5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
8 it is always safe to #include a header now.
11 *) Removed the aged BC-32 config and all its supporting scripts
14 *) Removed support for Ultrix, Netware, and OS/2.
17 *) Add support for HKDF.
20 *) Add support for blake2b and blake2s
23 *) Added support for "pipelining". Ciphers that have the
24 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
25 encryptions/decryptions simultaneously. There are currently no built-in
26 ciphers with this property but the expectation is that engines will be able
27 to offer it to significantly improve throughput. Support has been extended
28 into libssl so that multiple records for a single connection can be
29 processed in one go (for >=TLS 1.1).
32 *) Added the AFALG engine. This is an async capable engine which is able to
33 offload work to the Linux kernel. In this initial version it only supports
34 AES128-CBC. The kernel must be version 4.1.0 or greater.
37 *) OpenSSL now uses a new threading API. It is no longer necessary to
38 set locking callbacks to use OpenSSL in a multi-threaded environment. There
39 are two supported threading models: pthreads and windows threads. It is
40 also possible to configure OpenSSL at compile time for "no-threads". The
41 old threading API should no longer be used. The functions have been
42 replaced with "no-op" compatibility macros.
43 [Alessandro Ghedini, Matt Caswell]
45 *) Modify behavior of ALPN to invoke callback after SNI/servername
46 callback, such that updates to the SSL_CTX affect ALPN.
49 *) Add SSL_CIPHER queries for authentication and key-exchange.
52 *) Changes to the DEFAULT cipherlist:
53 - Prefer (EC)DHE handshakes over plain RSA.
54 - Prefer AEAD ciphers over legacy ciphers.
55 - Prefer ECDSA over RSA when both certificates are available.
56 - Prefer TLSv1.2 ciphers/PRF.
57 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
61 *) Change the ECC default curve list to be this, in order: x25519,
62 secp256r1, secp521r1, secp384r1.
65 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
66 disabled by default. They can be re-enabled using the
67 enable-weak-ssl-ciphers option to Configure.
70 *) If the server has ALPN configured, but supports no protocols that the
71 client advertises, send a fatal "no_application_protocol" alert.
72 This behaviour is SHALL in RFC 7301, though it isn't universally
73 implemented by other servers.
76 *) Add X25519 support.
77 Integrate support for X25519 into EC library. This includes support
78 for public and private key encoding using the format documented in
79 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
80 OID from that draft, encodes public keys using little endian
81 format in the ECPoint structure and private keys using
82 little endian form in the privateKey field of the ECPrivateKey
83 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
86 Note: the current version supports key generation, public and
87 private key encoding and ECDH key agreement using the EC API.
88 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
92 *) Deprecate SRP_VBASE_get_by_user.
93 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
94 In order to fix an unavoidable memory leak (CVE-2016-0798),
95 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
96 seed, even if the seed is configured.
98 Users should use SRP_VBASE_get1_by_user instead. Note that in
99 SRP_VBASE_get1_by_user, caller must free the returned value. Note
100 also that even though configuring the SRP seed attempts to hide
101 invalid usernames by continuing the handshake with fake
102 credentials, this behaviour is not constant time and no strong
103 guarantees are made that the handshake is indistinguishable from
104 that of a valid user.
107 *) Configuration change; it's now possible to build dynamic engines
108 without having to build shared libraries and vice versa. This
109 only applies to the engines in engines/, those in crypto/engine/
110 will always be built into libcrypto (i.e. "static").
112 Building dynamic engines is enabled by default; to disable, use
113 the configuration option "disable-dynamic-engine".
115 The only requirements for building dynamic engines are the
116 presence of the DSO module and building with position independent
117 code, so they will also automatically be disabled if configuring
118 with "disable-dso" or "disable-pic".
120 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
121 are also taken away from openssl/opensslconf.h, as they are
125 *) Configuration change; if there is a known flag to compile
126 position independent code, it will always be applied on the
127 libcrypto and libssl object files, and never on the application
128 object files. This means other libraries that use routines from
129 libcrypto / libssl can be made into shared libraries regardless
130 of how OpenSSL was configured.
132 If this isn't desirable, the configuration options "disable-pic"
133 or "no-pic" can be used to disable the use of PIC. This will
134 also disable building shared libraries and dynamic engines.
137 *) Removed JPAKE code. It was experimental and has no wide use.
140 *) The INSTALL_PREFIX Makefile variable has been renamed to
141 DESTDIR. That makes for less confusion on what this variable
142 is for. Also, the configuration option --install_prefix is
146 *) Heartbeat for TLS has been removed and is disabled by default
147 for DTLS; configure with enable-heartbeats. Code that uses the
148 old #define's might need to be updated.
149 [Emilia Käsper, Rich Salz]
151 *) Rename REF_CHECK to REF_DEBUG.
154 *) New "unified" build system
156 The "unified" build system is aimed to be a common system for all
157 platforms we support. With it comes new support for VMS.
159 This system builds supports building in a different directory tree
160 than the source tree. It produces one Makefile (for unix family
161 or lookalikes), or one descrip.mms (for VMS).
163 The source of information to make the Makefile / descrip.mms is
164 small files called 'build.info', holding the necessary
165 information for each directory with source to compile, and a
166 template in Configurations, like unix-Makefile.tmpl or
169 We rely heavily on the perl module Text::Template.
172 *) Added support for auto-initialisation and de-initialisation of the library.
173 OpenSSL no longer requires explicit init or deinit routines to be called,
174 except in certain circumstances. See the OPENSSL_init_crypto() and
175 OPENSSL_init_ssl() man pages for further information.
178 *) The arguments to the DTLSv1_listen function have changed. Specifically the
179 "peer" argument is now expected to be a BIO_ADDR object.
181 *) Rewrite of BIO networking library. The BIO library lacked consistent
182 support of IPv6, and adding it required some more extensive
183 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
184 which hold all types of addresses and chains of address information.
185 It also introduces a new API, with functions like BIO_socket,
186 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
187 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
188 have been adapted accordingly.
191 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
195 *) CRIME protection: disable compression by default, even if OpenSSL is
196 compiled with zlib enabled. Applications can still enable compression
197 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
198 using the SSL_CONF library to configure compression.
201 *) The signature of the session callback configured with
202 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
203 was explicitly marked as 'const unsigned char*' instead of
207 *) Always DPURIFY. Remove the use of uninitialized memory in the
208 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
211 *) Removed many obsolete configuration items, including
212 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
213 MD2_CHAR, MD2_INT, MD2_LONG
215 IDEA_SHORT, IDEA_LONG
216 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
217 [Rich Salz, with advice from Andy Polyakov]
219 *) Many BN internals have been moved to an internal header file.
220 [Rich Salz with help from Andy Polyakov]
222 *) Configuration and writing out the results from it has changed.
223 Files such as Makefile include/openssl/opensslconf.h and are now
224 produced through general templates, such as Makefile.in and
225 crypto/opensslconf.h.in and some help from the perl module
228 Also, the center of configuration information is no longer
229 Makefile. Instead, Configure produces a perl module in
230 configdata.pm which holds most of the config data (in the hash
231 table %config), the target data that comes from the target
232 configuration in one of the Configurations/*.conf files (in
236 *) To clarify their intended purposes, the Configure options
237 --prefix and --openssldir change their semantics, and become more
238 straightforward and less interdependent.
240 --prefix shall be used exclusively to give the location INSTALLTOP
241 where programs, scripts, libraries, include files and manuals are
242 going to be installed. The default is now /usr/local.
244 --openssldir shall be used exclusively to give the default
245 location OPENSSLDIR where certificates, private keys, CRLs are
246 managed. This is also where the default openssl.cnf gets
248 If the directory given with this option is a relative path, the
249 values of both the --prefix value and the --openssldir value will
250 be combined to become OPENSSLDIR.
251 The default for --openssldir is INSTALLTOP/ssl.
253 Anyone who uses --openssldir to specify where OpenSSL is to be
254 installed MUST change to use --prefix instead.
257 *) The GOST engine was out of date and therefore it has been removed. An up
258 to date GOST engine is now being maintained in an external repository.
259 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
260 support for GOST ciphersuites (these are only activated if a GOST engine
264 *) EGD is no longer supported by default; use enable-egd when
266 [Ben Kaduk and Rich Salz]
268 *) The distribution now has Makefile.in files, which are used to
269 create Makefile's when Configure is run. *Configure must be run
270 before trying to build now.*
273 *) The return value for SSL_CIPHER_description() for error conditions
277 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
279 Obtaining and performing DNSSEC validation of TLSA records is
280 the application's responsibility. The application provides
281 the TLSA records of its choice to OpenSSL, and these are then
282 used to authenticate the peer.
284 The TLSA records need not even come from DNS. They can, for
285 example, be used to implement local end-entity certificate or
286 trust-anchor "pinning", where the "pin" data takes the form
287 of TLSA records, which can augment or replace verification
288 based on the usual WebPKI public certification authorities.
291 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
292 continues to support deprecated interfaces in default builds.
293 However, applications are strongly advised to compile their
294 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
295 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
296 or the 1.1.0 releases.
298 In environments in which all applications have been ported to
299 not use any deprecated interfaces OpenSSL's Configure script
300 should be used with the --api=1.1.0 option to entirely remove
301 support for the deprecated features from the library and
302 unconditionally disable them in the installed headers.
303 Essentially the same effect can be achieved with the "no-deprecated"
304 argument to Configure, except that this will always restrict
305 the build to just the latest API, rather than a fixed API
308 As applications are ported to future revisions of the API,
309 they should update their compile-time OPENSSL_API_COMPAT define
310 accordingly, but in most cases should be able to continue to
311 compile with later releases.
313 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
314 0x10000000L and 0x00908000L, respectively. However those
315 versions did not support the OPENSSL_API_COMPAT feature, and
316 so applications are not typically tested for explicit support
317 of just the undeprecated features of either release.
320 *) Add support for setting the minimum and maximum supported protocol.
321 It can bet set via the SSL_set_min_proto_version() and
322 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
323 MaxProtcol. It's recommended to use the new APIs to disable
324 protocols instead of disabling individual protocols using
325 SSL_set_options() or SSL_CONF's Protocol. This change also
326 removes support for disabling TLS 1.2 in the OpenSSL TLS
327 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
330 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
333 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
334 and integrates ECDSA and ECDH functionality into EC. Implementations can
335 now redirect key generation and no longer need to convert to or from
338 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
339 include the ec.h header file instead.
342 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
343 ciphers who are no longer supported and drops support the ephemeral RSA key
344 exchange. The LOW ciphers currently doesn't have any ciphers in it.
347 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
348 opaque. For HMAC_CTX, the following constructors and destructors
351 HMAC_CTX *HMAC_CTX_new(void);
352 void HMAC_CTX_free(HMAC_CTX *ctx);
354 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
355 destroy such methods has been added. See EVP_MD_meth_new(3) and
356 EVP_CIPHER_meth_new(3) for documentation.
359 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
360 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
361 EVP_MD_CTX_reset() should be called instead to reinitialise
362 an already created structure.
363 2) For consistency with the majority of our object creators and
364 destructors, EVP_MD_CTX_(create|destroy) were renamed to
365 EVP_MD_CTX_(new|free). The old names are retained as macros
366 for deprecated builds.
369 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
370 cryptographic operations to be performed asynchronously as long as an
371 asynchronous capable engine is used. See the ASYNC_start_job() man page for
372 further details. Libssl has also had this capability integrated with the
373 introduction of the new mode SSL_MODE_ASYNC and associated error
374 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
375 pages. This work was developed in partnership with Intel Corp.
378 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
379 always enabled now. If you want to disable the support you should
380 exclude it using the list of supported ciphers. This also means that the
381 "-no_ecdhe" option has been removed from s_server.
384 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
385 SSL_{CTX_}set1_curves() which can set a list.
388 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
389 curve you want to support using SSL_{CTX_}set1_curves().
392 *) State machine rewrite. The state machine code has been significantly
393 refactored in order to remove much duplication of code and solve issues
394 with the old code (see ssl/statem/README for further details). This change
395 does have some associated API changes. Notably the SSL_state() function
396 has been removed and replaced by SSL_get_state which now returns an
397 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
398 altogether. The previous handshake states defined in ssl.h and ssl3.h have
402 *) All instances of the string "ssleay" in the public API were replaced
403 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
404 Some error codes related to internal RSA_eay API's were renamed.
407 *) The demo files in crypto/threads were moved to demo/threads.
410 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
412 [Matt Caswell, Rich Salz]
414 *) New ASN.1 embed macro.
416 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
417 structure is not allocated: it is part of the parent. That is instead of
425 This reduces memory fragmentation and make it impossible to accidentally
426 set a mandatory field to NULL.
428 This currently only works for some fields specifically a SEQUENCE, CHOICE,
429 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
430 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
434 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
437 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
438 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
439 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
440 DES and RC4 ciphersuites.
443 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
444 This changes the decoding behaviour for some invalid messages,
445 though the change is mostly in the more lenient direction, and
446 legacy behaviour is preserved as much as possible.
449 *) Fix no-stdio build.
450 [ David Woodhouse <David.Woodhouse@intel.com> and also
451 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
453 *) New testing framework
454 The testing framework has been largely rewritten and is now using
455 perl and the perl modules Test::Harness and an extended variant of
456 Test::More called OpenSSL::Test to do its work. All test scripts in
457 test/ have been rewritten into test recipes, and all direct calls to
458 executables in test/Makefile have become individual recipes using the
459 simplified testing OpenSSL::Test::Simple.
