number of bytes from 'state' included into the hash
8049 independent from the number of PRNG bytes requested.
8051 The first measure alone would be sufficient to avoid
8052 Markku-Juhani's attack. (Actually it had never occurred
8053 to me that the half of 'md_local' used for chaining was the
8054 half from which PRNG output bytes were taken -- I had always
8055 assumed that the secret half would be used.) The second
8056 measure makes sure that additional data from 'state' is never
8057 mixed into 'md_local' in small portions; this heuristically
8058 further strengthens the PRNG.
8061 *) Fix crypto/bn/asm/mips3.s.
8064 *) When only the key is given to "enc", the IV is undefined. Print out
8065 an error message in this case.
8068 *) Handle special case when X509_NAME is empty in X509 printing routines.
8071 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8072 positive and less than q.
8075 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8076 used: it isn't thread safe and the add_lock_callback should handle
8078 [Paul Rose <Paul.Rose@bridge.com>]
8080 *) Verify that incoming data obeys the block size in
8081 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8085 [Ulf Möller, Bodo Möller]
8087 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8088 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8089 when fixing the server behaviour for backwards-compatible 'client
8090 hello' messages. (Note that the attack is impractical against
8091 SSL 3.0 and TLS 1.0 anyway because length and version checking
8092 means that the probability of guessing a valid ciphertext is
8093 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8096 Before 0.9.5, the countermeasure (hide the error by generating a
8097 random 'decryption result') did not work properly because
8098 ERR_clear_error() was missing, meaning that SSL_get_error() would
8099 detect the supposedly ignored error.
8101 Both problems are now fixed.
8104 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8105 (previously it was 1024).
8108 *) Fix for compatibility mode trust settings: ignore trust settings
8109 unless some valid trust or reject settings are present.
8112 *) Fix for blowfish EVP: its a variable length cipher.
8115 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8116 parameters in DSA public key structures and return an error in the
8117 DSA routines if parameters are absent.
8120 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8121 in the current directory if neither $RANDFILE nor $HOME was set.
8122 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8123 caused some confusion to Windows users who haven't defined $HOME.
8124 Thus RAND_file_name() is changed again: e_os.h can define a
8125 DEFAULT_HOME, which will be used if $HOME is not set.
8126 For Windows, we use "C:"; on other platforms, we still require
8127 environment variables.
8129 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8130 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8131 having multiple threads call RAND_poll() concurrently.
8134 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8135 combination of a flag and a thread ID variable.
8136 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8137 flag), *other* threads can enter ssleay_add_bytes without obeying
8138 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8139 that they do not hold after the first thread unsets add_do_not_lock).
8142 *) Change bctest again: '-x' expressions are not available in all
8146 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8148 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8149 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8151 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8152 the default extension for executables, if any. Also, make the perl
8153 scripts that use symlink() to test if it really exists and use "cp"
8154 if it doesn't. All this made OpenSSL compilable and installable in
8158 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8159 If SEQUENCE is length is indefinite just set c->slen to the total
8160 amount of data available.
8161 [Steve Henson, reported by shige@FreeBSD.org]
8162 [This change does not apply to 0.9.7.]
8164 *) Change bctest to avoid here-documents inside command substitution
8165 (workaround for FreeBSD /bin/sh bug).
8166 For compatibility with Ultrix, avoid shell functions (introduced
8167 in the bctest version that searches along $PATH).
8170 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8171 with des_encrypt() defined on some operating systems, like Solaris
8175 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8176 On the Importance of Eliminating Errors in Cryptographic
8177 Computations, J. Cryptology 14 (2001) 2, 101-119,
8178 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8181 *) MIPS assembler BIGNUM division bug fix.
8184 *) Disabled incorrect Alpha assembler code.
8187 *) Fix PKCS#7 decode routines so they correctly update the length
8188 after reading an EOC for the EXPLICIT tag.
8190 [This change does not apply to 0.9.7.]
8192 *) Fix bug in PKCS#12 key generation routines. This was triggered
8193 if a 3DES key was generated with a 0 initial byte. Include
8194 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8195 (but broken) behaviour.
8198 *) Enhance bctest to search for a working bc along $PATH and print
8200 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8202 *) Fix memory leaks in err.c: free err_data string if necessary;
8203 don't write to the wrong index in ERR_set_error_data.
8206 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8210 *) Replace rdtsc with _emit statements for VC++ version 5.
8211 [Jeremy Cooper <jeremy@baymoo.org>]
8213 *) Make it possible to reuse SSLv2 sessions.
8216 *) In copy_email() check for >= 0 as a return value for
8217 X509_NAME_get_index_by_NID() since 0 is a valid index.
8218 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8220 *) Avoid coredump with unsupported or invalid public keys by checking if
8221 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8222 PKCS7_verify() fails with non detached data.
8225 *) Don't use getenv in library functions when run as setuid/setgid.
8226 New function OPENSSL_issetugid().
8229 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8230 due to incorrect handling of multi-threading:
8232 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8234 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8236 3. Count how many times MemCheck_off() has been called so that
8237 nested use can be treated correctly. This also avoids
8238 inband-signalling in the previous code (which relied on the
8239 assumption that thread ID 0 is impossible).
8242 *) Add "-rand" option also to s_client and s_server.
8245 *) Fix CPU detection on Irix 6.x.
8246 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8247 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8249 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8252 [This change does not apply to 0.9.7.]
8254 *) Use the cached encoding of an X509_NAME structure rather than
8255 copying it. This is apparently the reason for the libsafe "errors"
8256 but the code is actually correct.
8259 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8260 Bleichenbacher's DSA attack.
8261 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8262 to be set and top=0 forces the highest bit to be set; top=-1 is new
8263 and leaves the highest bit random.
8264 [Ulf Moeller, Bodo Moeller]
8266 *) In the NCONF_...-based implementations for CONF_... queries
8267 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8268 a temporary CONF structure with the data component set to NULL
8269 (which gives segmentation faults in lh_retrieve).
8270 Instead, use NULL for the CONF pointer in CONF_get_string and
8271 CONF_get_number (which may use environment variables) and directly
8272 return NULL from CONF_get_section.
8275 *) Fix potential buffer overrun for EBCDIC.
8278 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8279 keyUsage if basicConstraints absent for a CA.
8282 *) Make SMIME_write_PKCS7() write mail header values with a format that
8283 is more generally accepted (no spaces before the semicolon), since
8284 some programs can't parse those values properly otherwise. Also make
8285 sure BIO's that break lines after each write do not create invalid
8289 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8290 macros previously used would not encode an empty SEQUENCE OF
8291 and break the signature.
8293 [This change does not apply to 0.9.7.]
8295 *) Zero the premaster secret after deriving the master secret in
8299 *) Add some EVP_add_digest_alias registrations (as found in
8300 OpenSSL_add_all_digests()) to SSL_library_init()
8301 aka OpenSSL_add_ssl_algorithms(). This provides improved
8302 compatibility with peers using X.509 certificates
8303 with unconventional AlgorithmIdentifier OIDs.
8306 *) Fix for Irix with NO_ASM.
8307 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8309 *) ./config script fixes.
8310 [Ulf Moeller, Richard Levitte]
8312 *) Fix 'openssl passwd -1'.
8315 *) Change PKCS12_key_gen_asc() so it can cope with non null
8316 terminated strings whose length is passed in the passlen
8317 parameter, for example from PEM callbacks. This was done
8318 by adding an extra length parameter to asc2uni().
8319 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8321 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8322 call failed, free the DSA structure.
8325 *) Fix to uni2asc() to cope with zero length Unicode strings.
8326 These are present in some PKCS#12 files.
8329 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8330 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8331 when writing a 32767 byte record.
8332 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8334 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8335 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8337 (RSA objects have a reference count access to which is protected
8338 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8339 so they are meant to be shared between threads.)
8340 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8341 "Reddie, Steven" <Steven.Reddie@ca.com>]
8343 *) Fix a deadlock in CRYPTO_mem_leaks().
8346 *) Use better test patterns in bntest.
8349 *) rand_win.c fix for Borland C.
8352 *) BN_rshift bugfix for n == 0.
8355 *) Add a 'bctest' script that checks for some known 'bc' bugs
8356 so that 'make test' does not abort just because 'bc' is broken.
8359 *) Store verify_result within SSL_SESSION also for client side to
8360 avoid potential security hole. (Re-used sessions on the client side
8361 always resulted in verify_result==X509_V_OK, not using the original
8362 result of the server certificate verification.)
8365 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8366 SSL3_RT_APPLICATION_DATA, return 0.
8367 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8371 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8372 releases, have been re-implemented by renaming the previous
8373 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8374 and ssl3_read_internal, respectively, and adding 'peek' parameters
8375 to them. The new ssl[23]_{read,peek} functions are calls to
8376 ssl[23]_read_internal with the 'peek' flag set appropriately.
8377 A 'peek' parameter has also been added to ssl3_read_bytes, which
8378 does the actual work for ssl3_read_internal.
8381 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8382 the method-specific "init()" handler. Also clean up ex_data after
8383 calling the method-specific "finish()" handler. Previously, this was
8384 happening the other way round.
8387 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8388 The previous value, 12, was not always sufficient for BN_mod_exp().
8391 *) Make sure that shared libraries get the internal name engine with
8392 the full version number and not just 0. This should mark the
8393 shared libraries as not backward compatible. Of course, this should
8394 be changed again when we can guarantee backward binary compatibility.
8397 *) Fix typo in get_cert_by_subject() in by_dir.c
8398 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8400 *) Rework the system to generate shared libraries:
8402 - Make note of the expected extension for the shared libraries and
8403 if there is a need for symbolic links from for example libcrypto.so.0
8404 to libcrypto.so.0.9.7. There is extended info in Configure for
8407 - Make as few rebuilds of the shared libraries as possible.
8409 - Still avoid linking the OpenSSL programs with the shared libraries.
8411 - When installing, install the shared libraries separately from the
8415 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8417 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8418 and not in SSL_clear because the latter is also used by the
8419 accept/connect functions; previously, the settings made by
8420 SSL_set_read_ahead would be lost during the handshake.
8421 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8423 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8424 Previously, it would create entries for disabled algorithms no
8428 *) Added several new manual pages for SSL_* function.
8431 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8433 *) In ssl23_get_client_hello, generate an error message when faced
8434 with an initial SSL 3.0/TLS record that is too small to contain the
8435 first two bytes of the ClientHello message, i.e. client_version.
8436 (Note that this is a pathologic case that probably has never happened
8437 in real life.) The previous approach was to use the version number
8438 from the record header as a substitute; but our protocol choice
8439 should not depend on that one because it is not authenticated
8440 by the Finished messages.
8443 *) More robust randomness gathering functions for Windows.
8444 [Jeffrey Altman <jaltman@columbia.edu>]
8446 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8447 not set then we don't setup the error code for issuer check errors
8448 to avoid possibly overwriting other errors which the callback does
8449 handle. If an application does set the flag then we assume it knows
8450 what it is doing and can handle the new informational codes
8454 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8455 a general "ANY" type, as such it should be able to decode anything
8456 including tagged types. However it didn't check the class so it would
8457 wrongly interpret tagged types in the same way as their universal
8458 counterpart and unknown types were just rejected. Changed so that the
8459 tagged and unknown types are handled in the same way as a SEQUENCE:
8460 that is the encoding is stored intact. There is also a new type
8461 "V_ASN1_OTHER" which is used when the class is not universal, in this
8462 case we have no idea what the actual type is so we just lump them all
8466 *) On VMS, stdout may very well lead to a file that is written to
8467 in a record-oriented fashion. That means that every write() will
8468 write a separate record, which will be read separately by the
8469 programs trying to read from it. This can be very confusing.
8471 The solution is to put a BIO filter in the way that will buffer
8472 text until a linefeed is reached, and then write everything a
8473 line at a time, so every record written will be an actual line,
8474 not chunks of lines and not (usually doesn't happen, but I've
8475 seen it once) several lines in one record. BIO_f_linebuffer() is
8478 Currently, it's a VMS-only method, because that's where it has
8479 been tested well enough.
8482 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8483 it can return incorrect results.
8484 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8485 but it was in 0.9.6-beta[12].)
8488 *) Disable the check for content being present when verifying detached
8489 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8490 include zero length content when signing messages.
8493 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8494 BIO_ctrl (for BIO pairs).
8497 *) Add DSO method for VMS.
8500 *) Bug fix: Montgomery multiplication could produce results with the
8504 *) Add RPM specification openssl.spec and modify it to build three
8505 packages. The default package contains applications, application
8506 documentation and run-time libraries. The devel package contains
8507 include files, static libraries and function documentation. The
8508 doc package contains the contents of the doc directory. The original
8509 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8512 *) Add a large number of documentation files for many SSL routines.
8513 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8515 *) Add a configuration entry for Sony News 4.
8516 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8518 *) Don't set the two most significant bits to one when generating a
8519 random number < q in the DSA library.
8522 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8523 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8524 the underlying transport is blocking) if a handshake took place.
8525 (The default behaviour is needed by applications such as s_client
8526 and s_server that use select() to determine when to use SSL_read;
8527 but for applications that know in advance when to expect data, it
8528 just makes things more complicated.)
8531 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8535 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8536 work better on such systems.
8537 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8539 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8540 Update PKCS12_parse() so it copies the friendlyName and the
8541 keyid to the certificates aux info.
8544 *) Fix bug in PKCS7_verify() which caused an infinite loop
8545 if there was more than one signature.
8546 [Sven Uszpelkat <su@celocom.de>]
8548 *) Major change in util/mkdef.pl to include extra information
8549 about each symbol, as well as presenting variables as well
8550 as functions. This change means that there's n more need
8551 to rebuild the .num files when some algorithms are excluded.