461 For documentation on our testing modules, do:
463 perldoc test/testlib/OpenSSL/Test/Simple.pm
464 perldoc test/testlib/OpenSSL/Test.pm
468 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
469 are used; the latter aborts on memory leaks (usually checked on exit).
470 Some undocumented "set malloc, etc., hooks" functions were removed
471 and others were changed. All are now documented.
474 *) In DSA_generate_parameters_ex, if the provided seed is too short,
476 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
478 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
479 from RFC4279, RFC4785, RFC5487, RFC5489.
481 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
482 original RSA_PSK patch.
485 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
486 era flag was never set throughout the codebase (only read). Also removed
487 SSL3_FLAGS_POP_BUFFER which was only used if
488 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
491 *) Changed the default name options in the "ca", "crl", "req" and "x509"
492 to be "oneline" instead of "compat".
495 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
496 not aware of clients that still exhibit this bug, and the workaround
497 hasn't been working properly for a while.
500 *) The return type of BIO_number_read() and BIO_number_written() as well as
501 the corresponding num_read and num_write members in the BIO structure has
502 changed from unsigned long to uint64_t. On platforms where an unsigned
503 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
507 *) Given the pervasive nature of TLS extensions it is inadvisable to run
508 OpenSSL without support for them. It also means that maintaining
509 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
510 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
513 *) Removed support for the two export grade static DH ciphersuites
514 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
515 were newly added (along with a number of other static DH ciphersuites) to
516 1.0.2. However the two export ones have *never* worked since they were
517 introduced. It seems strange in any case to be adding new export
518 ciphersuites, and given "logjam" it also does not seem correct to fix them.
521 *) Version negotiation has been rewritten. In particular SSLv23_method(),
522 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
523 and turned into macros which simply call the new preferred function names
524 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
525 should use the new names instead. Also as part of this change the ssl23.h
526 header file has been removed.
529 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
530 code and the associated standard is no longer considered fit-for-purpose.
533 *) RT2547 was closed. When generating a private key, try to make the
534 output file readable only by the owner. This behavior change might
535 be noticeable when interacting with other software.
537 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
541 *) Added HTTP GET support to the ocsp command.
544 *) Changed default digest for the dgst and enc commands from MD5 to
548 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
551 *) Added support for TLS extended master secret from
552 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
553 initial patch which was a great help during development.
556 *) All libssl internal structures have been removed from the public header
557 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
558 now redundant). Users should not attempt to access internal structures
559 directly. Instead they should use the provided API functions.
562 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
563 Access to deprecated functions can be re-enabled by running config with
564 "enable-deprecated". In addition applications wishing to use deprecated
565 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
566 will, by default, disable some transitive includes that previously existed
567 in the header files (e.g. ec.h will no longer, by default, include bn.h)
570 *) Added support for OCB mode. OpenSSL has been granted a patent license
571 compatible with the OpenSSL license for use of OCB. Details are available
572 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
573 for OCB can be removed by calling config with no-ocb.
576 *) SSLv2 support has been removed. It still supports receiving a SSLv2
577 compatible client hello.
580 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
581 done while fixing the error code for the key-too-small case.
582 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
584 *) CA.sh has been removmed; use CA.pl instead.
587 *) Removed old DES API.
590 *) Remove various unsupported platforms:
596 Sinix/ReliantUNIX RM400
601 16-bit platforms such as WIN16
604 *) Clean up OPENSSL_NO_xxx #define's
605 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
606 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
607 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
608 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
609 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
610 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
611 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
612 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
613 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
614 Remove MS_STATIC; it's a relic from platforms <32 bits.
617 *) Cleaned up dead code
618 Remove all but one '#ifdef undef' which is to be looked at.
621 *) Clean up calling of xxx_free routines.
622 Just like free(), fix most of the xxx_free routines to accept
623 NULL. Remove the non-null checks from callers. Save much code.
626 *) Add secure heap for storage of private keys (when possible).
627 Add BIO_s_secmem(), CBIGNUM, etc.
628 Contributed by Akamai Technologies under our Corporate CLA.
631 *) Experimental support for a new, fast, unbiased prime candidate generator,
632 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
633 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
635 *) New output format NSS in the sess_id command line tool. This allows
636 exporting the session id and the master key in NSS keylog format.
637 [Martin Kaiser <martin@kaiser.cx>]
639 *) Harmonize version and its documentation. -f flag is used to display
641 [mancha <mancha1@zoho.com>]
643 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
644 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
645 [mancha <mancha1@zoho.com>]
647 *) Fix some double frees. These are not thought to be exploitable.
648 [mancha <mancha1@zoho.com>]
650 *) A missing bounds check in the handling of the TLS heartbeat extension
651 can be used to reveal up to 64k of memory to a connected client or
654 Thanks for Neel Mehta of Google Security for discovering this bug and to
655 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
656 preparing the fix (CVE-2014-0160)
657 [Adam Langley, Bodo Moeller]
659 *) Fix for the attack described in the paper "Recovering OpenSSL
660 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
661 by Yuval Yarom and Naomi Benger. Details can be obtained from:
662 http://eprint.iacr.org/2014/140
664 Thanks to Yuval Yarom and Naomi Benger for discovering this
665 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
666 [Yuval Yarom and Naomi Benger]
668 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
669 this fixes a limitation in previous versions of OpenSSL.
672 *) Experimental encrypt-then-mac support.
674 Experimental support for encrypt then mac from
675 draft-gutmann-tls-encrypt-then-mac-02.txt
677 To enable it set the appropriate extension number (0x42 for the test
678 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
680 For non-compliant peers (i.e. just about everything) this should have no
683 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
687 *) Add EVP support for key wrapping algorithms, to avoid problems with
688 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
689 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
690 algorithms and include tests cases.
693 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
697 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
698 MGF1 digest and OAEP label.
701 *) Make openssl verify return errors.
702 [Chris Palmer <palmer@google.com> and Ben Laurie]
704 *) New function ASN1_TIME_diff to calculate the difference between two
705 ASN1_TIME structures or one structure and the current time.
708 *) Update fips_test_suite to support multiple command line options. New
709 test to induce all self test errors in sequence and check expected
713 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
714 sign or verify all in one operation.
717 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
718 test programs and fips_test_suite. Includes functionality to parse
719 the minimal script output of fipsalgest.pl directly.
722 *) Add authorisation parameter to FIPS_module_mode_set().
725 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
728 *) Use separate DRBG fields for internal and external flags. New function
729 FIPS_drbg_health_check() to perform on demand health checking. Add
730 generation tests to fips_test_suite with reduced health check interval to
731 demonstrate periodic health checking. Add "nodh" option to
732 fips_test_suite to skip very slow DH test.
735 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
739 *) More extensive health check for DRBG checking many more failure modes.
740 New function FIPS_selftest_drbg_all() to handle every possible DRBG
741 combination: call this in fips_test_suite.
744 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
745 and POST to handle Dual EC cases.
748 *) Add support for canonical generation of DSA parameter 'g'. See
751 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
752 POST to handle HMAC cases.
755 *) Add functions FIPS_module_version() and FIPS_module_version_text()
756 to return numerical and string versions of the FIPS module number.
759 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
760 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
761 outside the validated module in the FIPS capable OpenSSL.
764 *) Minor change to DRBG entropy callback semantics. In some cases
765 there is no multiple of the block length between min_len and
766 max_len. Allow the callback to return more than max_len bytes
767 of entropy but discard any extra: it is the callback's responsibility
768 to ensure that the extra data discarded does not impact the
769 requested amount of entropy.
772 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
773 information in FIPS186-3, SP800-57 and SP800-131A.
776 *) CCM support via EVP. Interface is very similar to GCM case except we
777 must supply all data in one chunk (i.e. no update, final) and the
778 message length must be supplied if AAD is used. Add algorithm test
782 *) Initial version of POST overhaul. Add POST callback to allow the status
783 of POST to be monitored and/or failures induced. Modify fips_test_suite
784 to use callback. Always run all selftests even if one fails.
787 *) XTS support including algorithm test driver in the fips_gcmtest program.
788 Note: this does increase the maximum key length from 32 to 64 bytes but
789 there should be no binary compatibility issues as existing applications
790 will never use XTS mode.
793 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
794 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
795 performs algorithm blocking for unapproved PRNG types. Also do not
796 set PRNG type in FIPS_mode_set(): leave this to the application.
797 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
798 the standard OpenSSL PRNG: set additional data to a date time vector.
801 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
802 This shouldn't present any incompatibility problems because applications
803 shouldn't be using these directly and any that are will need to rethink
804 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
807 *) Extensive self tests and health checking required by SP800-90 DRBG.
808 Remove strength parameter from FIPS_drbg_instantiate and always
809 instantiate at maximum supported strength.
812 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
815 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
818 *) New function DH_compute_key_padded() to compute a DH key and pad with
819 leading zeroes if needed: this complies with SP800-56A et al.
822 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
823 anything, incomplete, subject to change and largely untested at present.
826 *) Modify fipscanisteronly build option to only build the necessary object
827 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
830 *) Add experimental option FIPSSYMS to give all symbols in
831 fipscanister.o and FIPS or fips prefix. This will avoid
832 conflicts with future versions of OpenSSL. Add perl script
833 util/fipsas.pl to preprocess assembly language source files
834 and rename any affected symbols.
837 *) Add selftest checks and algorithm block of non-fips algorithms in
838 FIPS mode. Remove DES2 from selftests.
841 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
842 return internal method without any ENGINE dependencies. Add new
843 tiny fips sign and verify functions.
846 *) New build option no-ec2m to disable characteristic 2 code.
849 *) New build option "fipscanisteronly". This only builds fipscanister.o
850 and (currently) associated fips utilities. Uses the file Makefile.fips
851 instead of Makefile.org as the prototype.
854 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
855 Update fips_gcmtest to use IV generator.
858 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
859 setting output buffer to NULL. The *Final function must be
860 called although it will not retrieve any additional data. The tag
861 can be set or retrieved with a ctrl. The IV length is by default 12
862 bytes (96 bits) but can be set to an alternative value. If the IV
863 length exceeds the maximum IV length (currently 16 bytes) it cannot be
867 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
868 underlying do_cipher function handles all cipher semantics itself
869 including padding and finalisation. This is useful if (for example)
870 an ENGINE cipher handles block padding itself. The behaviour of
871 do_cipher is subtly changed if this flag is set: the return value
872 is the number of characters written to the output buffer (zero is
873 no longer an error code) or a negative error code. Also if the
874 input buffer is NULL and length 0 finalisation should be performed.
877 *) If a candidate issuer certificate is already part of the constructed
878 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
881 *) Improve forward-security support: add functions
883 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
884 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
886 for use by SSL/TLS servers; the callback function will be called whenever a
887 new session is created, and gets to decide whether the session may be
888 cached to make it resumable (return 0) or not (return 1). (As by the
889 SSL/TLS protocol specifications, the session_id sent by the server will be
890 empty to indicate that the session is not resumable; also, the server will
891 not generate RFC 4507 (RFC 5077) session tickets.)
893 A simple reasonable callback implementation is to return is_forward_secure.
894 This parameter will be set to 1 or 0 depending on the ciphersuite selected
895 by the SSL/TLS server library, indicating whether it can provide forward
897 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
899 *) New -verify_name option in command line utilities to set verification
903 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
904 Add CMAC pkey methods.
907 *) Experimental renegotiation in s_server -www mode. If the client
908 browses /reneg connection is renegotiated. If /renegcert it is
909 renegotiated requesting a certificate.
912 *) Add an "external" session cache for debugging purposes to s_server. This
913 should help trace issues which normally are only apparent in deployed
914 multi-process servers.
917 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
918 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
919 BIO_set_cipher() and some obscure PEM functions were changed so they
920 can now return an error. The RAND changes required a change to the
921 RAND_METHOD structure.
924 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
925 a gcc attribute to warn if the result of a function is ignored. This
926 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
927 whose return value is often ignored.
930 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
931 These allow SCTs (signed certificate timestamps) to be requested and
932 validated when establishing a connection.
933 [Rob Percival <robpercival@google.com>]
935 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
937 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
938 Builds that are not configured with "enable-weak-ssl-ciphers" will not
939 provide any "EXPORT" or "LOW" strength ciphers.
942 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
943 is by default disabled at build-time. Builds that are not configured with
944 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
945 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
946 will need to explicitly call either of:
948 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
950 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
952 as appropriate. Even if either of those is used, or the application
953 explicitly uses the version-specific SSLv2_method() or its client and
954 server variants, SSLv2 ciphers vulnerable to exhaustive search key
955 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
956 ciphers, and SSLv2 56-bit DES are no longer available.
960 *) Fix a double-free in DSA code
962 A double free bug was discovered when OpenSSL parses malformed DSA private
963 keys and could lead to a DoS attack or memory corruption for applications
964 that receive DSA private keys from untrusted sources. This scenario is
967 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
972 *) Disable SRP fake user seed to address a server memory leak.
974 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
976 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
977 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
978 was changed to ignore the "fake user" SRP seed, even if the seed
981 Users should use SRP_VBASE_get1_by_user instead. Note that in
982 SRP_VBASE_get1_by_user, caller must free the returned value. Note
983 also that even though configuring the SRP seed attempts to hide
984 invalid usernames by continuing the handshake with fake
985 credentials, this behaviour is not constant time and no strong
986 guarantees are made that the handshake is indistinguishable from
987 that of a valid user.