8554 *) Allow the verify time to be set by an application,
8555 rather than always using the current time.
8558 *) Phase 2 verify code reorganisation. The certificate
8559 verify code now looks up an issuer certificate by a
8560 number of criteria: subject name, authority key id
8561 and key usage. It also verifies self signed certificates
8562 by the same criteria. The main comparison function is
8563 X509_check_issued() which performs these checks.
8565 Lot of changes were necessary in order to support this
8566 without completely rewriting the lookup code.
8568 Authority and subject key identifier are now cached.
8570 The LHASH 'certs' is X509_STORE has now been replaced
8571 by a STACK_OF(X509_OBJECT). This is mainly because an
8572 LHASH can't store or retrieve multiple objects with
8573 the same hash value.
8575 As a result various functions (which were all internal
8576 use only) have changed to handle the new X509_STORE
8577 structure. This will break anything that messed round
8578 with X509_STORE internally.
8580 The functions X509_STORE_add_cert() now checks for an
8581 exact match, rather than just subject name.
8583 The X509_STORE API doesn't directly support the retrieval
8584 of multiple certificates matching a given criteria, however
8585 this can be worked round by performing a lookup first
8586 (which will fill the cache with candidate certificates)
8587 and then examining the cache for matches. This is probably
8588 the best we can do without throwing out X509_LOOKUP
8589 entirely (maybe later...).
8591 The X509_VERIFY_CTX structure has been enhanced considerably.
8593 All certificate lookup operations now go via a get_issuer()
8594 callback. Although this currently uses an X509_STORE it
8595 can be replaced by custom lookups. This is a simple way
8596 to bypass the X509_STORE hackery necessary to make this
8597 work and makes it possible to use more efficient techniques
8598 in future. A very simple version which uses a simple
8599 STACK for its trusted certificate store is also provided
8600 using X509_STORE_CTX_trusted_stack().
8602 The verify_cb() and verify() callbacks now have equivalents
8603 in the X509_STORE_CTX structure.
8605 X509_STORE_CTX also has a 'flags' field which can be used
8606 to customise the verify behaviour.
8609 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8610 excludes S/MIME capabilities.
8613 *) When a certificate request is read in keep a copy of the
8614 original encoding of the signed data and use it when outputing
8615 again. Signatures then use the original encoding rather than
8616 a decoded, encoded version which may cause problems if the
8617 request is improperly encoded.
8620 *) For consistency with other BIO_puts implementations, call
8621 buffer_write(b, ...) directly in buffer_puts instead of calling
8624 In BIO_puts, increment b->num_write as in BIO_write.
8625 [Peter.Sylvester@EdelWeb.fr]
8627 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8628 BN_zero, we may not return a BIGNUM with an array consisting of
8632 *) Avoid calling abort() from within the library when problems are
8633 detected, except if preprocessor symbols have been defined
8634 (such as REF_CHECK, BN_DEBUG etc.).
8637 *) New openssl application 'rsautl'. This utility can be
8638 used for low level RSA operations. DER public key
8639 BIO/fp routines also added.
8642 *) New Configure entry and patches for compiling on QNX 4.
8643 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8645 *) A demo state-machine implementation was sponsored by
8646 Nuron (http://www.nuron.com/) and is now available in
8647 demos/state_machine.
8650 *) New options added to the 'dgst' utility for signature
8651 generation and verification.
8654 *) Unrecognized PKCS#7 content types are now handled via a
8655 catch all ASN1_TYPE structure. This allows unsupported
8656 types to be stored as a "blob" and an application can
8657 encode and decode it manually.
8660 *) Fix various signed/unsigned issues to make a_strex.c
8662 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8664 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8665 length if passed a buffer. ASN1_INTEGER_to_BN failed
8666 if passed a NULL BN and its argument was negative.
8667 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8669 *) Modification to PKCS#7 encoding routines to output definite
8670 length encoding. Since currently the whole structures are in
8671 memory there's not real point in using indefinite length
8672 constructed encoding. However if OpenSSL is compiled with
8673 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8676 *) Added BIO_vprintf() and BIO_vsnprintf().
8679 *) Added more prefixes to parse for in the the strings written
8680 through a logging bio, to cover all the levels that are available
8681 through syslog. The prefixes are now:
8683 PANIC, EMERG, EMR => LOG_EMERG
8684 ALERT, ALR => LOG_ALERT
8685 CRIT, CRI => LOG_CRIT
8686 ERROR, ERR => LOG_ERR
8687 WARNING, WARN, WAR => LOG_WARNING
8688 NOTICE, NOTE, NOT => LOG_NOTICE
8689 INFO, INF => LOG_INFO
8690 DEBUG, DBG => LOG_DEBUG
8692 and as before, if none of those prefixes are present at the
8693 beginning of the string, LOG_ERR is chosen.
8695 On Win32, the LOG_* levels are mapped according to this:
8697 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8698 LOG_WARNING => EVENTLOG_WARNING_TYPE
8699 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8703 *) Made it possible to reconfigure with just the configuration
8704 argument "reconf" or "reconfigure". The command line arguments
8705 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8706 and are retrieved from there when reconfiguring.
8710 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8712 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8715 *) The obj_dat.pl script was messing up the sorting of object
8716 names. The reason was that it compared the quoted version
8717 of strings as a result "OCSP" > "OCSP Signing" because
8718 " > SPACE. Changed script to store unquoted versions of
8719 names and add quotes on output. It was also omitting some
8720 names from the lookup table if they were given a default
8721 value (that is if SN is missing it is given the same
8722 value as LN and vice versa), these are now added on the
8723 grounds that if an object has a name we should be able to
8724 look it up. Finally added warning output when duplicate
8725 short or long names are found.
8728 *) Changes needed for Tandem NSK.
8729 [Scott Uroff <scott@xypro.com>]
8731 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8732 RSA_padding_check_SSLv23(), special padding was never detected
8733 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8734 version rollback attacks was not effective.
8736 In s23_clnt.c, don't use special rollback-attack detection padding
8737 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8738 client; similarly, in s23_srvr.c, don't do the rollback check if
8739 SSL 2.0 is the only protocol enabled in the server.
8742 *) Make it possible to get hexdumps of unprintable data with 'openssl
8743 asn1parse'. By implication, the functions ASN1_parse_dump() and
8744 BIO_dump_indent() are added.
8747 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8748 these print out strings and name structures based on various
8749 flags including RFC2253 support and proper handling of
8750 multibyte characters. Added options to the 'x509' utility
8751 to allow the various flags to be set.
8754 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8755 Also change the functions X509_cmp_current_time() and
8756 X509_gmtime_adj() work with an ASN1_TIME structure,
8757 this will enable certificates using GeneralizedTime in validity
8758 dates to be checked.
8761 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8762 negative public key encodings) on by default,
8763 NO_NEG_PUBKEY_BUG can be set to disable it.
8766 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8767 content octets. An i2c_ASN1_OBJECT is unnecessary because
8768 the encoding can be trivially obtained from the structure.
8771 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8772 not read locks (CRYPTO_r_[un]lock).
8775 *) A first attempt at creating official support for shared
8776 libraries through configuration. I've kept it so the
8777 default is static libraries only, and the OpenSSL programs
8778 are always statically linked for now, but there are
8779 preparations for dynamic linking in place.
8780 This has been tested on Linux and Tru64.
8783 *) Randomness polling function for Win9x, as described in:
8784 Peter Gutmann, Software Generation of Practically Strong
8788 *) Fix so PRNG is seeded in req if using an already existing
8792 *) New options to smime application. -inform and -outform
8793 allow alternative formats for the S/MIME message including
8794 PEM and DER. The -content option allows the content to be
8795 specified separately. This should allow things like Netscape
8796 form signing output easier to verify.
8799 *) Fix the ASN1 encoding of tags using the 'long form'.
8802 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8803 STRING types. These convert content octets to and from the
8804 underlying type. The actual tag and length octets are
8805 already assumed to have been read in and checked. These
8806 are needed because all other string types have virtually
8807 identical handling apart from the tag. By having versions
8808 of the ASN1 functions that just operate on content octets
8809 IMPLICIT tagging can be handled properly. It also allows
8810 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8811 and ASN1_INTEGER are identical apart from the tag.
8814 *) Change the handling of OID objects as follows:
8816 - New object identifiers are inserted in objects.txt, following
8817 the syntax given in objects.README.
8818 - objects.pl is used to process obj_mac.num and create a new
8820 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8823 This is currently kind of a hack, and the perl code in objects.pl
8824 isn't very elegant, but it works as I intended. The simplest way
8825 to check that it worked correctly is to look in obj_dat.h and
8826 check the array nid_objs and make sure the objects haven't moved
8827 around (this is important!). Additions are OK, as well as
8828 consistent name changes.
8831 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8834 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8835 The given file adds to whatever has already been seeded into the
8836 random pool through the RANDFILE configuration file option or
8837 environment variable, or the default random state file.
8840 *) mkstack.pl now sorts each macro group into lexical order.
8841 Previously the output order depended on the order the files
8842 appeared in the directory, resulting in needless rewriting
8846 *) Patches to make OpenSSL compile under Win32 again. Mostly
8847 work arounds for the VC++ problem that it treats func() as
8848 func(void). Also stripped out the parts of mkdef.pl that
8849 added extra typesafe functions: these no longer exist.
8852 *) Reorganisation of the stack code. The macros are now all
8853 collected in safestack.h . Each macro is defined in terms of
8854 a "stack macro" of the form SKM_<name>(type, a, b). The
8855 DEBUG_SAFESTACK is now handled in terms of function casts,
8856 this has the advantage of retaining type safety without the
8857 use of additional functions. If DEBUG_SAFESTACK is not defined
8858 then the non typesafe macros are used instead. Also modified the
8859 mkstack.pl script to handle the new form. Needs testing to see
8860 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8861 the default if no major problems. Similar behaviour for ASN1_SET_OF
8862 and PKCS12_STACK_OF.
8865 *) When some versions of IIS use the 'NET' form of private key the
8866 key derivation algorithm is different. Normally MD5(password) is
8867 used as a 128 bit RC4 key. In the modified case
8868 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8869 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8870 as the old Netscape_RSA functions except they have an additional
8871 'sgckey' parameter which uses the modified algorithm. Also added
8872 an -sgckey command line option to the rsa utility. Thanks to
8873 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8874 algorithm to openssl-dev.
8877 *) The evp_local.h macros were using 'c.##kname' which resulted in
8878 invalid expansion on some systems (SCO 5.0.5 for example).
8879 Corrected to 'c.kname'.
8880 [Phillip Porch <root@theporch.com>]
8882 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8883 a STACK of email addresses from a certificate or request, these look
8884 in the subject name and the subject alternative name extensions and
8885 omit any duplicate addresses.
8888 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8889 This makes DSA verification about 2 % faster.
8892 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8893 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8894 plus overhead for 1024 bit moduli).
8895 This makes exponentiations about 0.5 % faster for 1024 bit
8896 exponents (as measured by "openssl speed rsa2048").
8899 *) Rename memory handling macros to avoid conflicts with other
8901 Malloc => OPENSSL_malloc
8902 Malloc_locked => OPENSSL_malloc_locked
8903 Realloc => OPENSSL_realloc
8904 Free => OPENSSL_free
8907 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8908 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8911 *) CygWin32 support.
8912 [John Jarvie <jjarvie@newsguy.com>]
8914 *) The type-safe stack code has been rejigged. It is now only compiled
8915 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8916 by default all type-specific stack functions are "#define"d back to
8917 standard stack functions. This results in more streamlined output
8918 but retains the type-safety checking possibilities of the original
8922 *) The STACK code has been cleaned up, and certain type declarations
8923 that didn't make a lot of sense have been brought in line. This has
8924 also involved a cleanup of sorts in safestack.h to more correctly
8925 map type-safe stack functions onto their plain stack counterparts.
8926 This work has also resulted in a variety of "const"ifications of
8927 lots of the code, especially "_cmp" operations which should normally
8928 be prototyped with "const" parameters anyway.
8931 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8932 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8933 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8934 where all of 'md' is used each time the PRNG is used, but 'state'
8935 is used only indexed by a cyclic counter. As entropy may not be
8936 well distributed from the beginning, 'md' is important as a
8937 chaining variable. However, the output function chains only half
8938 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8939 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8940 in all of 'state' being rewritten, with the new values depending
8941 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8944 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8945 the handshake is continued after ssl_verify_cert_chain();
8946 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8947 can lead to 'unexplainable' connection aborts later.
8948 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8950 *) Major EVP API cipher revision.
8951 Add hooks for extra EVP features. This allows various cipher
8952 parameters to be set in the EVP interface. Support added for variable
8953 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8954 setting of RC2 and RC5 parameters.
8956 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8959 Remove lots of duplicated code from the EVP library. For example *every*
8960 cipher init() function handles the 'iv' in the same way according to the
8961 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8962 for CFB and OFB modes they zero ctx->num.
8964 New functionality allows removal of S/MIME code RC2 hack.
8966 Most of the routines have the same form and so can be declared in terms
8969 By shifting this to the top level EVP_CipherInit() it can be removed from
8970 all individual ciphers. If the cipher wants to handle IVs or keys
8971 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8974 Change lots of functions like EVP_EncryptUpdate() to now return a
8975 value: although software versions of the algorithms cannot fail
8976 any installed hardware versions can.
8979 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8980 this option is set, tolerate broken clients that send the negotiated
8981 protocol version number instead of the requested protocol version
8985 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8986 i.e. non-zero for export ciphersuites, zero otherwise.
8987 Previous versions had this flag inverted, inconsistent with
8988 rsa_tmp_cb (..._TMP_RSA_CB).