991 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
993 In the BN_hex2bn function the number of hex digits is calculated using an
994 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
995 large values of |i| this can result in |bn_expand| not allocating any
996 memory because |i * 4| is negative. This can leave the internal BIGNUM data
997 field as NULL leading to a subsequent NULL ptr deref. For very large values
998 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
999 In this case memory is allocated to the internal BIGNUM data field, but it
1000 is insufficiently sized leading to heap corruption. A similar issue exists
1001 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1002 is ever called by user applications with very large untrusted hex/dec data.
1003 This is anticipated to be a rare occurrence.
1005 All OpenSSL internal usage of these functions use data that is not expected
1006 to be untrusted, e.g. config file data or application command line
1007 arguments. If user developed applications generate config file data based
1008 on untrusted data then it is possible that this could also lead to security
1009 consequences. This is also anticipated to be rare.
1011 This issue was reported to OpenSSL by Guido Vranken.
1015 *) Fix memory issues in BIO_*printf functions
1017 The internal |fmtstr| function used in processing a "%s" format string in
1018 the BIO_*printf functions could overflow while calculating the length of a
1019 string and cause an OOB read when printing very long strings.
1021 Additionally the internal |doapr_outch| function can attempt to write to an
1022 OOB memory location (at an offset from the NULL pointer) in the event of a
1023 memory allocation failure. In 1.0.2 and below this could be caused where
1024 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1025 could be in processing a very long "%s" format string. Memory leaks can
1028 The first issue may mask the second issue dependent on compiler behaviour.
1029 These problems could enable attacks where large amounts of untrusted data
1030 is passed to the BIO_*printf functions. If applications use these functions
1031 in this way then they could be vulnerable. OpenSSL itself uses these
1032 functions when printing out human-readable dumps of ASN.1 data. Therefore
1033 applications that print this data could be vulnerable if the data is from
1034 untrusted sources. OpenSSL command line applications could also be
1035 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1036 as command line arguments.
1038 Libssl is not considered directly vulnerable. Additionally certificates etc
1039 received via remote connections via libssl are also unlikely to be able to
1040 trigger these issues because of message size limits enforced within libssl.
1042 This issue was reported to OpenSSL Guido Vranken.
1046 *) Side channel attack on modular exponentiation
1048 A side-channel attack was found which makes use of cache-bank conflicts on
1049 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1050 of RSA keys. The ability to exploit this issue is limited as it relies on
1051 an attacker who has control of code in a thread running on the same
1052 hyper-threaded core as the victim thread which is performing decryptions.
1054 This issue was reported to OpenSSL by Yuval Yarom, The University of
1055 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1056 Nadia Heninger, University of Pennsylvania with more information at
1057 http://cachebleed.info.
1061 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1062 if no keysize is specified with default_bits. This fixes an
1063 omission in an earlier change that changed all RSA/DSA key generation
1064 apps to use 2048 bits by default.
1067 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1068 *) DH small subgroups
1070 Historically OpenSSL only ever generated DH parameters based on "safe"
1071 primes. More recently (in version 1.0.2) support was provided for
1072 generating X9.42 style parameter files such as those required for RFC 5114
1073 support. The primes used in such files may not be "safe". Where an
1074 application is using DH configured with parameters based on primes that are
1075 not "safe" then an attacker could use this fact to find a peer's private
1076 DH exponent. This attack requires that the attacker complete multiple
1077 handshakes in which the peer uses the same private DH exponent. For example
1078 this could be used to discover a TLS server's private DH exponent if it's
1079 reusing the private DH exponent or it's using a static DH ciphersuite.
1081 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1082 TLS. It is not on by default. If the option is not set then the server
1083 reuses the same private DH exponent for the life of the server process and
1084 would be vulnerable to this attack. It is believed that many popular
1085 applications do set this option and would therefore not be at risk.
1087 The fix for this issue adds an additional check where a "q" parameter is
1088 available (as is the case in X9.42 based parameters). This detects the
1089 only known attack, and is the only possible defense for static DH
1090 ciphersuites. This could have some performance impact.
1092 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1093 default and cannot be disabled. This could have some performance impact.
1095 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1099 *) SSLv2 doesn't block disabled ciphers
1101 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1102 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1103 been disabled, provided that the SSLv2 protocol was not also disabled via
1106 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1107 and Sebastian Schinzel.
1111 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1113 *) BN_mod_exp may produce incorrect results on x86_64
1115 There is a carry propagating bug in the x86_64 Montgomery squaring
1116 procedure. No EC algorithms are affected. Analysis suggests that attacks
1117 against RSA and DSA as a result of this defect would be very difficult to
1118 perform and are not believed likely. Attacks against DH are considered just
1119 feasible (although very difficult) because most of the work necessary to
1120 deduce information about a private key may be performed offline. The amount
1121 of resources required for such an attack would be very significant and
1122 likely only accessible to a limited number of attackers. An attacker would
1123 additionally need online access to an unpatched system using the target
1124 private key in a scenario with persistent DH parameters and a private
1125 key that is shared between multiple clients. For example this can occur by
1126 default in OpenSSL DHE based SSL/TLS ciphersuites.
1128 This issue was reported to OpenSSL by Hanno Böck.
1132 *) Certificate verify crash with missing PSS parameter
1134 The signature verification routines will crash with a NULL pointer
1135 dereference if presented with an ASN.1 signature using the RSA PSS
1136 algorithm and absent mask generation function parameter. Since these
1137 routines are used to verify certificate signature algorithms this can be
1138 used to crash any certificate verification operation and exploited in a
1139 DoS attack. Any application which performs certificate verification is
1140 vulnerable including OpenSSL clients and servers which enable client
1143 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1147 *) X509_ATTRIBUTE memory leak
1149 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1150 memory. This structure is used by the PKCS#7 and CMS routines so any
1151 application which reads PKCS#7 or CMS data from untrusted sources is
1152 affected. SSL/TLS is not affected.
1154 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1159 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1160 This changes the decoding behaviour for some invalid messages,
1161 though the change is mostly in the more lenient direction, and
1162 legacy behaviour is preserved as much as possible.
1165 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1167 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1169 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1171 *) Alternate chains certificate forgery
1173 During certificate verfification, OpenSSL will attempt to find an
1174 alternative certificate chain if the first attempt to build such a chain
1175 fails. An error in the implementation of this logic can mean that an
1176 attacker could cause certain checks on untrusted certificates to be
1177 bypassed, such as the CA flag, enabling them to use a valid leaf
1178 certificate to act as a CA and "issue" an invalid certificate.
1180 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1184 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1186 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1187 incompatibility in the handling of HMAC. The previous ABI has now been
1191 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1193 *) Malformed ECParameters causes infinite loop
1195 When processing an ECParameters structure OpenSSL enters an infinite loop
1196 if the curve specified is over a specially malformed binary polynomial
1199 This can be used to perform denial of service against any
1200 system which processes public keys, certificate requests or
1201 certificates. This includes TLS clients and TLS servers with
1202 client authentication enabled.
1204 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1208 *) Exploitable out-of-bounds read in X509_cmp_time
1210 X509_cmp_time does not properly check the length of the ASN1_TIME
1211 string and can read a few bytes out of bounds. In addition,
1212 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1215 An attacker can use this to craft malformed certificates and CRLs of
1216 various sizes and potentially cause a segmentation fault, resulting in
1217 a DoS on applications that verify certificates or CRLs. TLS clients
1218 that verify CRLs are affected. TLS clients and servers with client
1219 authentication enabled may be affected if they use custom verification
1222 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1223 independently by Hanno Böck.
1227 *) PKCS7 crash with missing EnvelopedContent
1229 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1230 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1231 with missing content and trigger a NULL pointer dereference on parsing.
1233 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1234 structures from untrusted sources are affected. OpenSSL clients and
1235 servers are not affected.
1237 This issue was reported to OpenSSL by Michal Zalewski (Google).
1241 *) CMS verify infinite loop with unknown hash function
1243 When verifying a signedData message the CMS code can enter an infinite loop
1244 if presented with an unknown hash function OID. This can be used to perform
1245 denial of service against any system which verifies signedData messages using
1247 This issue was reported to OpenSSL by Johannes Bauer.
1251 *) Race condition handling NewSessionTicket
1253 If a NewSessionTicket is received by a multi-threaded client when attempting to
1254 reuse a previous ticket then a race condition can occur potentially leading to
1255 a double free of the ticket data.
1259 *) Only support 256-bit or stronger elliptic curves with the
1260 'ecdh_auto' setting (server) or by default (client). Of supported
1261 curves, prefer P-256 (both).
1264 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1266 *) ClientHello sigalgs DoS fix
1268 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1269 invalid signature algorithms extension a NULL pointer dereference will
1270 occur. This can be exploited in a DoS attack against the server.
1272 This issue was was reported to OpenSSL by David Ramos of Stanford
1275 [Stephen Henson and Matt Caswell]
1277 *) Multiblock corrupted pointer fix
1279 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1280 feature only applies on 64 bit x86 architecture platforms that support AES
1281 NI instructions. A defect in the implementation of "multiblock" can cause
1282 OpenSSL's internal write buffer to become incorrectly set to NULL when
1283 using non-blocking IO. Typically, when the user application is using a
1284 socket BIO for writing, this will only result in a failed connection.
1285 However if some other BIO is used then it is likely that a segmentation
1286 fault will be triggered, thus enabling a potential DoS attack.
1288 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1292 *) Segmentation fault in DTLSv1_listen fix
1294 The DTLSv1_listen function is intended to be stateless and processes the
1295 initial ClientHello from many peers. It is common for user code to loop
1296 over the call to DTLSv1_listen until a valid ClientHello is received with
1297 an associated cookie. A defect in the implementation of DTLSv1_listen means
1298 that state is preserved in the SSL object from one invocation to the next
1299 that can lead to a segmentation fault. Errors processing the initial
1300 ClientHello can trigger this scenario. An example of such an error could be
1301 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1304 This issue was reported to OpenSSL by Per Allansson.
1308 *) Segmentation fault in ASN1_TYPE_cmp fix
1310 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1311 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1312 certificate signature algorithm consistency this can be used to crash any
1313 certificate verification operation and exploited in a DoS attack. Any
1314 application which performs certificate verification is vulnerable including
1315 OpenSSL clients and servers which enable client authentication.
1319 *) Segmentation fault for invalid PSS parameters fix
1321 The signature verification routines will crash with a NULL pointer
1322 dereference if presented with an ASN.1 signature using the RSA PSS
1323 algorithm and invalid parameters. Since these routines are used to verify
1324 certificate signature algorithms this can be used to crash any
1325 certificate verification operation and exploited in a DoS attack. Any
1326 application which performs certificate verification is vulnerable including
1327 OpenSSL clients and servers which enable client authentication.
1329 This issue was was reported to OpenSSL by Brian Carpenter.
1333 *) ASN.1 structure reuse memory corruption fix
1335 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1336 memory corruption via an invalid write. Such reuse is and has been
1337 strongly discouraged and is believed to be rare.
1339 Applications that parse structures containing CHOICE or ANY DEFINED BY
1340 components may be affected. Certificate parsing (d2i_X509 and related
1341 functions) are however not affected. OpenSSL clients and servers are
1346 *) PKCS7 NULL pointer dereferences fix
1348 The PKCS#7 parsing code does not handle missing outer ContentInfo
1349 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1350 missing content and trigger a NULL pointer dereference on parsing.
1352 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1353 otherwise parse PKCS#7 structures from untrusted sources are
1354 affected. OpenSSL clients and servers are not affected.
1356 This issue was reported to OpenSSL by Michal Zalewski (Google).
1360 *) DoS via reachable assert in SSLv2 servers fix
1362 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1363 servers that both support SSLv2 and enable export cipher suites by sending
1364 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1366 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1367 (OpenSSL development team).
1371 *) Empty CKE with client auth and DHE fix
1373 If client auth is used then a server can seg fault in the event of a DHE
1374 ciphersuite being selected and a zero length ClientKeyExchange message
1375 being sent by the client. This could be exploited in a DoS attack.
1379 *) Handshake with unseeded PRNG fix
1381 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1382 with an unseeded PRNG. The conditions are:
1383 - The client is on a platform where the PRNG has not been seeded
1384 automatically, and the user has not seeded manually
1385 - A protocol specific client method version has been used (i.e. not
1386 SSL_client_methodv23)
1387 - A ciphersuite is used that does not require additional random data from
1388 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1390 If the handshake succeeds then the client random that has been used will
1391 have been generated from a PRNG with insufficient entropy and therefore the
1392 output may be predictable.
1394 For example using the following command with an unseeded openssl will
1395 succeed on an unpatched platform:
1397 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1401 *) Use After Free following d2i_ECPrivatekey error fix
1403 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1404 could cause a use after free condition. This, in turn, could cause a double
1405 free in several private key parsing functions (such as d2i_PrivateKey
1406 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1407 for applications that receive EC private keys from untrusted
1408 sources. This scenario is considered rare.
1410 This issue was discovered by the BoringSSL project and fixed in their
1415 *) X509_to_X509_REQ NULL pointer deref fix
1417 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1418 the certificate key is invalid. This function is rarely used in practice.
1420 This issue was discovered by Brian Carpenter.
1424 *) Removed the export ciphers from the DEFAULT ciphers
1427 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1429 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1430 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1431 So far those who have to target multiple plaforms would compromise
1432 and argue that binary targeting say ARMv5 would still execute on
1433 ARMv8. "Universal" build resolves this compromise by providing
1434 near-optimal performance even on newer platforms.
1437 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1438 (other platforms pending).
1439 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1441 *) Add support for the SignedCertificateTimestampList certificate and
1442 OCSP response extensions from RFC6962.
1445 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1446 for corner cases. (Certain input points at infinity could lead to
1447 bogus results, with non-infinity inputs mapped to infinity too.)
1450 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1451 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1452 common cases are optimized and there still is room for further
1453 improvements. Vector Permutation AES for Altivec is also added.
1456 *) Add support for little-endian ppc64 Linux target.
1457 [Marcelo Cerri (IBM)]
1459 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1460 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1461 are optimized and there still is room for further improvements.
1462 Both 32- and 64-bit modes are supported.
1463 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1465 *) Improved ARMv7 NEON support.
1468 *) Support for SPARC Architecture 2011 crypto extensions, first
1469 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1470 SHA256/512, MD5, GHASH and modular exponentiation.
1471 [Andy Polyakov, David Miller]
1473 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1475 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1477 *) Support for new and upcoming Intel processors, including AVX2,
1478 BMI and SHA ISA extensions. This includes additional "stitched"
1479 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1482 This work was sponsored by Intel Corp.
1485 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1486 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1487 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1490 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1491 this fixes a limiation in previous versions of OpenSSL.
1494 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1495 MGF1 digest and OAEP label.
1498 *) Add EVP support for key wrapping algorithms, to avoid problems with
1499 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1500 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1501 algorithms and include tests cases.
1504 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1506 [Douglas E. Engert, Steve Henson]
1508 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1509 difference in days and seconds between two tm or ASN1_TIME structures.
1512 *) Add -rev test option to s_server to just reverse order of characters
1513 received by client and send back to server. Also prints an abbreviated
1514 summary of the connection parameters.
1517 *) New option -brief for s_client and s_server to print out a brief summary
1518 of connection parameters.
1521 *) Add callbacks for arbitrary TLS extensions.
1522 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1524 *) New option -crl_download in several openssl utilities to download CRLs
1525 from CRLDP extension in certificates.
1528 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1531 *) New function X509_CRL_diff to generate a delta CRL from the difference
1532 of two full CRLs. Add support to "crl" utility.
1535 *) New functions to set lookup_crls function and to retrieve
1536 X509_STORE from X509_STORE_CTX.
1539 *) Print out deprecated issuer and subject unique ID fields in
1543 *) Extend OCSP I/O functions so they can be used for simple general purpose
1544 HTTP as well as OCSP. New wrapper function which can be used to download
1545 CRLs using the OCSP API.
1548 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1551 *) SSL_CONF* functions. These provide a common framework for application
1552 configuration using configuration files or command lines.
1555 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1556 message callback and prints the results. Needs compile time option
1557 "enable-ssl-trace". New options to s_client and s_server to enable
1561 *) New ctrl and macro to retrieve supported points extensions.
1562 Print out extension in s_server and s_client.
1565 *) New functions to retrieve certificate signature and signature
1569 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1573 *) New Suite B modes for TLS code. These use and enforce the requirements
1574 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1575 only use Suite B curves. The Suite B modes can be set by using the
1576 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1579 *) New chain verification flags for Suite B levels of security. Check
1580 algorithms are acceptable when flags are set in X509_verify_cert.
1583 *) Make tls1_check_chain return a set of flags indicating checks passed
1584 by a certificate chain. Add additional tests to handle client
1585 certificates: checks for matching certificate type and issuer name
1589 *) If an attempt is made to use a signature algorithm not in the peer
1590 preference list abort the handshake. If client has no suitable
1591 signature algorithms in response to a certificate request do not
1592 use the certificate.
1595 *) If server EC tmp key is not in client preference list abort handshake.
1598 *) Add support for certificate stores in CERT structure. This makes it
1599 possible to have different stores per SSL structure or one store in
1600 the parent SSL_CTX. Include distint stores for certificate chain
1601 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1602 to build and store a certificate chain in CERT structure: returing
1603 an error if the chain cannot be built: this will allow applications
1604 to test if a chain is correctly configured.
1606 Note: if the CERT based stores are not set then the parent SSL_CTX
1607 store is used to retain compatibility with existing behaviour.
1611 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1612 mask based on the current session, check mask when sending client
1613 hello and checking the requested ciphersuite.
1616 *) New ctrls to retrieve and set certificate types in a certificate
1617 request message. Print out received values in s_client. If certificate
1618 types is not set with custom values set sensible values based on
1619 supported signature algorithms.
1622 *) Support for distinct client and server supported signature algorithms.
1625 *) Add certificate callback. If set this is called whenever a certificate
1626 is required by client or server. An application can decide which
1627 certificate chain to present based on arbitrary criteria: for example
1628 supported signature algorithms. Add very simple example to s_server.
1629 This fixes many of the problems and restrictions of the existing client
1630 certificate callback: for example you can now clear an existing
1631 certificate and specify the whole chain.
1634 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1635 the certificate can be used for (if anything). Set valid_flags field
1636 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1637 to have similar checks in it.
1639 Add new "cert_flags" field to CERT structure and include a "strict mode".
1640 This enforces some TLS certificate requirements (such as only permitting
1641 certificate signature algorithms contained in the supported algorithms
1642 extension) which some implementations ignore: this option should be used
1643 with caution as it could cause interoperability issues.
1646 *) Update and tidy signature algorithm extension processing. Work out
1647 shared signature algorithms based on preferences and peer algorithms
1648 and print them out in s_client and s_server. Abort handshake if no
1649 shared signature algorithms.
1652 *) Add new functions to allow customised supported signature algorithms
1653 for SSL and SSL_CTX structures. Add options to s_client and s_server
1657 *) New function SSL_certs_clear() to delete all references to certificates
1658 from an SSL structure. Before this once a certificate had been added
1659 it couldn't be removed.
1662 *) Integrate hostname, email address and IP address checking with certificate
1663 verification. New verify options supporting checking in opensl utility.
1666 *) Fixes and wildcard matching support to hostname and email checking
1667 functions. Add manual page.
1668 [Florian Weimer (Red Hat Product Security Team)]
1670 *) New functions to check a hostname email or IP address against a
1671 certificate. Add options x509 utility to print results of checks against
1675 *) Fix OCSP checking.
1676 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1678 *) Initial experimental support for explicitly trusted non-root CAs.
1679 OpenSSL still tries to build a complete chain to a root but if an
1680 intermediate CA has a trust setting included that is used. The first
1681 setting is used: whether to trust (e.g., -addtrust option to the x509
1685 *) Add -trusted_first option which attempts to find certificates in the
1686 trusted store even if an untrusted chain is also supplied.
1689 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1690 platform support for Linux and Android.
1693 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1696 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1697 When in FIPS mode the approved implementations are used as normal,
1698 when not in FIPS mode the internal unapproved versions are used instead.
1699 This means that the FIPS capable OpenSSL isn't forced to use the
1700 (often lower performance) FIPS implementations outside FIPS mode.
1703 *) Transparently support X9.42 DH parameters when calling
1704 PEM_read_bio_DHparameters. This means existing applications can handle
1705 the new parameter format automatically.
1708 *) Initial experimental support for X9.42 DH parameter format: mainly
1709 to support use of 'q' parameter for RFC5114 parameters.
1712 *) Add DH parameters from RFC5114 including test data to dhtest.
1715 *) Support for automatic EC temporary key parameter selection. If enabled
1716 the most preferred EC parameters are automatically used instead of
1717 hardcoded fixed parameters. Now a server just has to call:
1718 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1719 support ECDH and use the most appropriate parameters.
1722 *) Enhance and tidy EC curve and point format TLS extension code. Use
1723 static structures instead of allocation if default values are used.
1724 New ctrls to set curves we wish to support and to retrieve shared curves.
1725 Print out shared curves in s_server. New options to s_server and s_client
1726 to set list of supported curves.
1729 *) New ctrls to retrieve supported signature algorithms and
1730 supported curve values as an array of NIDs. Extend openssl utility
1731 to print out received values.
1734 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1735 between NIDs and the more common NIST names such as "P-256". Enhance
1736 ecparam utility and ECC method to recognise the NIST names for curves.
1739 *) Enhance SSL/TLS certificate chain handling to support different
1740 chains for each certificate instead of one chain in the parent SSL_CTX.
1743 *) Support for fixed DH ciphersuite client authentication: where both
1744 server and client use DH certificates with common parameters.
1747 *) Support for fixed DH ciphersuites: those requiring DH server
1751 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1753 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1754 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1755 X509_CINF_get_signature were reverted post internal team review.
1757 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1759 *) Build fixes for the Windows and OpenVMS platforms
1760 [Matt Caswell and Richard Levitte]
1762 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1764 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1765 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1766 dereference. This could lead to a Denial Of Service attack. Thanks to
1767 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1771 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1772 dtls1_buffer_record function under certain conditions. In particular this
1773 could occur if an attacker sent repeated DTLS records with the same
1774 sequence number but for the next epoch. The memory leak could be exploited
1775 by an attacker in a Denial of Service attack through memory exhaustion.
1776 Thanks to Chris Mueller for reporting this issue.
1780 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1781 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1782 method would be set to NULL which could later result in a NULL pointer
1783 dereference. Thanks to Frank Schmirler for reporting this issue.
1787 *) Abort handshake if server key exchange message is omitted for ephemeral
1790 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1791 reporting this issue.
1795 *) Remove non-export ephemeral RSA code on client and server. This code
1796 violated the TLS standard by allowing the use of temporary RSA keys in
1797 non-export ciphersuites and could be used by a server to effectively
1798 downgrade the RSA key length used to a value smaller than the server
1799 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1800 INRIA or reporting this issue.
1804 *) Fixed issue where DH client certificates are accepted without verification.
1805 An OpenSSL server will accept a DH certificate for client authentication
1806 without the certificate verify message. This effectively allows a client to
1807 authenticate without the use of a private key. This only affects servers
1808 which trust a client certificate authority which issues certificates
1809 containing DH keys: these are extremely rare and hardly ever encountered.
1810 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1815 *) Ensure that the session ID context of an SSL is updated when its
1816 SSL_CTX is updated via SSL_set_SSL_CTX.
1818 The session ID context is typically set from the parent SSL_CTX,
1819 and can vary with the CTX.
1822 *) Fix various certificate fingerprint issues.
1824 By using non-DER or invalid encodings outside the signed portion of a
1825 certificate the fingerprint can be changed without breaking the signature.
1826 Although no details of the signed portion of the certificate can be changed
1827 this can cause problems with some applications: e.g. those using the
1828 certificate fingerprint for blacklists.
1830 1. Reject signatures with non zero unused bits.
1832 If the BIT STRING containing the signature has non zero unused bits reject
1833 the signature. All current signature algorithms require zero unused bits.
1835 2. Check certificate algorithm consistency.
1837 Check the AlgorithmIdentifier inside TBS matches the one in the
1838 certificate signature. NB: this will result in signature failure
1839 errors for some broken certificates.
1841 Thanks to Konrad Kraszewski from Google for reporting this issue.
1843 3. Check DSA/ECDSA signatures use DER.
1845 Reencode DSA/ECDSA signatures and compare with the original received
1846 signature. Return an error if there is a mismatch.
1848 This will reject various cases including garbage after signature
1849 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1850 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1851 (negative or with leading zeroes).
1853 Further analysis was conducted and fixes were developed by Stephen Henson
1854 of the OpenSSL core team.
1859 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1860 results on some platforms, including x86_64. This bug occurs at random
1861 with a very low probability, and is not known to be exploitable in any
1862 way, though its exact impact is difficult to determine. Thanks to Pieter
1863 Wuille (Blockstream) who reported this issue and also suggested an initial
1864 fix. Further analysis was conducted by the OpenSSL development team and
1865 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1866 the OpenSSL core team.
1870 *) Do not resume sessions on the server if the negotiated protocol
1871 version does not match the session's version. Resuming with a different
1872 version, while not strictly forbidden by the RFC, is of questionable
1873 sanity and breaks all known clients.
1874 [David Benjamin, Emilia Käsper]
1876 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1877 early CCS messages during renegotiation. (Note that because
1878 renegotiation is encrypted, this early CCS was not exploitable.)
1881 *) Tighten client-side session ticket handling during renegotiation:
1882 ensure that the client only accepts a session ticket if the server sends
1883 the extension anew in the ServerHello. Previously, a TLS client would
1884 reuse the old extension state and thus accept a session ticket if one was
1885 announced in the initial ServerHello.
1887 Similarly, ensure that the client requires a session ticket if one
1888 was advertised in the ServerHello. Previously, a TLS client would
1889 ignore a missing NewSessionTicket message.
1892 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1894 *) SRTP Memory Leak.
1896 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1897 sends a carefully crafted handshake message, to cause OpenSSL to fail
1898 to free up to 64k of memory causing a memory leak. This could be
1899 exploited in a Denial Of Service attack. This issue affects OpenSSL
1900 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1901 whether SRTP is used or configured. Implementations of OpenSSL that
1902 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1904 The fix was developed by the OpenSSL team.
1908 *) Session Ticket Memory Leak.
1910 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1911 integrity of that ticket is first verified. In the event of a session
1912 ticket integrity check failing, OpenSSL will fail to free memory
1913 causing a memory leak. By sending a large number of invalid session
1914 tickets an attacker could exploit this issue in a Denial Of Service
1919 *) Build option no-ssl3 is incomplete.
1921 When OpenSSL is configured with "no-ssl3" as a build option, servers
1922 could accept and complete a SSL 3.0 handshake, and clients could be
1923 configured to send them.
1925 [Akamai and the OpenSSL team]
1927 *) Add support for TLS_FALLBACK_SCSV.
1928 Client applications doing fallback retries should call
1929 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1931 [Adam Langley, Bodo Moeller]
1933 *) Add additional DigestInfo checks.