8989 [Bodo Moeller; problem reported by Amit Chopra]
8991 *) Add missing DSA library text string. Work around for some IIS
8992 key files with invalid SEQUENCE encoding.
8995 *) Add a document (doc/standards.txt) that list all kinds of standards
8996 and so on that are implemented in OpenSSL.
8999 *) Enhance c_rehash script. Old version would mishandle certificates
9000 with the same subject name hash and wouldn't handle CRLs at all.
9001 Added -fingerprint option to crl utility, to support new c_rehash
9005 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9008 *) Fix for SSL server purpose checking. Server checking was
9009 rejecting certificates which had extended key usage present
9010 but no ssl client purpose.
9011 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9013 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9014 is a little unclear about how a blank password is handled.
9015 Since the password in encoded as a BMPString with terminating
9016 double NULL a zero length password would end up as just the
9017 double NULL. However no password at all is different and is
9018 handled differently in the PKCS#12 key generation code. NS
9019 treats a blank password as zero length. MSIE treats it as no
9020 password on export: but it will try both on import. We now do
9021 the same: PKCS12_parse() tries zero length and no password if
9022 the password is set to "" or NULL (NULL is now a valid password:
9023 it wasn't before) as does the pkcs12 application.
9026 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9027 perror when PEM_read_bio_X509_REQ fails, the error message must
9028 be obtained from the error queue.
9031 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9032 it in ERR_remove_state if appropriate, and change ERR_get_state
9033 accordingly to avoid race conditions (this is necessary because
9034 thread_hash is no longer constant once set).
9037 *) Bugfix for linux-elf makefile.one.
9040 *) RSA_get_default_method() will now cause a default
9041 RSA_METHOD to be chosen if one doesn't exist already.
9042 Previously this was only set during a call to RSA_new()
9043 or RSA_new_method(NULL) meaning it was possible for
9044 RSA_get_default_method() to return NULL.
9047 *) Added native name translation to the existing DSO code
9048 that will convert (if the flag to do so is set) filenames
9049 that are sufficiently small and have no path information
9050 into a canonical native form. Eg. "blah" converted to
9051 "libblah.so" or "blah.dll" etc.
9054 *) New function ERR_error_string_n(e, buf, len) which is like
9055 ERR_error_string(e, buf), but writes at most 'len' bytes
9056 including the 0 terminator. For ERR_error_string_n, 'buf'
9058 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9060 *) CONF library reworked to become more general. A new CONF
9061 configuration file reader "class" is implemented as well as a
9062 new functions (NCONF_*, for "New CONF") to handle it. The now
9063 old CONF_* functions are still there, but are reimplemented to
9064 work in terms of the new functions. Also, a set of functions
9065 to handle the internal storage of the configuration data is
9066 provided to make it easier to write new configuration file
9067 reader "classes" (I can definitely see something reading a
9068 configuration file in XML format, for example), called _CONF_*,
9069 or "the configuration storage API"...
9071 The new configuration file reading functions are:
9073 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9074 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9076 NCONF_default, NCONF_WIN32
9078 NCONF_dump_fp, NCONF_dump_bio
9080 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9081 NCONF_new creates a new CONF object. This works in the same way
9082 as other interfaces in OpenSSL, like the BIO interface.
9083 NCONF_dump_* dump the internal storage of the configuration file,
9084 which is useful for debugging. All other functions take the same
9085 arguments as the old CONF_* functions wth the exception of the
9086 first that must be a `CONF *' instead of a `LHASH *'.
9088 To make it easer to use the new classes with the old CONF_* functions,
9089 the function CONF_set_default_method is provided.
9092 *) Add '-tls1' option to 'openssl ciphers', which was already
9093 mentioned in the documentation but had not been implemented.
9094 (This option is not yet really useful because even the additional
9095 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9098 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9099 OpenSSL-based applications) load shared libraries and bind to
9100 them in a portable way.
9101 [Geoff Thorpe, with contributions from Richard Levitte]
9103 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9105 *) Make sure _lrotl and _lrotr are only used with MSVC.
9107 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9108 (the default implementation of RAND_status).
9110 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9111 to '-clrext' (= clear extensions), as intended and documented.
9112 [Bodo Moeller; inconsistency pointed out by Michael Attili
9115 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9116 was larger than the MD block size.
9117 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9119 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9120 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9121 using the passed key: if the passed key was a private key the result
9122 of X509_print(), for example, would be to print out all the private key
9126 *) des_quad_cksum() byte order bug fix.
9127 [Ulf Möller, using the problem description in krb4-0.9.7, where
9128 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9130 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9132 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9134 *) For easily testing in shell scripts whether some command
9135 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9136 returns with exit code 0 iff no command of the given name is available.
9137 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9138 the output goes to stdout and nothing is printed to stderr.
9139 Additional arguments are always ignored.
9141 Since for each cipher there is a command of the same name,
9142 the 'no-cipher' compilation switches can be tested this way.
9144 ('openssl no-XXX' is not able to detect pseudo-commands such
9145 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9148 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9151 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9152 is set; it will be thrown away anyway because each handshake creates
9154 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9155 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9156 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9157 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9160 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9161 'Q' and 'R' lose their special meanings (quit/renegotiate).
9162 This is part of what -quiet does; unlike -quiet, -ign_eof
9163 does not suppress any output.
9166 *) Add compatibility options to the purpose and trust code. The
9167 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9168 accepts a certificate or CA, this was the previous behaviour,
9169 with all the associated security issues.
9171 X509_TRUST_COMPAT is the old trust behaviour: only and
9172 automatically trust self signed roots in certificate store. A
9173 new trust setting X509_TRUST_DEFAULT is used to specify that
9174 a purpose has no associated trust setting and it should instead
9175 use the value in the default purpose.
9178 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9179 and fix a memory leak.
9182 *) In util/mkerr.pl (which implements 'make errors'), preserve
9183 reason strings from the previous version of the .c file, as
9184 the default to have only downcase letters (and digits) in
9185 automatically generated reasons codes is not always appropriate.
9188 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9189 using strerror. Previously, ERR_reason_error_string() returned
9190 library names as reason strings for SYSerr; but SYSerr is a special
9191 case where small numbers are errno values, not library numbers.
9194 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9195 converts DSA parameters into DH parameters. (When creating parameters,
9196 DSA_generate_parameters is used.)
9199 *) Include 'length' (recommended exponent length) in C code generated
9200 by 'openssl dhparam -C'.
9203 *) The second argument to set_label in perlasm was already being used
9204 so couldn't be used as a "file scope" flag. Moved to third argument
9208 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9209 instead of RAND_bytes for encryption IVs and salts.
9212 *) Include RAND_status() into RAND_METHOD instead of implementing
9213 it only for md_rand.c Otherwise replacing the PRNG by calling
9214 RAND_set_rand_method would be impossible.
9217 *) Don't let DSA_generate_key() enter an infinite loop if the random
9218 number generation fails.
9221 *) New 'rand' application for creating pseudo-random output.
9224 *) Added configuration support for Linux/IA64
9225 [Rolf Haberrecker <rolf@suse.de>]
9227 *) Assembler module support for Mingw32.
9230 *) Shared library support for HPUX (in shlib/).
9231 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9233 *) Shared library support for Solaris gcc.
9234 [Lutz Behnke <behnke@trustcenter.de>]
9236 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9238 *) PKCS7_encrypt() was adding text MIME headers twice because they
9239 were added manually and by SMIME_crlf_copy().
9242 *) In bntest.c don't call BN_rand with zero bits argument.
9243 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9245 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9246 case was implemented. This caused BN_div_recp() to fail occasionally.
9249 *) Add an optional second argument to the set_label() in the perl
9250 assembly language builder. If this argument exists and is set
9251 to 1 it signals that the assembler should use a symbol whose
9252 scope is the entire file, not just the current function. This
9253 is needed with MASM which uses the format label:: for this scope.
9254 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9256 *) Change the ASN1 types so they are typedefs by default. Before
9257 almost all types were #define'd to ASN1_STRING which was causing
9258 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9262 *) Change names of new functions to the new get1/get0 naming
9263 convention: After 'get1', the caller owns a reference count
9264 and has to call ..._free; 'get0' returns a pointer to some
9265 data structure without incrementing reference counters.
9266 (Some of the existing 'get' functions increment a reference
9267 counter, some don't.)
9268 Similarly, 'set1' and 'add1' functions increase reference
9269 counters or duplicate objects.
9272 *) Allow for the possibility of temp RSA key generation failure:
9273 the code used to assume it always worked and crashed on failure.
9276 *) Fix potential buffer overrun problem in BIO_printf().
9277 [Ulf Möller, using public domain code by Patrick Powell; problem
9278 pointed out by David Sacerdote <das33@cornell.edu>]
9280 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9281 RAND_egd() and RAND_status(). In the command line application,
9282 the EGD socket can be specified like a seed file using RANDFILE
9286 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9287 Some CAs (e.g. Verisign) distribute certificates in this form.
9290 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9291 list to exclude them. This means that no special compilation option
9292 is needed to use anonymous DH: it just needs to be included in the
9296 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9297 EVP_MD_type. The old functionality is available in a new macro called
9298 EVP_MD_md(). Change code that uses it and update docs.
9301 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9302 where the 'void *' argument is replaced by a function pointer argument.
9303 Previously 'void *' was abused to point to functions, which works on
9304 many platforms, but is not correct. As these functions are usually
9305 called by macros defined in OpenSSL header files, most source code
9306 should work without changes.
9309 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9310 sections with information on -D... compiler switches used for
9311 compiling the library so that applications can see them. To enable
9312 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9313 must be defined. E.g.,
9314 #define OPENSSL_ALGORITHM_DEFINES
9315 #include <openssl/opensslconf.h>
9316 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9317 [Richard Levitte, Ulf and Bodo Möller]
9319 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9323 *) Change the 'other' type in certificate aux info to a STACK_OF
9324 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9325 the required ASN1 format: arbitrary types determined by an OID.
9328 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9329 argument to 'req'. This is not because the function is newer or
9330 better than others it just uses the work 'NEW' in the certificate
9331 request header lines. Some software needs this.
9334 *) Reorganise password command line arguments: now passwords can be
9335 obtained from various sources. Delete the PEM_cb function and make
9336 it the default behaviour: i.e. if the callback is NULL and the
9337 usrdata argument is not NULL interpret it as a null terminated pass
9338 phrase. If usrdata and the callback are NULL then the pass phrase
9339 is prompted for as usual.
9342 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9343 the support is automatically enabled. The resulting binaries will
9344 autodetect the card and use it if present.
9345 [Ben Laurie and Compaq Inc.]
9347 *) Work around for Netscape hang bug. This sends certificate request
9348 and server done in one record. Since this is perfectly legal in the
9349 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9350 the bugs/SSLv3 entry for more info.
9353 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9356 *) Add -rand argument to smime and pkcs12 applications and read/write
9360 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9363 *) Add command line password options to the remaining applications.
9366 *) Bug fix for BN_div_recp() for numerators with an even number of
9370 *) More tests in bntest.c, and changed test_bn output.
9373 *) ./config recognizes MacOS X now.
9376 *) Bug fix for BN_div() when the first words of num and divsor are
9377 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9380 *) Add support for various broken PKCS#8 formats, and command line
9381 options to produce them.
9384 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9385 get temporary BIGNUMs from a BN_CTX.
9388 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9392 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9393 include a #define from the old name to the new. The original intent
9394 was that statically linked binaries could for example just call
9395 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9396 link with digests. This never worked because SSLeay_add_all_digests()
9397 and SSLeay_add_all_ciphers() were in the same source file so calling
9398 one would link with the other. They are now in separate source files.
9401 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9404 *) Use a less unusual form of the Miller-Rabin primality test (it used
9405 a binary algorithm for exponentiation integrated into the Miller-Rabin
9406 loop, our standard modexp algorithms are faster).
9409 *) Support for the EBCDIC character set completed.
9410 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9412 *) Source code cleanups: use const where appropriate, eliminate casts,
9413 use void * instead of char * in lhash.
9416 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9417 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9418 this the server could overwrite ephemeral keys that the client
9422 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9423 using 50 iterations of the Rabin-Miller test.
9425 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9426 iterations of the Rabin-Miller test as required by the appendix
9427 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9428 As BN_is_prime_fasttest includes trial division, DSA parameter
9429 generation becomes much faster.
9431 This implies a change for the callback functions in DSA_is_prime
9432 and DSA_generate_parameters: The callback function is called once
9433 for each positive witness in the Rabin-Miller test, not just
9434 occasionally in the inner loop; and the parameters to the
9435 callback function now provide an iteration count for the outer
9436 loop rather than for the current invocation of the inner loop.
9437 DSA_generate_parameters additionally can call the callback
9438 function with an 'iteration count' of -1, meaning that a
9439 candidate has passed the trial division test (when q is generated
9440 from an application-provided seed, trial division is skipped).
9443 *) New function BN_is_prime_fasttest that optionally does trial
9444 division before starting the Rabin-Miller test and has
9445 an additional BN_CTX * argument (whereas BN_is_prime always
9446 has to allocate at least one BN_CTX).
9447 'callback(1, -1, cb_arg)' is called when a number has passed the
9448 trial division stage.
9451 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9455 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9458 *) New function BN_pseudo_rand().
9461 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9462 bignum version of BN_from_montgomery() with the working code from
9463 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9467 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9468 made it impossible to use the same SSL_SESSION data structure in
9469 SSL2 clients in multiple threads.
9472 *) The return value of RAND_load_file() no longer counts bytes obtained
9473 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9474 to seed the PRNG (previously an explicit byte count was required).
9475 [Ulf Möller, Bodo Möller]
9477 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9478 used (char *) instead of (void *) and had casts all over the place.
9481 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9484 *) Retain source code compatibility for BN_prime_checks macro:
9485 BN_is_prime(..., BN_prime_checks, ...) now uses
9486 BN_prime_checks_for_size to determine the appropriate number of
9487 Rabin-Miller iterations.
9490 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9491 DH_CHECK_P_NOT_SAFE_PRIME.
9492 (Check if this is true? OpenPGP calls them "strong".)
9495 *) Merge the functionality of "dh" and "gendh" programs into a new program
9496 "dhparam". The old programs are retained for now but will handle DH keys
9497 (instead of parameters) in future.
9500 *) Make the ciphers, s_server and s_client programs check the return values
9501 when a new cipher list is set.
9504 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9505 ciphers. Before when the 56bit ciphers were enabled the sorting was
9508 The syntax for the cipher sorting has been extended to support sorting by
9509 cipher-strength (using the strength_bits hard coded in the tables).
9510 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9512 Fix a bug in the cipher-command parser: when supplying a cipher command
9513 string with an "undefined" symbol (neither command nor alphanumeric
9514 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9515 an error is flagged.
9517 Due to the strength-sorting extension, the code of the
9518 ssl_create_cipher_list() function was completely rearranged. I hope that
9519 the readability was also increased :-)
9520 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9522 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9523 for the first serial number and places 2 in the serial number file. This
9524 avoids problems when the root CA is created with serial number zero and
9525 the first user certificate has the same issuer name and serial number
9529 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9530 the new code. Add documentation for this stuff.
9533 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9534 X509_*() to X509at_*() on the grounds that they don't handle X509
9535 structures and behave in an analogous way to the X509v3 functions:
9536 they shouldn't be called directly but wrapper functions should be used
9539 So we also now have some wrapper functions that call the X509at functions
9540 when passed certificate requests. (TO DO: similar things can be done with
9541 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9542 things. Some of these need some d2i or i2d and print functionality
9543 because they handle more complex structures.)
9546 *) Add missing #ifndefs that caused missing symbols when building libssl
9547 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9548 NO_RSA in ssl/s2*.c.
9549 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9551 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9552 has a return value which indicates the quality of the random data
9553 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9554 error queue. New function RAND_pseudo_bytes() generates output that is
9555 guaranteed to be unique but not unpredictable. RAND_add is like
9556 RAND_seed, but takes an extra argument for an entropy estimate
9557 (RAND_seed always assumes full entropy).
9560 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9561 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9562 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9563 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9564 false-positive rate of at most 2^-80 for random input.
9567 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9570 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9571 in the 0.9.5 release), this returns the chain
9572 from an X509_CTX structure with a dup of the stack and all
9573 the X509 reference counts upped: so the stack will exist
9574 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9577 Also make SSL_SESSION_print() print out the verify return
9581 *) Add manpage for the pkcs12 command. Also change the default
9582 behaviour so MAC iteration counts are used unless the new
9583 -nomaciter option is used. This improves file security and
9584 only older versions of MSIE (4.0 for example) need it.
9587 *) Honor the no-xxx Configure options when creating .DEF files.
9590 *) Add PKCS#10 attributes to field table: challengePassword,
9591 unstructuredName and unstructuredAddress. These are taken from
9592 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9593 international characters are used.
9595 More changes to X509_ATTRIBUTE code: allow the setting of types
9596 based on strings. Remove the 'loc' parameter when adding
9597 attributes because these will be a SET OF encoding which is sorted
9601 *) Initial changes to the 'req' utility to allow request generation
9602 automation. This will allow an application to just generate a template
9603 file containing all the field values and have req construct the
9606 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9607 used all over the place including certificate requests and PKCS#7
9608 structures. They are currently handled manually where necessary with
9609 some primitive wrappers for PKCS#7. The new functions behave in a
9610 manner analogous to the X509 extension functions: they allow
9611 attributes to be looked up by NID and added.
9613 Later something similar to the X509V3 code would be desirable to
9614 automatically handle the encoding, decoding and printing of the
9615 more complex types. The string types like challengePassword can
9616 be handled by the string table functions.
9618 Also modified the multi byte string table handling. Now there is
9619 a 'global mask' which masks out certain types. The table itself
9620 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9621 is useful when for example there is only one permissible type
9622 (as in countryName) and using the mask might result in no valid
9626 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9627 SSL_get_peer_finished to allow applications to obtain the latest
9628 Finished messages sent to the peer or expected from the peer,
9629 respectively. (SSL_get_peer_finished is usually the Finished message
9630 actually received from the peer, otherwise the protocol will be aborted.)
9632 As the Finished message are message digests of the complete handshake
9633 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9634 be used for external authentication procedures when the authentication
9635 provided by SSL/TLS is not desired or is not enough.
9638 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9639 the host supports BWX extension and if Compaq C is present on the
9640 $PATH. Just exploiting of the BWX extension results in 20-30%
9641 performance kick for some algorithms, e.g. DES and RC4 to mention
9642 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9646 *) Add support for MS "fast SGC". This is arguably a violation of the
9647 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9648 weak crypto and after checking the certificate is SGC a second one
9649 with strong crypto. MS SGC stops the first handshake after receiving
9650 the server certificate message and sends a second client hello. Since
9651 a server will typically do all the time consuming operations before
9652 expecting any further messages from the client (server key exchange
9653 is the most expensive) there is little difference between the two.
9655 To get OpenSSL to support MS SGC we have to permit a second client
9656 hello message after we have sent server done. In addition we have to
9657 reset the MAC if we do get this second client hello.
9660 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9661 if a DER encoded private key is RSA or DSA traditional format. Changed
9662 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9663 format DER encoded private key. Newer code should use PKCS#8 format which
9664 has the key type encoded in the ASN1 structure. Added DER private key
9665 support to pkcs8 application.
9668 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9669 ciphersuites has been selected (as required by the SSL 3/TLS 1
9670 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9671 is set, we interpret this as a request to violate the specification
9672 (the worst that can happen is a handshake failure, and 'correct'
9673 behaviour would result in a handshake failure anyway).
9676 *) In SSL_CTX_add_session, take into account that there might be multiple
9677 SSL_SESSION structures with the same session ID (e.g. when two threads
9678 concurrently obtain them from an external cache).
9679 The internal cache can handle only one SSL_SESSION with a given ID,
9680 so if there's a conflict, we now throw out the old one to achieve
9684 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9685 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9686 some routines that use cipher OIDs: some ciphers do not have OIDs
9687 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9691 *) Simplify the trust setting structure and code. Now we just have
9692 two sequences of OIDs for trusted and rejected settings. These will
9693 typically have values the same as the extended key usage extension
9694 and any application specific purposes.
9696 The trust checking code now has a default behaviour: it will just
9697 check for an object with the same NID as the passed id. Functions can
9698 be provided to override either the default behaviour or the behaviour
9699 for a given id. SSL client, server and email already have functions
9700 in place for compatibility: they check the NID and also return "trusted"
9701 if the certificate is self signed.
9704 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9705 traditional format into an EVP_PKEY structure.
9708 *) Add a password callback function PEM_cb() which either prompts for
9709 a password if usr_data is NULL or otherwise assumes it is a null
9710 terminated password. Allow passwords to be passed on command line
9711 environment or config files in a few more utilities.
9714 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9715 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9716 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9717 Update documentation.
9720 *) Support for ASN1 "NULL" type. This could be handled before by using
9721 ASN1_TYPE but there wasn't any function that would try to read a NULL
9722 and produce an error if it couldn't. For compatibility we also have
9723 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9724 don't allocate anything because they don't need to.
9727 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9729 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9731 *) Rebuild of the memory allocation routines used by OpenSSL code and
9732 possibly others as well. The purpose is to make an interface that
9733 provide hooks so anyone can build a separate set of allocation and
9734 deallocation routines to be used by OpenSSL, for example memory
9735 pool implementations, or something else, which was previously hard
9736 since Malloc(), Realloc() and Free() were defined as macros having
9737 the values malloc, realloc and free, respectively (except for Win32
9738 compilations). The same is provided for memory debugging code.
9739 OpenSSL already comes with functionality to find memory leaks, but
9740 this gives people a chance to debug other memory problems.
9742 With these changes, a new set of functions and macros have appeared:
9744 CRYPTO_set_mem_debug_functions() [F]
9745 CRYPTO_get_mem_debug_functions() [F]
9746 CRYPTO_dbg_set_options() [F]
9747 CRYPTO_dbg_get_options() [F]
9748 CRYPTO_malloc_debug_init() [M]
9750 The memory debug functions are NULL by default, unless the library
9751 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9752 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9753 gives the standard debugging functions that come with OpenSSL) or
9754 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9755 provided by the library user) must be used. When the standard
9756 debugging functions are used, CRYPTO_dbg_set_options can be used to
9757 request additional information:
9758 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9759 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9761 Also, things like CRYPTO_set_mem_functions will always give the
9762 expected result (the new set of functions is used for allocation
9763 and deallocation) at all times, regardless of platform and compiler
9766 To finish it up, some functions that were never use in any other
9767 way than through macros have a new API and new semantic:
9770 CRYPTO_dbg_realloc()
9773 All macros of value have retained their old syntax.
9774 [Richard Levitte and Bodo Moeller]
9776 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9777 ordering of SMIMECapabilities wasn't in "strength order" and there
9778 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9782 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9783 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9784 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9786 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9787 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9788 functionality to handle multipart/signed properly) and a utility
9789 called 'smime' to call all this stuff. This is based on code I
9790 originally wrote for Celo who have kindly allowed it to be
9791 included in OpenSSL.
9794 *) Add variants des_set_key_checked and des_set_key_unchecked of
9795 des_set_key (aka des_key_sched). Global variable des_check_key
9796 decides which of these is called by des_set_key; this way
9797 des_check_key behaves as it always did, but applications and
9798 the library itself, which was buggy for des_check_key == 1,
9799 have a cleaner way to pick the version they need.
9802 *) New function PKCS12_newpass() which changes the password of a
9806 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9807 dynamic mix. In both cases the ids can be used as an index into the
9808 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9809 functions so they accept a list of the field values and the
9810 application doesn't need to directly manipulate the X509_TRUST
9814 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9818 *) Modify the way the V3 extension code looks up extensions. This now
9819 works in a similar way to the object code: we have some "standard"
9820 extensions in a static table which is searched with OBJ_bsearch()
9821 and the application can add dynamic ones if needed. The file
9822 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9823 updated whenever a new extension is added to the core code and kept
9824 in ext_nid order. There is a simple program 'tabtest.c' which checks
9825 this. New extensions are not added too often so this file can readily
9826 be maintained manually.
9828 There are two big advantages in doing things this way. The extensions
9829 can be looked up immediately and no longer need to be "added" using
9830 X509V3_add_standard_extensions(): this function now does nothing.
9831 [Side note: I get *lots* of email saying the extension code doesn't
9832 work because people forget to call this function]
9833 Also no dynamic allocation is done unless new extensions are added:
9834 so if we don't add custom extensions there is no need to call
9835 X509V3_EXT_cleanup().
9838 *) Modify enc utility's salting as follows: make salting the default. Add a
9839 magic header, so unsalted files fail gracefully instead of just decrypting
9840 to garbage. This is because not salting is a big security hole, so people
9841 should be discouraged from doing it.
9844 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9845 digest to be passed on the command line but it only used this
9846 parameter when signing a certificate. Modified so all relevant
9847 operations are affected by the digest parameter including the
9848 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9849 DSA key was used because it didn't fix the digest.
9852 *) Initial certificate chain verify code. Currently tests the untrusted
9853 certificates for consistency with the verify purpose (which is set
9854 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9856 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9857 this is because it will reject chains with invalid extensions whereas
9858 every previous version of OpenSSL and SSLeay made no checks at all.
9860 Trust code: checks the root CA for the relevant trust settings. Trust
9861 settings have an initial value consistent with the verify purpose: e.g.
9862 if the verify purpose is for SSL client use it expects the CA to be
9863 trusted for SSL client use. However the default value can be changed to
9864 permit custom trust settings: one example of this would be to only trust
9865 certificates from a specific "secure" set of CAs.
9867 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9868 which should be used for version portability: especially since the
9869 verify structure is likely to change more often now.
9871 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9872 to set them. If not set then assume SSL clients will verify SSL servers
9875 Two new options to the verify program: -untrusted allows a set of
9876 untrusted certificates to be passed in and -purpose which sets the
9877 intended purpose of the certificate. If a purpose is set then the
9878 new chain verify code is used to check extension consistency.
9881 *) Support for the authority information access extension.
9884 *) Modify RSA and DSA PEM read routines to transparently handle
9885 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9886 public keys in a format compatible with certificate
9887 SubjectPublicKeyInfo structures. Unfortunately there were already
9888 functions called *_PublicKey_* which used various odd formats so
9889 these are retained for compatibility: however the DSA variants were
9890 never in a public release so they have been deleted. Changed dsa/rsa
9891 utilities to handle the new format: note no releases ever handled public
9892 keys so we should be OK.
9894 The primary motivation for this change is to avoid the same fiasco
9895 that dogs private keys: there are several incompatible private key
9896 formats some of which are standard and some OpenSSL specific and
9897 require various evil hacks to allow partial transparent handling and
9898 even then it doesn't work with DER formats. Given the option anything
9899 other than PKCS#8 should be dumped: but the other formats have to
9900 stay in the name of compatibility.
9902 With public keys and the benefit of hindsight one standard format
9903 is used which works with EVP_PKEY, RSA or DSA structures: though
9904 it clearly returns an error if you try to read the wrong kind of key.