1935 Reencode DigestInto in DER and check against the original when
1936 verifying RSA signature: this will reject any improperly encoded
1937 DigestInfo structures.
1939 Note: this is a precautionary measure and no attacks are currently known.
1943 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1945 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1946 SRP code can be overrun an internal buffer. Add sanity check that
1947 g, A, B < N to SRP code.
1949 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1950 Group for discovering this issue.
1954 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1955 TLS 1.0 instead of higher protocol versions when the ClientHello message
1956 is badly fragmented. This allows a man-in-the-middle attacker to force a
1957 downgrade to TLS 1.0 even if both the server and the client support a
1958 higher protocol version, by modifying the client's TLS records.
1960 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1961 researching this issue.
1965 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1966 to a denial of service attack. A malicious server can crash the client
1967 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1968 ciphersuite and sending carefully crafted handshake messages.
1970 Thanks to Felix Gröbert (Google) for discovering and researching this
1975 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1976 to leak memory. This can be exploited through a Denial of Service attack.
1977 Thanks to Adam Langley for discovering and researching this issue.
1981 *) An attacker can force openssl to consume large amounts of memory whilst
1982 processing DTLS handshake messages. This can be exploited through a
1983 Denial of Service attack.
1984 Thanks to Adam Langley for discovering and researching this issue.
1988 *) An attacker can force an error condition which causes openssl to crash
1989 whilst processing DTLS packets due to memory being freed twice. This
1990 can be exploited through a Denial of Service attack.
1991 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1996 *) If a multithreaded client connects to a malicious server using a resumed
1997 session and the server sends an ec point format extension it could write
1998 up to 255 bytes to freed memory.
2000 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2005 *) A malicious server can crash an OpenSSL client with a null pointer
2006 dereference (read) by specifying an SRP ciphersuite even though it was not
2007 properly negotiated with the client. This can be exploited through a
2008 Denial of Service attack.
2010 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2011 discovering and researching this issue.
2015 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2016 X509_name_oneline, X509_name_print_ex et al. to leak some information
2017 from the stack. Applications may be affected if they echo pretty printing
2018 output to the attacker.
2020 Thanks to Ivan Fratric (Google) for discovering this issue.
2022 [Emilia Käsper, and Steve Henson]
2024 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2025 for corner cases. (Certain input points at infinity could lead to
2026 bogus results, with non-infinity inputs mapped to infinity too.)
2029 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2031 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2032 handshake can force the use of weak keying material in OpenSSL
2033 SSL/TLS clients and servers.
2035 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2036 researching this issue. (CVE-2014-0224)
2037 [KIKUCHI Masashi, Steve Henson]
2039 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2040 OpenSSL DTLS client the code can be made to recurse eventually crashing
2043 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2045 [Imre Rad, Steve Henson]
2047 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2048 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2049 client or server. This is potentially exploitable to run arbitrary
2050 code on a vulnerable client or server.
2052 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2053 [Jüri Aedla, Steve Henson]
2055 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2056 are subject to a denial of service attack.
2058 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2059 this issue. (CVE-2014-3470)
2060 [Felix Gröbert, Ivan Fratric, Steve Henson]
2062 *) Harmonize version and its documentation. -f flag is used to display
2064 [mancha <mancha1@zoho.com>]
2066 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2067 in i2d_ECPrivateKey.
2068 [mancha <mancha1@zoho.com>]
2070 *) Fix some double frees. These are not thought to be exploitable.
2071 [mancha <mancha1@zoho.com>]
2073 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2075 *) A missing bounds check in the handling of the TLS heartbeat extension
2076 can be used to reveal up to 64k of memory to a connected client or
2079 Thanks for Neel Mehta of Google Security for discovering this bug and to
2080 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2081 preparing the fix (CVE-2014-0160)
2082 [Adam Langley, Bodo Moeller]
2084 *) Fix for the attack described in the paper "Recovering OpenSSL
2085 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2086 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2087 http://eprint.iacr.org/2014/140
2089 Thanks to Yuval Yarom and Naomi Benger for discovering this
2090 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2091 [Yuval Yarom and Naomi Benger]
2093 *) TLS pad extension: draft-agl-tls-padding-03
2095 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2096 TLS client Hello record length value would otherwise be > 255 and
2097 less that 512 pad with a dummy extension containing zeroes so it
2098 is at least 512 bytes long.
2100 [Adam Langley, Steve Henson]
2102 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2104 *) Fix for TLS record tampering bug. A carefully crafted invalid
2105 handshake could crash OpenSSL with a NULL pointer exception.
2106 Thanks to Anton Johansson for reporting this issues.
2109 *) Keep original DTLS digest and encryption contexts in retransmission
2110 structures so we can use the previous session parameters if they need
2111 to be resent. (CVE-2013-6450)
2114 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2115 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2116 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2117 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2118 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2119 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2120 [Rob Stradling, Adam Langley]
2122 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2124 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2125 supporting platforms or when small records were transferred.
2126 [Andy Polyakov, Steve Henson]
2128 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2130 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2132 This addresses the flaw in CBC record processing discovered by
2133 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2134 at: http://www.isg.rhul.ac.uk/tls/
2136 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2137 Security Group at Royal Holloway, University of London
2138 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2139 Emilia Käsper for the initial patch.
2141 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2143 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2144 ciphersuites which can be exploited in a denial of service attack.
2145 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2146 and detecting this bug and to Wolfgang Ettlinger
2147 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2151 *) Return an error when checking OCSP signatures when key is NULL.
2152 This fixes a DoS attack. (CVE-2013-0166)
2155 *) Make openssl verify return errors.
2156 [Chris Palmer <palmer@google.com> and Ben Laurie]
2158 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2159 the right response is stapled. Also change SSL_get_certificate()
2160 so it returns the certificate actually sent.
2161 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2162 [Rob Stradling <rob.stradling@comodo.com>]
2164 *) Fix possible deadlock when decoding public keys.
2167 *) Don't use TLS 1.0 record version number in initial client hello
2171 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2173 *) Sanity check record length before skipping explicit IV in TLS
2174 1.2, 1.1 and DTLS to fix DoS attack.
2176 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2177 fuzzing as a service testing platform.
2181 *) Initialise tkeylen properly when encrypting CMS messages.
2182 Thanks to Solar Designer of Openwall for reporting this issue.
2185 *) In FIPS mode don't try to use composite ciphers as they are not
2189 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2191 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2192 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2193 mean any application compiled against OpenSSL 1.0.0 headers setting
2194 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2195 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2196 0x10000000L Any application which was previously compiled against
2197 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2198 will need to be recompiled as a result. Letting be results in
2199 inability to disable specifically TLS 1.1 and in client context,
2200 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2203 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2204 disable just protocol X, but all protocols above X *if* there are
2205 protocols *below* X still enabled. In more practical terms it means
2206 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2207 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2208 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2212 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2214 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2215 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2216 in CRYPTO_realloc_clean.
2218 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2219 issue and to Adam Langley <agl@chromium.org> for fixing it.
2221 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2223 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2226 *) Workarounds for some broken servers that "hang" if a client hello
2227 record length exceeds 255 bytes.
2229 1. Do not use record version number > TLS 1.0 in initial client
2230 hello: some (but not all) hanging servers will now work.
2231 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2232 the number of ciphers sent in the client hello. This should be
2233 set to an even number, such as 50, for example by passing:
2234 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2235 Most broken servers should now work.
2236 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2237 TLS 1.2 client support entirely.
2240 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2243 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2245 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2246 STRING form instead of a DigestInfo.
2249 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2250 and the RSA_sign/RSA_verify functions. This was made more apparent when
2251 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2252 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2253 the correct format in RSA_verify so both forms transparently work.
2256 *) Some servers which support TLS 1.0 can choke if we initially indicate
2257 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2258 encrypted premaster secret. As a workaround use the maximum permitted
2259 client version in client hello, this should keep such servers happy
2260 and still work with previous versions of OpenSSL.
2263 *) Add support for TLS/DTLS heartbeats.
2264 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2266 *) Add support for SCTP.
2267 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2269 *) Improved PRNG seeding for VOS.
2270 [Paul Green <Paul.Green@stratus.com>]
2272 *) Extensive assembler packs updates, most notably:
2274 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2275 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2276 - x86_64: bit-sliced AES implementation;
2277 - ARM: NEON support, contemporary platforms optimizations;
2278 - s390x: z196 support;
2279 - *: GHASH and GF(2^m) multiplication implementations;
2283 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2284 (removal of unnecessary code)
2285 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2287 *) Add TLS key material exporter from RFC 5705.
2290 *) Add DTLS-SRTP negotiation from RFC 5764.
2293 *) Add Next Protocol Negotiation,
2294 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2295 disabled with a no-npn flag to config or Configure. Code donated
2297 [Adam Langley <agl@google.com> and Ben Laurie]
2299 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2300 NIST-P256, NIST-P521, with constant-time single point multiplication on
2301 typical inputs. Compiler support for the nonstandard type __uint128_t is
2302 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2303 Code made available under Apache License version 2.0.
2305 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2306 line to include this in your build of OpenSSL, and run "make depend" (or
2307 "make update"). This enables the following EC_METHODs:
2309 EC_GFp_nistp224_method()
2310 EC_GFp_nistp256_method()
2311 EC_GFp_nistp521_method()
2313 EC_GROUP_new_by_curve_name() will automatically use these (while
2314 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2316 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2318 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2319 all platforms. Move ssize_t definition from e_os.h to the public
2320 header file e_os2.h as it now appears in public header file cms.h
2323 *) New -sigopt option to the ca, req and x509 utilities. Additional
2324 signature parameters can be passed using this option and in
2328 *) Add RSA PSS signing function. This will generate and set the
2329 appropriate AlgorithmIdentifiers for PSS based on those in the
2330 corresponding EVP_MD_CTX structure. No application support yet.
2333 *) Support for companion algorithm specific ASN1 signing routines.
2334 New function ASN1_item_sign_ctx() signs a pre-initialised
2335 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2336 the appropriate parameters.
2339 *) Add new algorithm specific ASN1 verification initialisation function
2340 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2341 handling will be the same no matter what EVP_PKEY_METHOD is used.
2342 Add a PSS handler to support verification of PSS signatures: checked
2343 against a number of sample certificates.
2346 *) Add signature printing for PSS. Add PSS OIDs.
2347 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2349 *) Add algorithm specific signature printing. An individual ASN1 method
2350 can now print out signatures instead of the standard hex dump.
2352 More complex signatures (e.g. PSS) can print out more meaningful
2353 information. Include DSA version that prints out the signature
2357 *) Password based recipient info support for CMS library: implementing
2361 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2362 neatly separates the code into cipher and PBE sections and is required
2363 for some algorithms that split PBES2 into separate pieces (such as
2364 password based CMS).
2367 *) Session-handling fixes:
2368 - Fix handling of connections that are resuming with a session ID,
2369 but also support Session Tickets.
2370 - Fix a bug that suppressed issuing of a new ticket if the client
2371 presented a ticket with an expired session.
2372 - Try to set the ticket lifetime hint to something reasonable.
2373 - Make tickets shorter by excluding irrelevant information.
2374 - On the client side, don't ignore renewed tickets.
2375 [Adam Langley, Bodo Moeller (Google)]
2377 *) Fix PSK session representation.
2380 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2382 This work was sponsored by Intel.
2385 *) Add GCM support to TLS library. Some custom code is needed to split
2386 the IV between the fixed (from PRF) and explicit (from TLS record)
2387 portions. This adds all GCM ciphersuites supported by RFC5288 and
2388 RFC5289. Generalise some AES* cipherstrings to include GCM and
2389 add a special AESGCM string for GCM only.
2392 *) Expand range of ctrls for AES GCM. Permit setting invocation
2393 field on decrypt and retrieval of invocation field only on encrypt.
2396 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2397 As required by RFC5289 these ciphersuites cannot be used if for
2398 versions of TLS earlier than 1.2.
2401 *) For FIPS capable OpenSSL interpret a NULL default public key method
2402 as unset and return the appropriate default but do *not* set the default.
2403 This means we can return the appropriate method in applications that
2404 switch between FIPS and non-FIPS modes.
2407 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2408 ENGINE is used then we cannot handle that in the FIPS module so we
2409 keep original code iff non-FIPS operations are allowed.
2412 *) Add -attime option to openssl utilities.
2413 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2415 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2418 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2419 FIPS EC methods unconditionally for now.
2422 *) New build option no-ec2m to disable characteristic 2 code.
2425 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2426 all cases can be covered as some introduce binary incompatibilities.
2429 *) Redirect RSA operations to FIPS module including keygen,
2430 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2433 *) Add similar low level API blocking to ciphers.
2436 *) Low level digest APIs are not approved in FIPS mode: any attempt
2437 to use these will cause a fatal error. Applications that *really* want
2438 to use them can use the private_* version instead.
2441 *) Redirect cipher operations to FIPS module for FIPS builds.
2444 *) Redirect digest operations to FIPS module for FIPS builds.
2447 *) Update build system to add "fips" flag which will link in fipscanister.o
2448 for static and shared library builds embedding a signature if needed.
2451 *) Output TLS supported curves in preference order instead of numerical
2452 order. This is currently hardcoded for the highest order curves first.
2453 This should be configurable so applications can judge speed vs strength.
2456 *) Add TLS v1.2 server support for client authentication.
2459 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2463 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2464 FIPS modules versions.
2467 *) Add TLS v1.2 client side support for client authentication. Keep cache
2468 of handshake records longer as we don't know the hash algorithm to use
2469 until after the certificate request message is received.