9906 Added a -pubkey option to the 'x509' utility to output the public key.
9907 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9908 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9909 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9910 that do the same as the EVP_PKEY_assign_*() except they up the
9911 reference count of the added key (they don't "swallow" the
9915 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9916 CRLs would fail if the file contained no certificates or no CRLs:
9917 added a new function to read in both types and return the number
9918 read: this means that if none are read it will be an error. The
9919 DER versions of the certificate and CRL reader would always fail
9920 because it isn't possible to mix certificates and CRLs in DER format
9921 without choking one or the other routine. Changed this to just read
9922 a certificate: this is the best we can do. Also modified the code
9923 in apps/verify.c to take notice of return codes: it was previously
9924 attempting to read in certificates from NULL pointers and ignoring
9925 any errors: this is one reason why the cert and CRL reader seemed
9926 to work. It doesn't check return codes from the default certificate
9927 routines: these may well fail if the certificates aren't installed.
9930 *) Code to support otherName option in GeneralName.
9933 *) First update to verify code. Change the verify utility
9934 so it warns if it is passed a self signed certificate:
9935 for consistency with the normal behaviour. X509_verify
9936 has been modified to it will now verify a self signed
9937 certificate if *exactly* the same certificate appears
9938 in the store: it was previously impossible to trust a
9939 single self signed certificate. This means that:
9940 openssl verify ss.pem
9941 now gives a warning about a self signed certificate but
9942 openssl verify -CAfile ss.pem ss.pem
9946 *) For servers, store verify_result in SSL_SESSION data structure
9947 (and add it to external session representation).
9948 This is needed when client certificate verifications fails,
9949 but an application-provided verification callback (set by
9950 SSL_CTX_set_cert_verify_callback) allows accepting the session
9951 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9952 but returns 1): When the session is reused, we have to set
9953 ssl->verify_result to the appropriate error code to avoid
9955 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9957 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9958 case in PKCS7_dataInit() where the signed PKCS7 structure
9959 didn't contain any existing data because it was being created.
9960 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9962 *) Add a salt to the key derivation routines in enc.c. This
9963 forms the first 8 bytes of the encrypted file. Also add a
9964 -S option to allow a salt to be input on the command line.
9967 *) New function X509_cmp(). Oddly enough there wasn't a function
9968 to compare two certificates. We do this by working out the SHA1
9969 hash and comparing that. X509_cmp() will be needed by the trust
9973 *) SSL_get1_session() is like SSL_get_session(), but increments
9974 the reference count in the SSL_SESSION returned.
9975 [Geoff Thorpe <geoff@eu.c2.net>]
9977 *) Fix for 'req': it was adding a null to request attributes.
9978 Also change the X509_LOOKUP and X509_INFO code to handle
9979 certificate auxiliary information.
9982 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9986 *) Add the possibility to add extra information to the memory leak
9987 detecting output, to form tracebacks, showing from where each
9988 allocation was originated: CRYPTO_push_info("constant string") adds
9989 the string plus current file name and line number to a per-thread
9990 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9991 is like calling CYRPTO_pop_info() until the stack is empty.
9992 Also updated memory leak detection code to be multi-thread-safe.
9995 *) Add options -text and -noout to pkcs7 utility and delete the
9996 encryption options which never did anything. Update docs.
9999 *) Add options to some of the utilities to allow the pass phrase
10000 to be included on either the command line (not recommended on
10001 OSes like Unix) or read from the environment. Update the
10002 manpages and fix a few bugs.
10005 *) Add a few manpages for some of the openssl commands.
10008 *) Fix the -revoke option in ca. It was freeing up memory twice,
10009 leaking and not finding already revoked certificates.
10012 *) Extensive changes to support certificate auxiliary information.
10013 This involves the use of X509_CERT_AUX structure and X509_AUX
10014 functions. An X509_AUX function such as PEM_read_X509_AUX()
10015 can still read in a certificate file in the usual way but it
10016 will also read in any additional "auxiliary information". By
10017 doing things this way a fair degree of compatibility can be
10018 retained: existing certificates can have this information added
10019 using the new 'x509' options.
10021 Current auxiliary information includes an "alias" and some trust
10022 settings. The trust settings will ultimately be used in enhanced
10023 certificate chain verification routines: currently a certificate
10024 can only be trusted if it is self signed and then it is trusted
10028 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10029 The problem was that one of the replacement routines had not been working
10030 since SSLeay releases. For now the offending routine has been replaced
10031 with non-optimised assembler. Even so, this now gives around 95%
10032 performance improvement for 1024 bit RSA signs.
10035 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10036 handling. Most clients have the effective key size in bits equal to
10037 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10038 A few however don't do this and instead use the size of the decrypted key
10039 to determine the RC2 key length and the AlgorithmIdentifier to determine
10040 the effective key length. In this case the effective key length can still
10041 be 40 bits but the key length can be 168 bits for example. This is fixed
10042 by manually forcing an RC2 key into the EVP_PKEY structure because the
10043 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10044 the key length and effective key length are equal.
10047 *) Add a bunch of functions that should simplify the creation of
10048 X509_NAME structures. Now you should be able to do:
10049 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10050 and have it automatically work out the correct field type and fill in
10051 the structures. The more adventurous can try:
10052 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10053 and it will (hopefully) work out the correct multibyte encoding.
10056 *) Change the 'req' utility to use the new field handling and multibyte
10057 copy routines. Before the DN field creation was handled in an ad hoc
10058 way in req, ca, and x509 which was rather broken and didn't support
10059 BMPStrings or UTF8Strings. Since some software doesn't implement
10060 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10061 using the dirstring_type option. See the new comment in the default
10062 openssl.cnf for more info.
10065 *) Make crypto/rand/md_rand.c more robust:
10066 - Assure unique random numbers after fork().
10067 - Make sure that concurrent threads access the global counter and
10068 md serializably so that we never lose entropy in them
10069 or use exactly the same state in multiple threads.
10070 Access to the large state is not always serializable because
10071 the additional locking could be a performance killer, and
10072 md should be large enough anyway.
10075 *) New file apps/app_rand.c with commonly needed functionality
10076 for handling the random seed file.
10078 Use the random seed file in some applications that previously did not:
10080 dsaparam -genkey (which also ignored its '-rand' option),
10083 x509 (when signing).
10084 Except on systems with /dev/urandom, it is crucial to have a random
10085 seed file at least for key creation, DSA signing, and for DH exchanges;
10086 for RSA signatures we could do without one.
10088 gendh and gendsa (unlike genrsa) used to read only the first byte
10089 of each file listed in the '-rand' option. The function as previously
10090 found in genrsa is now in app_rand.c and is used by all programs
10091 that support '-rand'.
10094 *) In RAND_write_file, use mode 0600 for creating files;
10095 don't just chmod when it may be too late.
10098 *) Report an error from X509_STORE_load_locations
10099 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10102 *) New function ASN1_mbstring_copy() this copies a string in either
10103 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10104 into an ASN1_STRING type. A mask of permissible types is passed
10105 and it chooses the "minimal" type to use or an error if not type
10109 *) Add function equivalents to the various macros in asn1.h. The old
10110 macros are retained with an M_ prefix. Code inside the library can
10111 use the M_ macros. External code (including the openssl utility)
10112 should *NOT* in order to be "shared library friendly".
10115 *) Add various functions that can check a certificate's extensions
10116 to see if it usable for various purposes such as SSL client,
10117 server or S/MIME and CAs of these types. This is currently
10118 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10119 verification. Also added a -purpose flag to x509 utility to
10120 print out all the purposes.
10123 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10127 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10128 for, obtain and decode and extension and obtain its critical flag.
10129 This allows all the necessary extension code to be handled in a
10130 single function call.
10133 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10134 platforms. See crypto/rc4/rc4_enc.c for further details.
10137 *) New -noout option to asn1parse. This causes no output to be produced
10138 its main use is when combined with -strparse and -out to extract data
10139 from a file (which may not be in ASN.1 format).
10142 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10143 when producing the local key id.
10144 [Richard Levitte <levitte@stacken.kth.se>]
10146 *) New option -dhparam in s_server. This allows a DH parameter file to be
10147 stated explicitly. If it is not stated then it tries the first server
10148 certificate file. The previous behaviour hard coded the filename
10152 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10153 a public key to be input or output. For example:
10154 openssl rsa -in key.pem -pubout -out pubkey.pem
10155 Also added necessary DSA public key functions to handle this.
10158 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10159 in the message. This was handled by allowing
10160 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10161 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10163 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10164 to the end of the strings whereas this didn't. This would cause problems
10165 if strings read with d2i_ASN1_bytes() were later modified.
10166 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10168 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10169 data and it contains EOF it will end up returning an error. This is
10170 caused by input 46 bytes long. The cause is due to the way base64
10171 BIOs find the start of base64 encoded data. They do this by trying a
10172 trial decode on each line until they find one that works. When they
10173 do a flag is set and it starts again knowing it can pass all the
10174 data directly through the decoder. Unfortunately it doesn't reset
10175 the context it uses. This means that if EOF is reached an attempt
10176 is made to pass two EOFs through the context and this causes the
10177 resulting error. This can also cause other problems as well. As is
10178 usual with these problems it takes *ages* to find and the fix is
10179 trivial: move one line.
10180 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10182 *) Ugly workaround to get s_client and s_server working under Windows. The
10183 old code wouldn't work because it needed to select() on sockets and the
10184 tty (for keypresses and to see if data could be written). Win32 only
10185 supports select() on sockets so we select() with a 1s timeout on the
10186 sockets and then see if any characters are waiting to be read, if none
10187 are present then we retry, we also assume we can always write data to
10188 the tty. This isn't nice because the code then blocks until we've
10189 received a complete line of data and it is effectively polling the
10190 keyboard at 1s intervals: however it's quite a bit better than not
10191 working at all :-) A dedicated Windows application might handle this
10192 with an event loop for example.
10195 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10196 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10197 will be called when RSA_sign() and RSA_verify() are used. This is useful
10198 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10199 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10200 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10201 This necessitated the support of an extra signature type NID_md5_sha1
10202 for SSL signatures and modifications to the SSL library to use it instead
10203 of calling RSA_public_decrypt() and RSA_private_encrypt().
10206 *) Add new -verify -CAfile and -CApath options to the crl program, these
10207 will lookup a CRL issuers certificate and verify the signature in a
10208 similar way to the verify program. Tidy up the crl program so it
10209 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10210 less strict. It will now permit CRL extensions even if it is not
10211 a V2 CRL: this will allow it to tolerate some broken CRLs.
10214 *) Initialize all non-automatic variables each time one of the openssl
10215 sub-programs is started (this is necessary as they may be started
10216 multiple times from the "OpenSSL>" prompt).
10217 [Lennart Bang, Bodo Moeller]
10219 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10220 removing all other RSA functionality (this is what NO_RSA does). This
10221 is so (for example) those in the US can disable those operations covered
10222 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10226 *) Non-copying interface to BIO pairs.
10227 (still largely untested)
10230 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10231 ASCII string. This was handled independently in various places before.
10234 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10235 UTF8 strings a character at a time.
10238 *) Use client_version from client hello to select the protocol
10239 (s23_srvr.c) and for RSA client key exchange verification
10240 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10243 *) Add various utility functions to handle SPKACs, these were previously
10244 handled by poking round in the structure internals. Added new function
10245 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10246 print, verify and generate SPKACs. Based on an original idea from
10247 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10250 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10253 *) Allow the config file extension section to be overwritten on the
10254 command line. Based on an original idea from Massimiliano Pala
10255 <madwolf@comune.modena.it>. The new option is called -extensions
10256 and can be applied to ca, req and x509. Also -reqexts to override
10257 the request extensions in req and -crlexts to override the crl extensions
10261 *) Add new feature to the SPKAC handling in ca. Now you can include
10262 the same field multiple times by preceding it by "XXXX." for example:
10265 this is the same syntax as used in the req config file.
10268 *) Allow certificate extensions to be added to certificate requests. These
10269 are specified in a 'req_extensions' option of the req section of the
10270 config file. They can be printed out with the -text option to req but
10271 are otherwise ignored at present.
10274 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10275 data read consists of only the final block it would not decrypted because
10276 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10277 A misplaced 'break' also meant the decrypted final block might not be
10278 copied until the next read.
10281 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10282 a few extra parameters to the DH structure: these will be useful if
10283 for example we want the value of 'q' or implement X9.42 DH.
10286 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10287 provides hooks that allow the default DSA functions or functions on a
10288 "per key" basis to be replaced. This allows hardware acceleration and
10289 hardware key storage to be handled without major modification to the
10290 library. Also added low level modexp hooks and CRYPTO_EX structure and
10291 associated functions.
10294 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10295 as "read only": it can't be written to and the buffer it points to will
10296 not be freed. Reading from a read only BIO is much more efficient than
10297 a normal memory BIO. This was added because there are several times when
10298 an area of memory needs to be read from a BIO. The previous method was
10299 to create a memory BIO and write the data to it, this results in two
10300 copies of the data and an O(n^2) reading algorithm. There is a new
10301 function BIO_new_mem_buf() which creates a read only memory BIO from
10302 an area of memory. Also modified the PKCS#7 routines to use read only
10306 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10307 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10308 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10309 but a retry condition occurred while trying to read the rest.
10312 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10313 NID_pkcs7_encrypted by default: this was wrong since this should almost
10314 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10315 the encrypted data type: this is a more sensible place to put it and it
10316 allows the PKCS#12 code to be tidied up that duplicated this
10320 *) Changed obj_dat.pl script so it takes its input and output files on
10321 the command line. This should avoid shell escape redirection problems
10325 *) Initial support for certificate extension requests, these are included
10326 in things like Xenroll certificate requests. Included functions to allow
10327 extensions to be obtained and added.