2472 *) Initial TLS v1.2 client support. Add a default signature algorithms
2473 extension including all the algorithms we support. Parse new signature
2474 format in client key exchange. Relax some ECC signing restrictions for
2475 TLS v1.2 as indicated in RFC5246.
2478 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2479 to new signature format when needed using client digest preference.
2480 All server ciphersuites should now work correctly in TLS v1.2. No client
2481 support yet and no support for client certificates.
2484 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2485 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2486 ciphersuites. At present only RSA key exchange ciphersuites work with
2487 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2488 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2489 and version checking.
2492 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2493 with this defined it will not be affected by any changes to ssl internal
2494 structures. Add several utility functions to allow openssl application
2495 to work with OPENSSL_NO_SSL_INTERN defined.
2499 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2501 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2504 *) Permit abbreviated handshakes when renegotiating using the function
2505 SSL_renegotiate_abbreviated().
2506 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2508 *) Add call to ENGINE_register_all_complete() to
2509 ENGINE_load_builtin_engines(), so some implementations get used
2510 automatically instead of needing explicit application support.
2513 *) Add support for TLS key exporter as described in RFC5705.
2514 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2516 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2517 a few changes are required:
2519 Add SSL_OP_NO_TLSv1_1 flag.
2520 Add TLSv1_1 methods.
2521 Update version checking logic to handle version 1.1.
2522 Add explicit IV handling (ported from DTLS code).
2523 Add command line options to s_client/s_server.
2526 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2528 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2529 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2530 content decryption and always return the same error. Note: this attack
2531 needs on average 2^20 messages so it only affects automated senders. The
2532 old behaviour can be reenabled in the CMS code by setting the
2533 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2534 an MMA defence is not necessary.
2535 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2536 this issue. (CVE-2012-0884)
2539 *) Fix CVE-2011-4619: make sure we really are receiving a
2540 client hello before rejecting multiple SGC restarts. Thanks to
2541 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2544 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2546 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2547 Thanks to Antonio Martin, Enterprise Secure Access Research and
2548 Development, Cisco Systems, Inc. for discovering this bug and
2549 preparing a fix. (CVE-2012-0050)
2552 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2554 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2555 of the Vaudenay padding oracle attack on CBC mode encryption
2556 which enables an efficient plaintext recovery attack against
2557 the OpenSSL implementation of DTLS. Their attack exploits timing
2558 differences arising during decryption processing. A research
2559 paper describing this attack can be found at:
2560 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2561 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2562 Security Group at Royal Holloway, University of London
2563 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2564 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2565 for preparing the fix. (CVE-2011-4108)
2566 [Robin Seggelmann, Michael Tuexen]
2568 *) Clear bytes used for block padding of SSL 3.0 records.
2570 [Adam Langley (Google)]
2572 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2573 Kadianakis <desnacked@gmail.com> for discovering this issue and
2574 Adam Langley for preparing the fix. (CVE-2011-4619)
2575 [Adam Langley (Google)]
2577 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2578 [Andrey Kulikov <amdeich@gmail.com>]
2580 *) Prevent malformed RFC3779 data triggering an assertion failure.
2581 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2582 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2583 [Rob Austein <sra@hactrn.net>]
2585 *) Improved PRNG seeding for VOS.
2586 [Paul Green <Paul.Green@stratus.com>]
2588 *) Fix ssl_ciph.c set-up race.
2589 [Adam Langley (Google)]
2591 *) Fix spurious failures in ecdsatest.c.
2592 [Emilia Käsper (Google)]
2594 *) Fix the BIO_f_buffer() implementation (which was mixing different
2595 interpretations of the '..._len' fields).
2596 [Adam Langley (Google)]
2598 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2599 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2600 threads won't reuse the same blinding coefficients.
2602 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2603 lock to call BN_BLINDING_invert_ex, and avoids one use of
2604 BN_BLINDING_update for each BN_BLINDING structure (previously,
2605 the last update always remained unused).
2606 [Emilia Käsper (Google)]
2608 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2609 [Bob Buckholz (Google)]
2611 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2613 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2614 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2615 [Kaspar Brand <ossl@velox.ch>]
2617 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2618 for multi-threaded use of ECDH. (CVE-2011-3210)
2619 [Adam Langley (Google)]
2621 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2624 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2625 signature public key algorithm by using OID xref utilities instead.
2626 Before this you could only use some ECC ciphersuites with SHA1 only.
2629 *) Add protection against ECDSA timing attacks as mentioned in the paper
2630 by Billy Bob Brumley and Nicola Tuveri, see:
2632 http://eprint.iacr.org/2011/232.pdf
2634 [Billy Bob Brumley and Nicola Tuveri]
2636 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2638 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2639 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2641 *) Fix bug in string printing code: if *any* escaping is enabled we must
2642 escape the escape character (backslash) or the resulting string is
2646 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2648 *) Disable code workaround for ancient and obsolete Netscape browsers
2649 and servers: an attacker can use it in a ciphersuite downgrade attack.
2650 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2653 *) Fixed J-PAKE implementation error, originally discovered by
2654 Sebastien Martini, further info and confirmation from Stefan
2655 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2658 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2660 *) Fix extension code to avoid race conditions which can result in a buffer
2661 overrun vulnerability: resumed sessions must not be modified as they can
2662 be shared by multiple threads. CVE-2010-3864
2665 *) Fix WIN32 build system to correctly link an ENGINE directory into
2669 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2671 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2673 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2675 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2677 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2678 context. The operation can be customised via the ctrl mechanism in
2679 case ENGINEs want to include additional functionality.
2682 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2685 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2686 output hashes compatible with older versions of OpenSSL.
2687 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2689 *) Fix compression algorithm handling: if resuming a session use the
2690 compression algorithm of the resumed session instead of determining
2691 it from client hello again. Don't allow server to change algorithm.
2694 *) Add load_crls() function to apps tidying load_certs() too. Add option
2695 to verify utility to allow additional CRLs to be included.
2698 *) Update OCSP request code to permit adding custom headers to the request:
2699 some responders need this.
2702 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2704 [Julia Lawall <julia@diku.dk>]
2706 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2707 needlessly dereferenced structures, used obsolete functions and
2708 didn't handle all updated verify codes correctly.
2711 *) Disable MD2 in the default configuration.
2714 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2715 indicate the initial BIO being pushed or popped. This makes it possible
2716 to determine whether the BIO is the one explicitly called or as a result
2717 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2718 it handles reference counts correctly and doesn't zero out the I/O bio
2719 when it is not being explicitly popped. WARNING: applications which
2720 included workarounds for the old buggy behaviour will need to be modified
2721 or they could free up already freed BIOs.
2724 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2725 renaming to all platforms (within the 0.9.8 branch, this was
2726 done conditionally on Netware platforms to avoid a name clash).
2727 [Guenter <lists@gknw.net>]
2729 *) Add ECDHE and PSK support to DTLS.
2730 [Michael Tuexen <tuexen@fh-muenster.de>]
2732 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2736 *) Add "missing" function EVP_MD_flags() (without this the only way to
2737 retrieve a digest flags is by accessing the structure directly. Update
2738 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2739 or cipher is registered as in the "from" argument. Print out all
2740 registered digests in the dgst usage message instead of manually
2741 attempting to work them out.
2744 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2745 this allows the use of compression and extensions. Change default cipher
2746 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2747 by default unless an application cipher string requests it.
2750 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2751 key ids to find matching certificates and keys but some PKCS#12 files
2752 don't follow the (somewhat unwritten) rules and this strategy fails.
2753 Now just gather all certificates together and the first private key
2754 then look for the first certificate that matches the key.
2757 *) Support use of registered digest and cipher names for dgst and cipher
2758 commands instead of having to add each one as a special case. So now
2765 openssl dgst -sha256 foo
2767 and this works for ENGINE based algorithms too.
2771 *) Update Gost ENGINE to support parameter files.
2772 [Victor B. Wagner <vitus@cryptocom.ru>]
2774 *) Support GeneralizedTime in ca utility.
2775 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2777 *) Enhance the hash format used for certificate directory links. The new
2778 form uses the canonical encoding (meaning equivalent names will work
2779 even if they aren't identical) and uses SHA1 instead of MD5. This form
2780 is incompatible with the older format and as a result c_rehash should
2781 be used to rebuild symbolic links.
2784 *) Make PKCS#8 the default write format for private keys, replacing the
2785 traditional format. This form is standardised, more secure and doesn't
2786 include an implicit MD5 dependency.
2789 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2790 committed to OpenSSL should pass this lot as a minimum.
2793 *) Add session ticket override functionality for use by EAP-FAST.
2794 [Jouni Malinen <j@w1.fi>]
2796 *) Modify HMAC functions to return a value. Since these can be implemented
2797 in an ENGINE errors can occur.
2800 *) Type-checked OBJ_bsearch_ex.
2803 *) Type-checked OBJ_bsearch. Also some constification necessitated
2804 by type-checking. Still to come: TXT_DB, bsearch(?),
2805 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2809 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2810 seconds to a tm structure directly, instead of going through OS
2811 specific date routines. This avoids any issues with OS routines such
2812 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2813 and X509_time_adj_ex() to cover the extended range. The existing
2814 X509_time_adj() is still usable and will no longer have any date issues.
2817 *) Delta CRL support. New use deltas option which will attempt to locate
2818 and search any appropriate delta CRLs available.
2820 This work was sponsored by Google.
2823 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2824 code and add additional score elements. Validate alternate CRL paths
2825 as part of the CRL checking and indicate a new error "CRL path validation
2826 error" in this case. Applications wanting additional details can use
2827 the verify callback and check the new "parent" field. If this is not
2828 NULL CRL path validation is taking place. Existing applications wont
2829 see this because it requires extended CRL support which is off by
2832 This work was sponsored by Google.
2835 *) Support for freshest CRL extension.
2837 This work was sponsored by Google.
2840 *) Initial indirect CRL support. Currently only supported in the CRLs
2841 passed directly and not via lookup. Process certificate issuer
2842 CRL entry extension and lookup CRL entries by bother issuer name
2843 and serial number. Check and process CRL issuer entry in IDP extension.
2845 This work was sponsored by Google.
2848 *) Add support for distinct certificate and CRL paths. The CRL issuer
2849 certificate is validated separately in this case. Only enabled if
2850 an extended CRL support flag is set: this flag will enable additional
2851 CRL functionality in future.
2853 This work was sponsored by Google.
2856 *) Add support for policy mappings extension.
2858 This work was sponsored by Google.
2861 *) Fixes to pathlength constraint, self issued certificate handling,
2862 policy processing to align with RFC3280 and PKITS tests.
2864 This work was sponsored by Google.
2867 *) Support for name constraints certificate extension. DN, email, DNS
2868 and URI types are currently supported.
2870 This work was sponsored by Google.
2873 *) To cater for systems that provide a pointer-based thread ID rather
2874 than numeric, deprecate the current numeric thread ID mechanism and
2875 replace it with a structure and associated callback type. This
2876 mechanism allows a numeric "hash" to be extracted from a thread ID in
2877 either case, and on platforms where pointers are larger than 'long',
2878 mixing is done to help ensure the numeric 'hash' is usable even if it
2879 can't be guaranteed unique. The default mechanism is to use "&errno"
2880 as a pointer-based thread ID to distinguish between threads.
2882 Applications that want to provide their own thread IDs should now use
2883 CRYPTO_THREADID_set_callback() to register a callback that will call
2884 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2886 Note that ERR_remove_state() is now deprecated, because it is tied
2887 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2888 to free the current thread's error state should be replaced by
2889 ERR_remove_thread_state(NULL).
2891 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2892 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2893 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2894 application was previously providing a numeric thread callback that
2895 was inappropriate for distinguishing threads, then uniqueness might
2896 have been obtained with &errno that happened immediately in the
2897 intermediate development versions of OpenSSL; this is no longer the
2898 case, the numeric thread callback will now override the automatic use
2900 [Geoff Thorpe, with help from Bodo Moeller]
2902 *) Initial support for different CRL issuing certificates. This covers a
2903 simple case where the self issued certificates in the chain exist and
2904 the real CRL issuer is higher in the existing chain.
2906 This work was sponsored by Google.
2909 *) Removed effectively defunct crypto/store from the build.
2912 *) Revamp of STACK to provide stronger type-checking. Still to come:
2913 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2914 ASN1_STRING, CONF_VALUE.
2917 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2918 RAM on SSL connections. This option can save about 34k per idle SSL.
2921 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2922 STACK, TXT_DB, bsearch, qsort.
2925 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2926 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2927 support for data, signedData, compressedData, digestedData and
2928 encryptedData, envelopedData types included. Scripts to check against
2929 RFC4134 examples draft and interop and consistency checks of many
2930 content types and variants.
2933 *) Add options to enc utility to support use of zlib compression BIO.
2936 *) Extend mk1mf to support importing of options and assembly language
2937 files from Configure script, currently only included in VC-WIN32.
2938 The assembly language rules can now optionally generate the source
2939 files from the associated perl scripts.
2942 *) Implement remaining functionality needed to support GOST ciphersuites.
2943 Interop testing has been performed using CryptoPro implementations.
2944 [Victor B. Wagner <vitus@cryptocom.ru>]
2946 *) s390x assembler pack.
2949 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2953 *) Implement Opaque PRF Input TLS extension as specified in
2954 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2955 official specification yet and no extension type assignment by
2956 IANA exists, this extension (for now) will have to be explicitly
2957 enabled when building OpenSSL by providing the extension number
2958 to use. For example, specify an option
2960 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2962 to the "config" or "Configure" script to enable the extension,
2963 assuming extension number 0x9527 (which is a completely arbitrary
2964 and unofficial assignment based on the MD5 hash of the Internet
2965 Draft). Note that by doing so, you potentially lose
2966 interoperability with other TLS implementations since these might
2967 be using the same extension number for other purposes.