10330 *) -crlf option to s_client and s_server for sending newlines as
10331 CRLF (as required by many protocols).
10334 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10336 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10337 [Ralf S. Engelschall]
10339 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10340 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10342 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10346 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10347 DH parameters/keys (q is lost during that conversion, but the resulting
10348 DH parameters contain its length).
10350 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10351 much faster than DH_generate_parameters (which creates parameters
10352 where p = 2*q + 1), and also the smaller q makes DH computations
10353 much more efficient (160-bit exponentiation instead of 1024-bit
10354 exponentiation); so this provides a convenient way to support DHE
10355 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10356 utter importance to use
10357 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10359 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10360 when such DH parameters are used, because otherwise small subgroup
10361 attacks may become possible!
10364 *) Avoid memory leak in i2d_DHparams.
10367 *) Allow the -k option to be used more than once in the enc program:
10368 this allows the same encrypted message to be read by multiple recipients.
10371 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10372 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10373 it will always use the numerical form of the OID, even if it has a short
10377 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10378 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10379 otherwise bn_mod_exp was called. In the case of hardware keys for example
10380 no private key components need be present and it might store extra data
10381 in the RSA structure, which cannot be accessed from bn_mod_exp.
10382 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10383 private key operations.
10386 *) Added support for SPARC Linux.
10389 *) pem_password_cb function type incompatibly changed from
10390 typedef int pem_password_cb(char *buf, int size, int rwflag);
10392 ....(char *buf, int size, int rwflag, void *userdata);
10393 so that applications can pass data to their callbacks:
10394 The PEM[_ASN1]_{read,write}... functions and macros now take an
10395 additional void * argument, which is just handed through whenever
10396 the password callback is called.
10397 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10399 New function SSL_CTX_set_default_passwd_cb_userdata.
10401 Compatibility note: As many C implementations push function arguments
10402 onto the stack in reverse order, the new library version is likely to
10403 interoperate with programs that have been compiled with the old
10404 pem_password_cb definition (PEM_whatever takes some data that
10405 happens to be on the stack as its last argument, and the callback
10406 just ignores this garbage); but there is no guarantee whatsoever that
10409 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10410 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10411 problems not only on Windows, but also on some Unix platforms.
10412 To avoid problematic command lines, these definitions are now in an
10413 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10414 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10417 *) MIPS III/IV assembler module is reimplemented.
10420 *) More DES library cleanups: remove references to srand/rand and
10421 delete an unused file.
10424 *) Add support for the the free Netwide assembler (NASM) under Win32,
10425 since not many people have MASM (ml) and it can be hard to obtain.
10426 This is currently experimental but it seems to work OK and pass all
10427 the tests. Check out INSTALL.W32 for info.
10430 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10431 without temporary keys kept an extra copy of the server key,
10432 and connections with temporary keys did not free everything in case
10436 *) New function RSA_check_key and new openssl rsa option -check
10437 for verifying the consistency of RSA keys.
10438 [Ulf Moeller, Bodo Moeller]
10440 *) Various changes to make Win32 compile work:
10441 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10442 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10443 comparison" warnings.
10444 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10447 *) Add a debugging option to PKCS#5 v2 key generation function: when
10448 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10449 derived keys are printed to stderr.
10452 *) Copy the flags in ASN1_STRING_dup().
10453 [Roman E. Pavlov <pre@mo.msk.ru>]
10455 *) The x509 application mishandled signing requests containing DSA
10456 keys when the signing key was also DSA and the parameters didn't match.
10458 It was supposed to omit the parameters when they matched the signing key:
10459 the verifying software was then supposed to automatically use the CA's
10460 parameters if they were absent from the end user certificate.
10462 Omitting parameters is no longer recommended. The test was also
10463 the wrong way round! This was probably due to unusual behaviour in
10464 EVP_cmp_parameters() which returns 1 if the parameters match.
10465 This meant that parameters were omitted when they *didn't* match and
10466 the certificate was useless. Certificates signed with 'ca' didn't have
10468 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10470 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10471 The interface is as follows:
10472 Applications can use
10473 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10474 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10475 "off" is now the default.
10476 The library internally uses
10477 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10478 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10479 to disable memory-checking temporarily.
10481 Some inconsistent states that previously were possible (and were
10482 even the default) are now avoided.
10484 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10485 with each memory chunk allocated; this is occasionally more helpful
10486 than just having a counter.
10488 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10490 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10494 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10495 which largely parallels "options", but is for changing API behaviour,
10496 whereas "options" are about protocol behaviour.
10497 Initial "mode" flags are:
10499 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10500 a single record has been written.
10501 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10502 retries use the same buffer location.
10503 (But all of the contents must be
10507 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10510 *) Fix problems with no-hmac etc.
10511 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10513 *) New functions RSA_get_default_method(), RSA_set_method() and
10514 RSA_get_method(). These allows replacement of RSA_METHODs without having
10515 to mess around with the internals of an RSA structure.
10518 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10519 Also really enable memory leak checks in openssl.c and in some
10521 [Chad C. Mulligan, Bodo Moeller]
10523 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10524 up the length of negative integers. This has now been simplified to just
10525 store the length when it is first determined and use it later, rather
10526 than trying to keep track of where data is copied and updating it to
10528 [Steve Henson, reported by Brien Wheeler
10529 <bwheeler@authentica-security.com>]
10531 *) Add a new function PKCS7_signatureVerify. This allows the verification
10532 of a PKCS#7 signature but with the signing certificate passed to the
10533 function itself. This contrasts with PKCS7_dataVerify which assumes the
10534 certificate is present in the PKCS#7 structure. This isn't always the
10535 case: certificates can be omitted from a PKCS#7 structure and be
10536 distributed by "out of band" means (such as a certificate database).
10539 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10540 function prototypes in pem.h, also change util/mkdef.pl to add the
10541 necessary function names.
10544 *) mk1mf.pl (used by Windows builds) did not properly read the
10545 options set by Configure in the top level Makefile, and Configure
10546 was not even able to write more than one option correctly.
10547 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10550 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10551 file to be loaded from a BIO or FILE pointer. The BIO version will
10552 for example allow memory BIOs to contain config info.
10555 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10556 Whoever hopes to achieve shared-library compatibility across versions
10557 must use this, not the compile-time macro.
10558 (Exercise 0.9.4: Which is the minimum library version required by
10560 Note: All this applies only to multi-threaded programs, others don't
10564 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10565 through a BIO pair triggered the default case, i.e.
10566 SSLerr(...,SSL_R_UNKNOWN_STATE).
10569 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10570 can use the SSL library even if none of the specific BIOs is
10574 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10575 for the encoded length.
10576 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10578 *) Add initial documentation of the X509V3 functions.
10581 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10582 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10583 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10584 secure PKCS#8 private key format with a high iteration count.
10587 *) Fix determination of Perl interpreter: A perl or perl5
10588 _directory_ in $PATH was also accepted as the interpreter.
10589 [Ralf S. Engelschall]
10591 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10592 wrong with it but it was very old and did things like calling
10593 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10594 unusual formatting.
10597 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10598 to use the new extension code.
10601 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10602 with macros. This should make it easier to change their form, add extra
10603 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10607 *) Add to configuration table a new entry that can specify an alternative
10608 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10609 according to Mark Crispin <MRC@Panda.COM>.
10613 *) DES CBC did not update the IV. Weird.
10616 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10617 Changing the behaviour of the former might break existing programs --
10618 where IV updating is needed, des_ncbc_encrypt can be used.
10621 *) When bntest is run from "make test" it drives bc to check its
10622 calculations, as well as internally checking them. If an internal check
10623 fails, it needs to cause bc to give a non-zero result or make test carries
10624 on without noticing the failure. Fixed.
10627 *) DES library cleanups.
10630 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10631 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10632 ciphers. NOTE: although the key derivation function has been verified
10633 against some published test vectors it has not been extensively tested
10634 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10638 *) Instead of "mkdir -p", which is not fully portable, use new
10639 Perl script "util/mkdir-p.pl".
10642 *) Rewrite the way password based encryption (PBE) is handled. It used to
10643 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10644 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10645 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10646 the 'parameter' field of the AlgorithmIdentifier is passed to the
10647 underlying key generation function so it must do its own ASN1 parsing.
10648 This has also changed the EVP_PBE_CipherInit() function which now has a
10649 'parameter' argument instead of literal salt and iteration count values
10650 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10653 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10654 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10655 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10656 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10657 value was just used as a "magic string" and not used directly its
10658 value doesn't matter.
10661 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10665 *) "linux-sparc64" configuration (ultrapenguin).
10666 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10667 "linux-sparc" configuration.
10668 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10670 *) config now generates no-xxx options for missing ciphers.
10673 *) Support the EBCDIC character set (work in progress).
10674 File ebcdic.c not yet included because it has a different license.
10675 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10677 *) Support BS2000/OSD-POSIX.
10678 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10680 *) Make callbacks for key generation use void * instead of char *.
10683 *) Make S/MIME samples compile (not yet tested).
10686 *) Additional typesafe stacks.
10689 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10693 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10695 *) New configuration variant "sco5-gcc".
10697 *) Updated some demos.
10698 [Sean O Riordain, Wade Scholine]
10700 *) Add missing BIO_free at exit of pkcs12 application.
10703 *) Fix memory leak in conf.c.
10706 *) Updates for Win32 to assembler version of MD5.
10709 *) Set #! path to perl in apps/der_chop to where we found it
10710 instead of using a fixed path.
10713 *) SHA library changes for irix64-mips4-cc.
10716 *) Improvements for VMS support.
10720 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10722 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10723 This also avoids the problems with SC4.2 and unpatched SC5.
10724 [Andy Polyakov <appro@fy.chalmers.se>]
10726 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10727 These are required because of the typesafe stack would otherwise break
10728 existing code. If old code used a structure member which used to be STACK
10729 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10730 sk_num or sk_value it would produce an error because the num, data members
10731 are not present in STACK_OF. Now it just produces a warning. sk_set
10732 replaces the old method of assigning a value to sk_value
10733 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10734 that does this will no longer work (and should use sk_set instead) but
10735 this could be regarded as a "questionable" behaviour anyway.
10738 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10739 correctly handle encrypted S/MIME data.
10742 *) Change type of various DES function arguments from des_cblock
10743 (which means, in function argument declarations, pointer to char)
10744 to des_cblock * (meaning pointer to array with 8 char elements),
10745 which allows the compiler to do more typechecking; it was like
10746 that back in SSLeay, but with lots of ugly casts.
10748 Introduce new type const_des_cblock.
10751 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10752 problems: find RecipientInfo structure that matches recipient certificate
10753 and initialise the ASN1 structures properly based on passed cipher.
10756 *) Belatedly make the BN tests actually check the results.
10759 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10760 to and from BNs: it was completely broken. New compilation option
10761 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10762 key elements as negative integers.
10765 *) Reorganize and speed up MD5.
10766 [Andy Polyakov <appro@fy.chalmers.se>]
10769 [Richard Levitte <richard@levitte.org>]
10771 *) New option -out to asn1parse to allow the parsed structure to be
10772 output to a file. This is most useful when combined with the -strparse
10773 option to examine the output of things like OCTET STRINGS.
10776 *) Make SSL library a little more fool-proof by not requiring any longer
10777 that SSL_set_{accept,connect}_state be called before
10778 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10779 in many applications because usually everything *appeared* to work as
10780 intended anyway -- now it really works as intended).
10783 *) Move openssl.cnf out of lib/.
10786 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10787 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10788 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10789 [Ralf S. Engelschall]
10791 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10792 handle PKCS#7 enveloped data properly.
10793 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10795 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10796 copying pointers. The cert_st handling is changed by this in
10797 various ways (and thus what used to be known as ctx->default_cert
10798 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10799 any longer when s->cert does not give us what we need).
10800 ssl_cert_instantiate becomes obsolete by this change.
10801 As soon as we've got the new code right (possibly it already is?),
10802 we have solved a couple of bugs of the earlier code where s->cert
10803 was used as if it could not have been shared with other SSL structures.
10805 Note that using the SSL API in certain dirty ways now will result
10806 in different behaviour than observed with earlier library versions:
10807 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10808 does not influence s as it used to.
10810 In order to clean up things more thoroughly, inside SSL_SESSION
10811 we don't use CERT any longer, but a new structure SESS_CERT
10812 that holds per-session data (if available); currently, this is
10813 the peer's certificate chain and, for clients, the server's certificate
10814 and temporary key. CERT holds only those values that can have
10815 meaningful defaults in an SSL_CTX.
10818 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10819 from the internal representation. Various PKCS#7 fixes: remove some
10820 evil casts and set the enc_dig_alg field properly based on the signing
10824 *) Allow PKCS#12 password to be set from the command line or the
10825 environment. Let 'ca' get its config file name from the environment
10826 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10830 *) Allow certificate policies extension to use an IA5STRING for the
10831 organization field. This is contrary to the PKIX definition but
10832 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10836 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10837 without disallowing inline assembler and the like for non-pedantic builds.
10840 *) Support Borland C++ builder.
10841 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10843 *) Support Mingw32.
10846 *) SHA-1 cleanups and performance enhancements.
10847 [Andy Polyakov <appro@fy.chalmers.se>]
10849 *) Sparc v8plus assembler for the bignum library.
10850 [Andy Polyakov <appro@fy.chalmers.se>]
10852 *) Accept any -xxx and +xxx compiler options in Configure.
10855 *) Update HPUX configuration.
10858 *) Add missing sk_<type>_unshift() function to safestack.h
10859 [Ralf S. Engelschall]
10861 *) New function SSL_CTX_use_certificate_chain_file that sets the
10862 "extra_cert"s in addition to the certificate. (This makes sense
10863 only for "PEM" format files, as chains as a whole are not
10867 *) Support verify_depth from the SSL API.