2969 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2970 opaque PRF input value to use in the handshake. This will create
2971 an interal copy of the length-'len' string at 'src', and will
2972 return non-zero for success.
2974 To get more control and flexibility, provide a callback function
2977 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2978 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2982 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2985 Callback function 'cb' will be called in handshakes, and is
2986 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2987 Argument 'arg' is for application purposes (the value as given to
2988 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2989 be provided to the callback function). The callback function
2990 has to return non-zero to report success: usually 1 to use opaque
2991 PRF input just if possible, or 2 to enforce use of the opaque PRF
2992 input. In the latter case, the library will abort the handshake
2993 if opaque PRF input is not successfully negotiated.
2995 Arguments 'peerinput' and 'len' given to the callback function
2996 will always be NULL and 0 in the case of a client. A server will
2997 see the client's opaque PRF input through these variables if
2998 available (NULL and 0 otherwise). Note that if the server
2999 provides an opaque PRF input, the length must be the same as the
3000 length of the client's opaque PRF input.
3002 Note that the callback function will only be called when creating
3003 a new session (session resumption can resume whatever was
3004 previously negotiated), and will not be called in SSL 2.0
3005 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3006 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3007 for applications that need to enforce opaque PRF input.
3011 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3014 [Victor B. Wagner <vitus@cryptocom.ru>]
3016 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3017 RFC4507bis. The encrypted ticket format is an encrypted encoded
3018 SSL_SESSION structure, that way new session features are automatically
3021 If a client application caches session in an SSL_SESSION structure
3022 support is transparent because tickets are now stored in the encoded
3025 The SSL_CTX structure automatically generates keys for ticket
3026 protection in servers so again support should be possible
3027 with no application modification.
3029 If a client or server wishes to disable RFC4507 support then the option
3030 SSL_OP_NO_TICKET can be set.
3032 Add a TLS extension debugging callback to allow the contents of any client
3033 or server extensions to be examined.
3035 This work was sponsored by Google.
3038 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3039 OpenSSL should now compile cleanly on gcc 4.2
3040 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3042 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3043 support including streaming MAC support: this is required for GOST
3044 ciphersuite support.
3045 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3047 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3048 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3049 to output in BER and PEM format.
3052 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3053 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3054 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3055 ENGINE support for HMAC keys which are unextractable. New -mac and
3056 -macopt options to dgst utility.
3059 *) New option -sigopt to dgst utility. Update dgst to use
3060 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3061 alternative signing parameters such as X9.31 or PSS in the dgst
3065 *) Change ssl_cipher_apply_rule(), the internal function that does
3066 the work each time a ciphersuite string requests enabling
3067 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3068 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3069 the order of disabled ciphersuites such that those ciphersuites
3070 that most recently went from enabled to disabled not only stay
3071 in order with respect to each other, but also have higher priority
3072 than other disabled ciphersuites the next time ciphersuites are
3075 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3076 the same ciphersuites as with "HIGH" alone, but in a specific
3077 order where the PSK ciphersuites come first (since they are the
3078 most recently disabled ciphersuites when "HIGH" is parsed).
3080 Also, change ssl_create_cipher_list() (using this new
3081 funcionality) such that between otherwise identical
3082 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3086 *) Change ssl_create_cipher_list() so that it automatically
3087 arranges the ciphersuites in reasonable order before starting
3088 to process the rule string. Thus, the definition for "DEFAULT"
3089 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3090 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3091 This makes it much easier to arrive at a reasonable default order
3092 in applications for which anonymous ciphers are OK (meaning
3093 that you can't actually use DEFAULT).
3094 [Bodo Moeller; suggested by Victor Duchovni]
3096 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3097 processing) into multiple integers instead of setting
3098 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3099 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3100 (These masks as well as the individual bit definitions are hidden
3101 away into the non-exported interface ssl/ssl_locl.h, so this
3102 change to the definition of the SSL_CIPHER structure shouldn't
3103 affect applications.) This give us more bits for each of these
3104 categories, so there is no longer a need to coagulate AES128 and
3105 AES256 into a single algorithm bit, and to coagulate Camellia128
3106 and Camellia256 into a single algorithm bit, which has led to all
3109 Thus, among other things, the kludge introduced in 0.9.7m and
3110 0.9.8e for masking out AES256 independently of AES128 or masking
3111 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3113 With the change, we also introduce new ciphersuite aliases that
3114 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3118 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3119 Use the leftmost N bytes of the signature input if the input is
3120 larger than the prime q (with N being the size in bytes of q).
3123 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3124 it yet and it is largely untested.
3127 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3130 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3131 some compilers (gcc 4.2 and later) reject their use. Safestack is
3132 reimplemented. Update ASN1 to avoid use of legacy functions.
3135 *) Win32/64 targets are linked with Winsock2.
3138 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3139 to external functions. This can be used to increase CRL handling
3140 efficiency especially when CRLs are very large by (for example) storing
3141 the CRL revoked certificates in a database.
3144 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3145 new CRLs added to a directory can be used. New command line option
3146 -verify_return_error to s_client and s_server. This causes real errors
3147 to be returned by the verify callback instead of carrying on no matter
3148 what. This reflects the way a "real world" verify callback would behave.
3151 *) GOST engine, supporting several GOST algorithms and public key formats.
3152 Kindly donated by Cryptocom.
3155 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3156 partitioned by DP are handled but no indirect CRL or reason partitioning
3157 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3158 selected via a scoring technique which handles IDP and AKID in CRLs.
3161 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3162 will ultimately be used for all verify operations: this will remove the
3163 X509_STORE dependency on certificate verification and allow alternative
3164 lookup methods. X509_STORE based implementations of these two callbacks.
3167 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3168 Modify get_crl() to find a valid (unexpired) CRL if possible.
3171 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3172 this would be called X509_CRL_cmp() but that name is already used by
3173 a function that just compares CRL issuer names. Cache several CRL
3174 extensions in X509_CRL structure and cache CRLDP in X509.
3177 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3178 this maps equivalent X509_NAME structures into a consistent structure.
3179 Name comparison can then be performed rapidly using memcmp().
3182 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3186 *) Allow digests to supply their own micalg string for S/MIME type using
3187 the ctrl EVP_MD_CTRL_MICALG.
3190 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3191 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3192 ctrl. It can then customise the structure before and/or after signing
3196 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3197 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3198 to free up any added signature OIDs.
3201 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3202 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3203 digest and cipher tables. New options added to openssl utility:
3204 list-message-digest-algorithms and list-cipher-algorithms.
3207 *) Change the array representation of binary polynomials: the list
3208 of degrees of non-zero coefficients is now terminated with -1.
3209 Previously it was terminated with 0, which was also part of the
3210 value; thus, the array representation was not applicable to
3211 polynomials where t^0 has coefficient zero. This change makes
3212 the array representation useful in a more general context.
3215 *) Various modifications and fixes to SSL/TLS cipher string
3216 handling. For ECC, the code now distinguishes between fixed ECDH
3217 with RSA certificates on the one hand and with ECDSA certificates
3218 on the other hand, since these are separate ciphersuites. The
3219 unused code for Fortezza ciphersuites has been removed.
3221 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3222 (not "ECDHE"). For consistency with the code for DH
3223 certificates, use of ECDH certificates is now considered ECDH
3224 authentication, not RSA or ECDSA authentication (the latter is
3225 merely the CA's signing algorithm and not actively used in the
3228 The temporary ciphersuite alias "ECCdraft" is no longer
3229 available, and ECC ciphersuites are no longer excluded from "ALL"
3230 and "DEFAULT". The following aliases now exist for RFC 4492
3231 ciphersuites, most of these by analogy with the DH case:
3233 kECDHr - ECDH cert, signed with RSA
3234 kECDHe - ECDH cert, signed with ECDSA
3235 kECDH - ECDH cert (signed with either RSA or ECDSA)
3236 kEECDH - ephemeral ECDH
3237 ECDH - ECDH cert or ephemeral ECDH
3243 AECDH - anonymous ECDH
3244 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3248 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3249 Use correct micalg parameters depending on digest(s) in signed message.
3252 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3253 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3256 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3257 an engine to register a method. Add ENGINE lookups for methods and
3258 functional reference processing.
3261 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3262 EVP_{Sign,Verify}* which allow an application to customise the signature
3266 *) New -resign option to smime utility. This adds one or more signers
3267 to an existing PKCS#7 signedData structure. Also -md option to use an
3268 alternative message digest algorithm for signing.
3271 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3272 create PKCS7 structures containing multiple signers. Update smime
3273 application to support multiple signers.
3276 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3280 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3281 Reorganize PBE internals to lookup from a static table using NIDs,
3282 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3283 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3284 PRF which will be automatically used with PBES2.
3287 *) Replace the algorithm specific calls to generate keys in "req" with the
3291 *) Update PKCS#7 enveloped data routines to use new API. This is now
3292 supported by any public key method supporting the encrypt operation. A
3293 ctrl is added to allow the public key algorithm to examine or modify
3294 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3298 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3299 a default digest type to use. In most cases this will be SHA1 but some
3300 algorithms (such as GOST) need to specify an alternative digest. The
3301 return value indicates how strong the preference is 1 means optional and
3302 2 is mandatory (that is it is the only supported type). Modify
3303 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3304 use the default md. Update openssl utilities to use the default digest
3305 type for signing if it is not explicitly indicated.
3308 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3309 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3310 signing method from the key type. This effectively removes the link
3311 between digests and public key types.
3314 *) Add an OID cross reference table and utility functions. Its purpose is to
3315 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3316 rsaEncryption. This will allow some of the algorithm specific hackery
3317 needed to use the correct OID to be removed.
3320 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3321 structures for PKCS7_sign(). They are now set up by the relevant public
3325 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3328 *) Add support for key derivation (agreement) in the API, DH method and
3332 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3333 public and private key formats. As a side effect these add additional
3334 command line functionality not previously available: DSA signatures can be
3335 generated and verified using pkeyutl and DH key support and generation in
3340 [Oliver Tappe <zooey@hirschkaefer.de>]
3342 *) New make target "install_html_docs" installs HTML renditions of the
3344 [Oliver Tappe <zooey@hirschkaefer.de>]
3346 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3347 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3348 support key and parameter generation and add initial key generation
3349 functionality for RSA.
3352 *) Add functions for main EVP_PKEY_method operations. The undocumented
3353 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3354 EVP_PKEY_{encrypt,decrypt}_old.
3357 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3358 key API, doesn't do much yet.
3361 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3362 public key algorithms. New option to openssl utility:
3363 "list-public-key-algorithms" to print out info.
3366 *) Implement the Supported Elliptic Curves Extension for
3367 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3370 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3371 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3374 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3375 utilities such as rsa, dsa, dsaparam etc except they process any key
3379 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3380 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3381 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3385 *) Initial support for pluggable public key ASN1.
3386 De-spaghettify the public key ASN1 handling. Move public and private
3387 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3388 algorithm specific handling to a single module within the relevant
3389 algorithm directory. Add functions to allow (near) opaque processing
3390 of public and private key structures.
3393 *) Implement the Supported Point Formats Extension for
3394 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3397 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3398 for the psk identity [hint] and the psk callback functions to the
3399 SSL_SESSION, SSL and SSL_CTX structure.
3402 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3406 SSL_CTX_use_psk_identity_hint
3407 SSL_get_psk_identity_hint
3408 SSL_get_psk_identity
3409 SSL_use_psk_identity_hint
3411 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3413 *) Add RFC 3161 compliant time stamp request creation, response generation
3414 and response verification functionality.
3415 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3417 *) Add initial support for TLS extensions, specifically for the server_name
3418 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3419 have new members for a host name. The SSL data structure has an
3420 additional member SSL_CTX *initial_ctx so that new sessions can be
3421 stored in that context to allow for session resumption, even after the
3422 SSL has been switched to a new SSL_CTX in reaction to a client's
3423 server_name extension.
3425 New functions (subject to change):
3427 SSL_get_servername()
3428 SSL_get_servername_type()
3431 New CTRL codes and macros (subject to change):
3433 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3434 - SSL_CTX_set_tlsext_servername_callback()
3435 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3436 - SSL_CTX_set_tlsext_servername_arg()
3437 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3439 openssl s_client has a new '-servername ...' option.
3441 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3442 '-key2 ...', '-servername_fatal' (subject to change). This allows
3443 testing the HostName extension for a specific single host name ('-cert'
3444 and '-key' remain fallbacks for handshakes without HostName
3445 negotiation). If the unrecognized_name alert has to be sent, this by
3446 default is a warning; it becomes fatal with the '-servername_fatal'
3449 [Peter Sylvester, Remy Allais, Christophe Renou]
3451 *) Whirlpool hash implementation is added.
3454 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3455 bn(64,32). Because of instruction set limitations it doesn't have
3456 any negative impact on performance. This was done mostly in order
3457 to make it possible to share assembler modules, such as bn_mul_mont
3458 implementations, between 32- and 64-bit builds without hassle.
3461 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3462 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3466 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3467 dedicated Montgomery multiplication procedure, is introduced.
3468 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3469 "64-bit" performance on certain 32-bit targets.
3472 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3473 in SSL structures. New SSL ctrl to set maximum send fragment size.
3474 Save memory by seeting the I/O buffer sizes dynamically instead of
3475 using the maximum available value.
3478 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3479 in addition to the text details.