10868 x509_vfy.c had what can be considered an off-by-one-error:
10869 Its depth (which was not part of the external interface)
10870 was actually counting the number of certificates in a chain;
10871 now it really counts the depth.
10874 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10875 instead of X509err, which often resulted in confusing error
10876 messages since the error codes are not globally unique
10877 (e.g. an alleged error in ssl3_accept when a certificate
10878 didn't match the private key).
10880 *) New function SSL_CTX_set_session_id_context that allows to set a default
10881 value (so that you don't need SSL_set_session_id_context for each
10882 connection using the SSL_CTX).
10885 *) OAEP decoding bug fix.
10888 *) Support INSTALL_PREFIX for package builders, as proposed by
10892 *) New Configure options "threads" and "no-threads". For systems
10893 where the proper compiler options are known (currently Solaris
10894 and Linux), "threads" is the default.
10897 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10900 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10901 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10902 such as /usr/local/bin.
10905 *) "make linux-shared" to build shared libraries.
10906 [Niels Poppe <niels@netbox.org>]
10908 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10911 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10912 extension adding in x509 utility.
10915 *) Remove NOPROTO sections and error code comments.
10918 *) Partial rewrite of the DEF file generator to now parse the ANSI
10922 *) New Configure options --prefix=DIR and --openssldir=DIR.
10925 *) Complete rewrite of the error code script(s). It is all now handled
10926 by one script at the top level which handles error code gathering,
10927 header rewriting and C source file generation. It should be much better
10928 than the old method: it now uses a modified version of Ulf's parser to
10929 read the ANSI prototypes in all header files (thus the old K&R definitions
10930 aren't needed for error creation any more) and do a better job of
10931 translating function codes into names. The old 'ASN1 error code imbedded
10932 in a comment' is no longer necessary and it doesn't use .err files which
10933 have now been deleted. Also the error code call doesn't have to appear all
10934 on one line (which resulted in some large lines...).
10937 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10940 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10941 0 (which usually indicates a closed connection), but continue reading.
10944 *) Fix some race conditions.
10947 *) Add support for CRL distribution points extension. Add Certificate
10948 Policies and CRL distribution points documentation.
10951 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10954 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10955 8 of keying material. Merlin has also confirmed interop with this fix
10956 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10957 [Merlin Hughes <merlin@baltimore.ie>]
10959 *) Fix lots of warnings.
10960 [Richard Levitte <levitte@stacken.kth.se>]
10962 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10963 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10964 [Richard Levitte <levitte@stacken.kth.se>]
10966 *) Fix problems with sizeof(long) == 8.
10967 [Andy Polyakov <appro@fy.chalmers.se>]
10969 *) Change functions to ANSI C.
10972 *) Fix typos in error codes.
10973 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10975 *) Remove defunct assembler files from Configure.
10978 *) SPARC v8 assembler BIGNUM implementation.
10979 [Andy Polyakov <appro@fy.chalmers.se>]
10981 *) Support for Certificate Policies extension: both print and set.
10982 Various additions to support the r2i method this uses.
10985 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10986 return a const string when you are expecting an allocated buffer.
10989 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10990 types DirectoryString and DisplayText.
10993 *) Add code to allow r2i extensions to access the configuration database,
10994 add an LHASH database driver and add several ctx helper functions.
10997 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10998 fail when they extended the size of a BIGNUM.
11001 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11002 support typesafe stack.
11005 *) Fix typo in SSL_[gs]et_options().
11006 [Nils Frostberg <nils@medcom.se>]
11008 *) Delete various functions and files that belonged to the (now obsolete)
11009 old X509V3 handling code.
11012 *) New Configure option "rsaref".
11015 *) Don't auto-generate pem.h.
11018 *) Introduce type-safe ASN.1 SETs.
11021 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11022 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11024 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11025 that links with OpenSSL (well at least cause lots of warnings), but fear
11026 not: the conversion is trivial, and it eliminates loads of evil casts. A
11027 few STACKed things have been converted already. Feel free to convert more.
11028 In the fullness of time, I'll do away with the STACK type altogether.
11031 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11032 specified in <certfile> by updating the entry in the index.txt file.
11033 This way one no longer has to edit the index.txt file manually for
11034 revoking a certificate. The -revoke option does the gory details now.
11035 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11037 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11038 `-text' option at all and this way the `-noout -text' combination was
11039 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11040 [Ralf S. Engelschall]
11042 *) Make sure a corresponding plain text error message exists for the
11043 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11044 verify callback function determined that a certificate was revoked.
11045 [Ralf S. Engelschall]
11047 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11048 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11049 all available cipers including rc5, which was forgotten until now.
11050 In order to let the testing shell script know which algorithms
11051 are available, a new (up to now undocumented) command
11052 "openssl list-cipher-commands" is used.
11055 *) Bugfix: s_client occasionally would sleep in select() when
11056 it should have checked SSL_pending() first.
11059 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11060 the raw DSA values prior to ASN.1 encoding.
11063 *) Tweaks to Configure
11064 [Niels Poppe <niels@netbox.org>]
11066 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11070 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11073 *) New config option to avoid instructions that are illegal on the 80386.
11074 The default code is faster, but requires at least a 486.
11077 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11078 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11079 same as SSL2_VERSION anyway.
11082 *) New "-showcerts" option for s_client.
11085 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11086 application. Various cleanups and fixes.
11089 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11090 modify error routines to work internally. Add error codes and PBE init
11091 to library startup routines.
11094 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11095 packing functions to asn1 and evp. Changed function names and error
11096 codes along the way.
11099 *) PKCS12 integration: and so it begins... First of several patches to
11100 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11101 objects to objects.h
11104 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11105 and display support for Thawte strong extranet extension.
11108 *) Add LinuxPPC support.
11109 [Jeff Dubrule <igor@pobox.org>]
11111 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11112 bn_div_words in alpha.s.
11113 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11115 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11116 OAEP isn't supported when OpenSSL is built with RSAref.
11117 [Ulf Moeller <ulf@fitug.de>]
11119 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11120 so they no longer are missing under -DNOPROTO.
11121 [Soren S. Jorvang <soren@t.dk>]
11124 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11126 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11127 doesn't work when the session is reused. Coming soon!
11130 *) Fix a security hole, that allows sessions to be reused in the wrong
11131 context thus bypassing client cert protection! All software that uses
11132 client certs and session caches in multiple contexts NEEDS PATCHING to
11133 allow session reuse! A fuller solution is in the works.
11134 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11136 *) Some more source tree cleanups (removed obsolete files
11137 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11138 permission on "config" script to be executable) and a fix for the INSTALL
11140 [Ulf Moeller <ulf@fitug.de>]
11142 *) Remove some legacy and erroneous uses of malloc, free instead of
11144 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11146 *) Make rsa_oaep_test return non-zero on error.
11147 [Ulf Moeller <ulf@fitug.de>]
11149 *) Add support for native Solaris shared libraries. Configure
11150 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11151 if someone would make that last step automatic.
11152 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11154 *) ctx_size was not built with the right compiler during "make links". Fixed.
11157 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11158 except NULL ciphers". This means the default cipher list will no longer
11159 enable NULL ciphers. They need to be specifically enabled e.g. with
11160 the string "DEFAULT:eNULL".
11163 *) Fix to RSA private encryption routines: if p < q then it would
11164 occasionally produce an invalid result. This will only happen with
11165 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11168 *) Be less restrictive and allow also `perl util/perlpath.pl
11169 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11170 because this way one can also use an interpreter named `perl5' (which is
11171 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11172 installed as `perl').
11173 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11175 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11176 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11178 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11179 advapi32.lib to Win32 build and change the pem test comparison
11180 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11181 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11182 and crypto/des/ede_cbcm_enc.c.
11185 *) DES quad checksum was broken on big-endian architectures. Fixed.
11188 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11189 Win32 test batch file so it (might) work again. The Win32 test batch file
11190 is horrible: I feel ill....
11193 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11194 in e_os.h. Audit of header files to check ANSI and non ANSI
11195 sections: 10 functions were absent from non ANSI section and not exported
11196 from Windows DLLs. Fixed up libeay.num for new functions.
11199 *) Make `openssl version' output lines consistent.
11200 [Ralf S. Engelschall]
11202 *) Fix Win32 symbol export lists for BIO functions: Added
11203 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11204 to ms/libeay{16,32}.def.
11205 [Ralf S. Engelschall]
11207 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11208 fine under Unix and passes some trivial tests I've now added. But the
11209 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11210 added to make sure no one expects that this stuff really works in the
11211 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11212 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11214 [Ralf S. Engelschall]
11216 *) Fix the generation of two part addresses in perl.
11217 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11219 *) Add config entry for Linux on MIPS.
11220 [John Tobey <jtobey@channel1.com>]
11222 *) Make links whenever Configure is run, unless we are on Windoze.
11225 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11226 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11230 *) Add a useful kludge to allow package maintainers to specify compiler and
11231 other platforms details on the command line without having to patch the
11232 Configure script everytime: One now can use ``perl Configure
11233 <id>:<details>'', i.e. platform ids are allowed to have details appended
11234 to them (separated by colons). This is treated as there would be a static
11235 pre-configured entry in Configure's %table under key <id> with value
11236 <details> and ``perl Configure <id>'' is called. So, when you want to
11237 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11238 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11239 now, which overrides the FreeBSD-elf entry on-the-fly.
11240 [Ralf S. Engelschall]
11242 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11245 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11246 on the `perl Configure ...' command line. This way one can compile
11247 OpenSSL libraries with Position Independent Code (PIC) which is needed
11248 for linking it into DSOs.
11249 [Ralf S. Engelschall]
11251 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11255 *) Cleaned up the LICENSE document: The official contact for any license
11256 questions now is the OpenSSL core team under openssl-core@openssl.org.
11257 And add a paragraph about the dual-license situation to make sure people
11258 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11259 to the OpenSSL toolkit.
11260 [Ralf S. Engelschall]
11262 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11263 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11264 Additionally cleaned up the `make links' target: Remove unnecessary
11265 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11266 to speed processing and no longer clutter the display with confusing
11267 stuff. Instead only the actually done links are displayed.
11268 [Ralf S. Engelschall]
11270 *) Permit null encryption ciphersuites, used for authentication only. It used
11271 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11272 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11276 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11277 signed attributes when verifying signatures (this would break them),
11278 the detached data encoding was wrong and public keys obtained using
11279 X509_get_pubkey() weren't freed.
11282 *) Add text documentation for the BUFFER functions. Also added a work around
11283 to a Win95 console bug. This was triggered by the password read stuff: the
11284 last character typed gets carried over to the next fread(). If you were
11285 generating a new cert request using 'req' for example then the last
11286 character of the passphrase would be CR which would then enter the first
11290 *) Added the new `Includes OpenSSL Cryptography Software' button as
11291 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11292 button and can be used by applications based on OpenSSL to show the
11293 relationship to the OpenSSL project.
11294 [Ralf S. Engelschall]
11296 *) Remove confusing variables in function signatures in files
11297 ssl/ssl_lib.c and ssl/ssl.h.
11298 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11300 *) Don't install bss_file.c under PREFIX/include/
11301 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11303 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11304 functions that return function pointers and has support for NT specific
11305 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11306 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11307 unsigned to signed types: this was killing the Win32 compile.
11310 *) Add new certificate file to stack functions,
11311 SSL_add_dir_cert_subjects_to_stack() and
11312 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11313 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11314 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11315 This means that Apache-SSL and similar packages don't have to mess around
11316 to add as many CAs as they want to the preferred list.
11319 *) Experiment with doxygen documentation. Currently only partially applied to
11321 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11322 openssl.doxy as the configuration file.
11325 *) Get rid of remaining C++-style comments which strict C compilers hate.
11326 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11328 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11329 compiled in by default: it has problems with large keys.
11332 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11333 DH private keys and/or callback functions which directly correspond to
11334 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11335 is needed for applications which have to configure certificates on a
11336 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11338 For the RSA certificate situation is makes no difference, but
11339 for the DSA certificate situation this fixes the "no shared cipher"
11340 problem where the OpenSSL cipher selection procedure failed because the
11341 temporary keys were not overtaken from the context and the API provided
11342 no way to reconfigure them.
11343 The new functions now let applications reconfigure the stuff and they
11344 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11345 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11346 non-public-API function ssl_cert_instantiate() is used as a helper
11347 function and also to reduce code redundancy inside ssl_rsa.c.
11348 [Ralf S. Engelschall]
11350 *) Move s_server -dcert and -dkey options out of the undocumented feature
11351 area because they are useful for the DSA situation and should be
11352 recognized by the users.
11353 [Ralf S. Engelschall]
11355 *) Fix the cipher decision scheme for export ciphers: the export bits are
11356 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11357 SSL_EXP_MASK. So, the original variable has to be used instead of the
11358 already masked variable.
11359 [Richard Levitte <levitte@stacken.kth.se>]
11361 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11362 [Richard Levitte <levitte@stacken.kth.se>]
11364 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11365 from `int' to `unsigned int' because it's a length and initialized by
11366 EVP_DigestFinal() which expects an `unsigned int *'.
11367 [Richard Levitte <levitte@stacken.kth.se>]
11369 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11370 script. Instead use the usual Shell->Perl transition trick.
11371 [Ralf S. Engelschall]
11373 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11374 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11375 -noout -modulus' as it's already the case for `openssl rsa -noout
11376 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11377 currently the public key is printed (a decision which was already done by
11378 `openssl dsa -modulus' in the past) which serves a similar purpose.
11379 Additionally the NO_RSA no longer completely removes the whole -modulus
11380 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11382 [Ralf S. Engelschall]
11384 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11385 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11386 [Arne Ansper <arne@ats.cyber.ee>]
11388 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11389 to be added. Now both 'req' and 'ca' can use new objects defined in the
11393 *) Add cool BIO that does syslog (or event log on NT).
11394 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11396 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11397 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11398 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11399 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11402 *) Add preliminary config info for new extension code.
11405 *) Make RSA_NO_PADDING really use no padding.
11406 [Ulf Moeller <ulf@fitug.de>]
11408 *) Generate errors when private/public key check is done.
11411 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11412 for some CRL extensions and new objects added.
11415 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11416 key usage extension and fuller support for authority key id.
11419 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11420 padding method for RSA, which is recommended for new applications in PKCS
11421 #1 v2.0 (RFC 2437, October 1998).
11422 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11423 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11424 against Bleichbacher's attack on RSA.
11425 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11428 *) Updates to the new SSL compression code
11429 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11431 *) Fix so that the version number in the master secret, when passed
11432 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11433 (because the server will not accept higher), that the version number
11434 is 0x03,0x01, not 0x03,0x00
11435 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11437 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11438 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11439 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11442 *) Support for RAW extensions where an arbitrary extension can be
11443 created by including its DER encoding. See apps/openssl.cnf for
11447 *) Make sure latest Perl versions don't interpret some generated C array
11448 code as Perl array code in the crypto/err/err_genc.pl script.
11449 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11451 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11452 not many people have the assembler. Various Win32 compilation fixes and
11453 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11454 build instructions.
11457 *) Modify configure script 'Configure' to automatically create crypto/date.h
11458 file under Win32 and also build pem.h from pem.org. New script
11459 util/mkfiles.pl to create the MINFO file on environments that can't do a
11460 'make files': perl util/mkfiles.pl >MINFO should work.
11463 *) Major rework of DES function declarations, in the pursuit of correctness
11464 and purity. As a result, many evil casts evaporated, and some weirdness,
11465 too. You may find this causes warnings in your code. Zapping your evil
11466 casts will probably fix them. Mostly.
11469 *) Fix for a typo in asn1.h. Bug fix to object creation script
11470 obj_dat.pl. It considered a zero in an object definition to mean
11471 "end of object": none of the objects in objects.h have any zeros
11472 so it wasn't spotted.
11473 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11475 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11476 Masking (CBCM). In the absence of test vectors, the best I have been able
11477 to do is check that the decrypt undoes the encrypt, so far. Send me test
11478 vectors if you have them.
11481 *) Correct calculation of key length for export ciphers (too much space was
11482 allocated for null ciphers). This has not been tested!
11485 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11486 message is now correct (it understands "crypto" and "ssl" on its
11487 command line). There is also now an "update" option. This will update
11488 the util/ssleay.num and util/libeay.num files with any new functions.
11490 perl util/mkdef.pl crypto ssl update
11491 it will update them.
11494 *) Overhauled the Perl interface (perl/*):
11495 - ported BN stuff to OpenSSL's different BN library
11496 - made the perl/ source tree CVS-aware
11497 - renamed the package from SSLeay to OpenSSL (the files still contain
11498 their history because I've copied them in the repository)
11499 - removed obsolete files (the test scripts will be replaced
11500 by better Test::Harness variants in the future)
11501 [Ralf S. Engelschall]
11503 *) First cut for a very conservative source tree cleanup:
11504 1. merge various obsolete readme texts into doc/ssleay.txt
11505 where we collect the old documents and readme texts.
11506 2. remove the first part of files where I'm already sure that we no
11507 longer need them because of three reasons: either they are just temporary
11508 files which were left by Eric or they are preserved original files where
11509 I've verified that the diff is also available in the CVS via "cvs diff
11510 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11511 the crypto/md/ stuff).
11512 [Ralf S. Engelschall]
11514 *) More extension code. Incomplete support for subject and issuer alt
11515 name, issuer and authority key id. Change the i2v function parameters
11516 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11517 what that's for :-) Fix to ASN1 macro which messed up
11518 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11521 *) Preliminary support for ENUMERATED type. This is largely copied from the
11525 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11526 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11528 *) Make sure `make rehash' target really finds the `openssl' program.
11529 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11531 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11532 like to hear about it if this slows down other processors.
11535 *) Add CygWin32 platform information to Configure script.
11536 [Alan Batie <batie@aahz.jf.intel.com>]
11538 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11539 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11541 *) New program nseq to manipulate netscape certificate sequences
11544 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11548 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11549 but the BN code had some problems that would cause failures when
11550 doing certificate verification and some other functions.
11551 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11553 *) Add ASN1 and PEM code to support netscape certificate sequences.
11556 *) Add ASN1 and PEM code to support netscape certificate sequences.
11559 *) Add several PKIX and private extended key usage OIDs.
11562 *) Modify the 'ca' program to handle the new extension code. Modify
11563 openssl.cnf for new extension format, add comments.
11566 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11567 and add a sample to openssl.cnf so req -x509 now adds appropriate
11571 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11572 error code, add initial support to X509_print() and x509 application.
11575 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11576 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11577 stuff is currently isolated and isn't even compiled yet.
11580 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11581 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11582 Removed the versions check from X509 routines when loading extensions:
11583 this allows certain broken certificates that don't set the version
11584 properly to be processed.
11587 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11588 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11589 can still be regenerated with "make depend".
11592 *) Spelling mistake in C version of CAST-128.
11593 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11595 *) Changes to the error generation code. The perl script err-code.pl
11596 now reads in the old error codes and retains the old numbers, only
11597 adding new ones if necessary. It also only changes the .err files if new
11598 codes are added. The makefiles have been modified to only insert errors
11599 when needed (to avoid needlessly modifying header files). This is done
11600 by only inserting errors if the .err file is newer than the auto generated
11601 C file. To rebuild all the error codes from scratch (the old behaviour)
11602 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11603 or delete all the .err files.
11606 *) CAST-128 was incorrectly implemented for short keys. The C version has
11607 been fixed, but is untested. The assembler versions are also fixed, but
11608 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11609 to regenerate it if needed.
11610 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11611 Hagino <itojun@kame.net>]
11613 *) File was opened incorrectly in randfile.c.
11614 [Ulf Möller <ulf@fitug.de>]
11616 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11617 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11618 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11619 al: it's just almost always a UTCTime. Note this patch adds new error
11620 codes so do a "make errors" if there are problems.
11623 *) Correct Linux 1 recognition in config.
11624 [Ulf Möller <ulf@fitug.de>]
11626 *) Remove pointless MD5 hash when using DSA keys in ca.
11627 [Anonymous <nobody@replay.com>]
11629 *) Generate an error if given an empty string as a cert directory. Also
11630 generate an error if handed NULL (previously returned 0 to indicate an
11631 error, but didn't set one).
11632 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11634 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11637 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11638 parameters. This was causing a warning which killed off the Win32 compile.
11641 *) Remove C++ style comments from crypto/bn/bn_local.h.
11642 [Neil Costigan <neil.costigan@celocom.com>]
11644 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11645 based on a text string, looking up short and long names and finally
11646 "dot" format. The "dot" format stuff didn't work. Added new function
11647 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11648 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11649 OID is not part of the table.
11652 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11653 X509_LOOKUP_by_alias().
11656 *) Sort openssl functions by name.
11659 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11660 encryption from sample DSA keys (in case anyone is interested the password
11664 *) Make _all_ *_free functions accept a NULL pointer.
11665 [Frans Heymans <fheymans@isaserver.be>]
11667 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11669 [Anonymous <nobody@replay.com>]
11671 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11672 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11674 *) Don't blow it for numeric -newkey arguments to apps/req.
11675 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11677 *) Temp key "for export" tests were wrong in s3_srvr.c.
11678 [Anonymous <nobody@replay.com>]
11680 *) Add prototype for temp key callback functions
11681 SSL_CTX_set_tmp_{rsa,dh}_callback().
11684 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11685 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11688 *) X509_name_add_entry() freed the wrong thing after an error.
11689 [Arne Ansper <arne@ats.cyber.ee>]
11691 *) rsa_eay.c would attempt to free a NULL context.
11692 [Arne Ansper <arne@ats.cyber.ee>]
11694 *) BIO_s_socket() had a broken should_retry() on Windoze.
11695 [Arne Ansper <arne@ats.cyber.ee>]
11697 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11698 [Arne Ansper <arne@ats.cyber.ee>]
11700 *) Make sure the already existing X509_STORE->depth variable is initialized
11701 in X509_STORE_new(), but document the fact that this variable is still
11702 unused in the certificate verification process.
11703 [Ralf S. Engelschall]
11705 *) Fix the various library and apps files to free up pkeys obtained from
11706 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11709 *) Fix reference counting in X509_PUBKEY_get(). This makes
11710 demos/maurice/example2.c work, amongst others, probably.
11711 [Steve Henson and Ben Laurie]
11713 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11714 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11715 are no longer created. This way we have a single and consistent command
11716 line interface `openssl <command>', similar to `cvs <command>'.
11717 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11719 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11720 BIT STRING wrapper always have zero unused bits.
11723 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11726 *) Make the top-level INSTALL documentation easier to understand.
11729 *) Makefiles updated to exit if an error occurs in a sub-directory
11730 make (including if user presses ^C) [Paul Sutton]
11732 *) Make Montgomery context stuff explicit in RSA data structure.
11735 *) Fix build order of pem and err to allow for generated pem.h.
11738 *) Fix renumbering bug in X509_NAME_delete_entry().
11741 *) Enhanced the err-ins.pl script so it makes the error library number
11742 global and can add a library name. This is needed for external ASN1 and
11743 other error libraries.
11746 *) Fixed sk_insert which never worked properly.
11749 *) Fix ASN1 macros so they can handle indefinite length constructed
11750 EXPLICIT tags. Some non standard certificates use these: they can now
11754 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11755 into a single doc/ssleay.txt bundle. This way the information is still
11756 preserved but no longer messes up this directory. Now it's new room for
11757 the new set of documentation files.
11758 [Ralf S. Engelschall]
11760 *) SETs were incorrectly DER encoded. This was a major pain, because they
11761 shared code with SEQUENCEs, which aren't coded the same. This means that
11762 almost everything to do with SETs or SEQUENCEs has either changed name or
11763 number of arguments.
11764 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11766 *) Fix test data to work with the above.
11769 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11770 was already fixed by Eric for 0.9.1 it seems.
11771 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11773 *) Autodetect FreeBSD3.
11776 *) Fix various bugs in Configure. This affects the following platforms:
11780 unixware-2.0-pentium
11784 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11785 before they are needed.
11788 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11792 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11794 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11795 changed SSLeay to OpenSSL in version strings.
11796 [Ralf S. Engelschall]
11798 *) Some fixups to the top-level documents.
11801 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11802 because the symlink to include/ was missing.
11803 [Ralf S. Engelschall]
11805 *) Incorporated the popular no-RSA/DSA-only patches
11806 which allow to compile a RSA-free SSLeay.
11807 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11809 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11810 when "ssleay" is still not found.
11811 [Ralf S. Engelschall]
11813 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11814 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11816 *) Updated the README file.
11817 [Ralf S. Engelschall]
11819 *) Added various .cvsignore files in the CVS repository subdirs
11820 to make a "cvs update" really silent.
11821 [Ralf S. Engelschall]
11823 *) Recompiled the error-definition header files and added
11824 missing symbols to the Win32 linker tables.
11825 [Ralf S. Engelschall]
11827 *) Cleaned up the top-level documents;
11828 o new files: CHANGES and LICENSE
11829 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11830 o merged COPYRIGHT into LICENSE
11831 o removed obsolete TODO file
11832 o renamed MICROSOFT to INSTALL.W32
11833 [Ralf S. Engelschall]
11835 *) Removed dummy files from the 0.9.1b source tree:
11836 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11837 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11838 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11839 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11840 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11841 [Ralf S. Engelschall]
11843 *) Added various platform portability fixes.
11846 *) The Genesis of the OpenSSL rpject:
11847 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11848 Young and Tim J. Hudson created while they were working for C2Net until
11850 [The OpenSSL Project]
11853 Changes between 0.9.0b and 0.9.1b [not released]
11855 *) Updated a few CA certificates under certs/
11858 *) Changed some BIGNUM api stuff.
11861 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11862 DGUX x86, Linux Alpha, etc.
11865 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11866 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11870 *) Add -strparse option to asn1pars program which parses nested
11872 [Dr Stephen Henson <shenson@bigfoot.com>]
11874 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11877 *) DSA fix for "ca" program.
11880 *) Added "-genkey" option to "dsaparam" program.
11883 *) Added RIPE MD160 (rmd160) message digest.
11886 *) Added -a (all) option to "ssleay version" command.
11889 *) Added PLATFORM define which is the id given to Configure.
11892 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11895 *) Extended the ASN.1 parser routines.
11898 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11901 *) Added a BN_CTX to the BN library.
11904 *) Fixed the weak key values in DES library
11907 *) Changed API in EVP library for cipher aliases.
11910 *) Added support for RC2/64bit cipher.
11913 *) Converted the lhash library to the crypto/mem.c functions.
11916 *) Added more recognized ASN.1 object ids.
11919 *) Added more RSA padding checks for SSL/TLS.
11922 *) Added BIO proxy/filter functionality.
11925 *) Added extra_certs to SSL_CTX which can be used
11926 send extra CA certificates to the client in the CA cert chain sending
11927 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11930 *) Now Fortezza is denied in the authentication phase because
11931 this is key exchange mechanism is not supported by SSLeay at all.
11934 *) Additional PKCS1 checks.
11937 *) Support the string "TLSv1" for all TLS v1 ciphers.
11940 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11941 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11944 *) Fixed a few memory leaks.
11947 *) Fixed various code and comment typos.
11950 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11951 bytes sent in the client random.
11952 [Edward Bishop <ebishop@spyglass.com>]