3482 *) Very, very preliminary EXPERIMENTAL support for printing of general
3483 ASN1 structures. This currently produces rather ugly output and doesn't
3484 handle several customised structures at all.
3487 *) Integrated support for PVK file format and some related formats such
3488 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3489 these in the 'rsa' and 'dsa' utilities.
3492 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3495 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3496 place for the (very old) "NETSCAPE" format certificates which are now
3497 handled using new ASN1 code equivalents.
3500 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3501 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3502 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3505 *) Modify CRL distribution points extension code to print out previously
3506 unsupported fields. Enhance extension setting code to allow setting of
3510 *) Add print and set support for Issuing Distribution Point CRL extension.
3513 *) Change 'Configure' script to enable Camellia by default.
3516 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3518 *) When rejecting SSL/TLS records due to an incorrect version number, never
3519 update s->server with a new major version number. As of
3520 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3521 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3522 the previous behavior could result in a read attempt at NULL when
3523 receiving specific incorrect SSL/TLS records once record payload
3524 protection is active. (CVE-2010-0740)
3525 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3527 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3528 could be crashed if the relevant tables were not present (e.g. chrooted).
3529 [Tomas Hoger <thoger@redhat.com>]
3531 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3533 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3534 [Martin Olsson, Neel Mehta]
3536 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3537 accommodate for stack sorting, always a write lock!).
3540 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3541 excessive delays in the RAND_poll(): over a minute. As a workaround
3542 include a time check in the inner Heap32Next loop too.
3545 *) The code that handled flushing of data in SSL/TLS originally used the
3546 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3547 the problem outlined in PR#1949. The fix suggested there however can
3548 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3549 of Apache). So instead simplify the code to flush unconditionally.
3550 This should be fine since flushing with no data to flush is a no op.
3553 *) Handle TLS versions 2.0 and later properly and correctly use the
3554 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3555 off ancient servers have a habit of sticking around for a while...
3558 *) Modify compression code so it frees up structures without using the
3559 ex_data callbacks. This works around a problem where some applications
3560 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3561 restarting) then use compression (e.g. SSL with compression) later.
3562 This results in significant per-connection memory leaks and
3563 has caused some security issues including CVE-2008-1678 and
3567 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3568 change when encrypting or decrypting.
3571 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3572 connect and renegotiate with servers which do not support RI.
3573 Until RI is more widely deployed this option is enabled by default.
3576 *) Add "missing" ssl ctrls to clear options and mode.
3579 *) If client attempts to renegotiate and doesn't support RI respond with
3580 a no_renegotiation alert as required by RFC5746. Some renegotiating
3581 TLS clients will continue a connection gracefully when they receive
3582 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3583 waiting for a server hello which it will never receive. Now we treat a
3584 received no_renegotiation alert as a fatal error. This is because
3585 applications requesting a renegotiation might well expect it to succeed
3586 and would have no code in place to handle the server denying it so the
3587 only safe thing to do is to terminate the connection.
3590 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3591 peer supports secure renegotiation and 0 otherwise. Print out peer
3592 renegotiation support in s_client/s_server.
3595 *) Replace the highly broken and deprecated SPKAC certification method with
3596 the updated NID creation version. This should correctly handle UTF8.
3599 *) Implement RFC5746. Re-enable renegotiation but require the extension
3600 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3601 turns out to be a bad idea. It has been replaced by
3602 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3603 SSL_CTX_set_options(). This is really not recommended unless you
3604 know what you are doing.
3605 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3607 *) Fixes to stateless session resumption handling. Use initial_ctx when
3608 issuing and attempting to decrypt tickets in case it has changed during
3609 servername handling. Use a non-zero length session ID when attempting
3610 stateless session resumption: this makes it possible to determine if
3611 a resumption has occurred immediately after receiving server hello
3612 (several places in OpenSSL subtly assume this) instead of later in
3616 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3617 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3618 fixes for a few places where the return code is not checked
3620 [Julia Lawall <julia@diku.dk>]
3622 *) Add --strict-warnings option to Configure script to include devteam
3623 warnings in other configurations.
3626 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3627 makes it possible to install openssl libraries in locations which
3628 have names other than "lib", for example "/usr/lib64" which some
3630 [Steve Henson, based on patch from Jeremy Utley]
3632 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3633 X690 8.9.12 and can produce some misleading textual output of OIDs.
3634 [Steve Henson, reported by Dan Kaminsky]
3636 *) Delete MD2 from algorithm tables. This follows the recommendation in
3637 several standards that it is not used in new applications due to
3638 several cryptographic weaknesses. For binary compatibility reasons
3639 the MD2 API is still compiled in by default.
3642 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3646 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3647 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3649 [Guenter <lists@gknw.net>]
3651 *) Fix the server certificate chain building code to use X509_verify_cert(),
3652 it used to have an ad-hoc builder which was unable to cope with anything
3653 other than a simple chain.
3654 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3656 *) Don't check self signed certificate signatures in X509_verify_cert()
3657 by default (a flag can override this): it just wastes time without
3658 adding any security. As a useful side effect self signed root CAs
3659 with non-FIPS digests are now usable in FIPS mode.
3662 *) In dtls1_process_out_of_seq_message() the check if the current message
3663 is already buffered was missing. For every new message was memory
3664 allocated, allowing an attacker to perform an denial of service attack
3665 with sending out of seq handshake messages until there is no memory
3666 left. Additionally every future messege was buffered, even if the
3667 sequence number made no sense and would be part of another handshake.
3668 So only messages with sequence numbers less than 10 in advance will be
3669 buffered. (CVE-2009-1378)
3670 [Robin Seggelmann, discovered by Daniel Mentz]
3672 *) Records are buffered if they arrive with a future epoch to be
3673 processed after finishing the corresponding handshake. There is
3674 currently no limitation to this buffer allowing an attacker to perform
3675 a DOS attack with sending records with future epochs until there is no
3676 memory left. This patch adds the pqueue_size() function to determine
3677 the size of a buffer and limits the record buffer to 100 entries.
3679 [Robin Seggelmann, discovered by Daniel Mentz]
3681 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3682 parent structure is freed. (CVE-2009-1379)
3685 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3686 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3689 [Ilya O. <vrghost@gmail.com>]
3691 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3693 *) Disable renegotiation completely - this fixes a severe security
3694 problem (CVE-2009-3555) at the cost of breaking all
3695 renegotiation. Renegotiation can be re-enabled by setting
3696 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3697 run-time. This is really not recommended unless you know what
3701 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3703 *) Don't set val to NULL when freeing up structures, it is freed up by
3704 underlying code. If sizeof(void *) > sizeof(long) this can result in
3705 zeroing past the valid field. (CVE-2009-0789)
3706 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3708 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3709 checked correctly. This would allow some invalid signed attributes to
3710 appear to verify correctly. (CVE-2009-0591)
3711 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3713 *) Reject UniversalString and BMPString types with invalid lengths. This
3714 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3715 a legal length. (CVE-2009-0590)
3718 *) Set S/MIME signing as the default purpose rather than setting it
3719 unconditionally. This allows applications to override it at the store
3723 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3724 to handle some structures.
3727 *) Improve efficiency of mem_gets: don't search whole buffer each time
3729 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3731 *) New -hex option for openssl rand.
3734 *) Print out UTF8String and NumericString when parsing ASN1.
3737 *) Support NumericString type for name components.
3740 *) Allow CC in the environment to override the automatically chosen
3741 compiler. Note that nothing is done to ensure flags work with the
3745 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3747 *) Properly check EVP_VerifyFinal() and similar return values
3749 [Ben Laurie, Bodo Moeller, Google Security Team]
3751 *) Enable TLS extensions by default.
3754 *) Allow the CHIL engine to be loaded, whether the application is
3755 multithreaded or not. (This does not release the developer from the
3756 obligation to set up the dynamic locking callbacks.)
3757 [Sander Temme <sander@temme.net>]
3759 *) Use correct exit code if there is an error in dgst command.
3760 [Steve Henson; problem pointed out by Roland Dirlewanger]
3762 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3763 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3766 *) Add experimental JPAKE support, including demo authentication in
3767 s_client and s_server.
3770 *) Set the comparison function in v3_addr_canonize().
3771 [Rob Austein <sra@hactrn.net>]
3773 *) Add support for XMPP STARTTLS in s_client.
3774 [Philip Paeps <philip@freebsd.org>]
3776 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3777 to ensure that even with this option, only ciphersuites in the
3778 server's preference list will be accepted. (Note that the option
3779 applies only when resuming a session, so the earlier behavior was
3780 just about the algorithm choice for symmetric cryptography.)
3783 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3785 *) Fix NULL pointer dereference if a DTLS server received
3786 ChangeCipherSpec as first record (CVE-2009-1386).
3789 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3790 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3793 *) The fix in 0.9.8c that supposedly got rid of unsafe
3794 double-checked locking was incomplete for RSA blinding,
3795 addressing just one layer of what turns out to have been
3796 doubly unsafe triple-checked locking.
3798 So now fix this for real by retiring the MONT_HELPER macro
3799 in crypto/rsa/rsa_eay.c.
3801 [Bodo Moeller; problem pointed out by Marius Schilder]
3803 *) Various precautionary measures:
3805 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3807 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3808 (NB: This would require knowledge of the secret session ticket key
3809 to exploit, in which case you'd be SOL either way.)
3811 - Change bn_nist.c so that it will properly handle input BIGNUMs
3812 outside the expected range.
3814 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3817 [Neel Mehta, Bodo Moeller]
3819 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3820 the load fails. Useful for distros.
3821 [Ben Laurie and the FreeBSD team]
3823 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3826 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3829 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3831 This work was sponsored by Logica.
3834 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3835 keystores. Support for SSL/TLS client authentication too.
3836 Not compiled unless enable-capieng specified to Configure.
3838 This work was sponsored by Logica.
3841 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3842 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3843 attribute creation routines such as certifcate requests and PKCS#12
3847 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3849 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3850 handshake which could lead to a cilent crash as found using the
3851 Codenomicon TLS test suite (CVE-2008-1672)
3852 [Steve Henson, Mark Cox]
3854 *) Fix double free in TLS server name extensions which could lead to
3855 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3858 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3860 Clear the error queue to ensure that error entries left from
3861 older function calls do not interfere with the correct operation.
3862 [Lutz Jaenicke, Erik de Castro Lopo]
3864 *) Remove root CA certificates of commercial CAs:
3866 The OpenSSL project does not recommend any specific CA and does not
3867 have any policy with respect to including or excluding any CA.
3868 Therefore it does not make any sense to ship an arbitrary selection
3869 of root CA certificates with the OpenSSL software.
3872 *) RSA OAEP patches to fix two separate invalid memory reads.
3873 The first one involves inputs when 'lzero' is greater than
3874 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3875 before the beginning of from). The second one involves inputs where
3876 the 'db' section contains nothing but zeroes (there is a one-byte
3877 invalid read after the end of 'db').
3878 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3880 *) Partial backport from 0.9.9-dev:
3882 Introduce bn_mul_mont (dedicated Montgomery multiplication
3883 procedure) as a candidate for BIGNUM assembler implementation.
3884 While 0.9.9-dev uses assembler for various architectures, only
3885 x86_64 is available by default here in the 0.9.8 branch, and
3886 32-bit x86 is available through a compile-time setting.
3888 To try the 32-bit x86 assembler implementation, use Configure
3889 option "enable-montasm" (which exists only for this backport).
3891 As "enable-montasm" for 32-bit x86 disclaims code stability
3892 anyway, in this constellation we activate additional code
3893 backported from 0.9.9-dev for further performance improvements,
3894 namely BN_from_montgomery_word. (To enable this otherwise,
3895 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3897 [Andy Polyakov (backport partially by Bodo Moeller)]
3899 *) Add TLS session ticket callback. This allows an application to set
3900 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3901 values. This is useful for key rollover for example where several key
3902 sets may exist with different names.
3905 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3906 This was broken until now in 0.9.8 releases, such that the only way
3907 a registered ENGINE could be used (assuming it initialises
3908 successfully on the host) was to explicitly set it as the default
3909 for the relevant algorithms. This is in contradiction with 0.9.7
3910 behaviour and the documentation. With this fix, when an ENGINE is
3911 registered into a given algorithm's table of implementations, the
3912 'uptodate' flag is reset so that auto-discovery will be used next
3913 time a new context for that algorithm attempts to select an
3915 [Ian Lister (tweaked by Geoff Thorpe)]
3917 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3918 implemention in the following ways:
3920 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3923 Lack of BER streaming support means one pass streaming processing is
3924 only supported if data is detached: setting the streaming flag is
3925 ignored for embedded content.
3927 CMS support is disabled by default and must be explicitly enabled
3928 with the enable-cms configuration option.
3931 *) Update the GMP engine glue to do direct copies between BIGNUM and
3932 mpz_t when openssl and GMP use the same limb size. Otherwise the
3933 existing "conversion via a text string export" trick is still used.
3934 [Paul Sheer <paulsheer@gmail.com>]
3936 *) Zlib compression BIO. This is a filter BIO which compressed and
3937 uncompresses any data passed through it.
3940 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3941 RFC3394 compatible AES key wrapping.
3944 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3945 sets string data without copying. X509_ALGOR_set0() and
3946 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3947 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3948 from an X509_ATTRIBUTE structure optionally checking it occurs only
3949 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3953 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3954 to get the expected BN_FLG_CONSTTIME behavior.
3955 [Bodo Moeller (Google)]
3959 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3960 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3961 - added some more tests to do_tests.pl
3962 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3963 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3964 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3965 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3966 - various changes to netware.pl to enable gcc-cross builds on Win32
3968 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3969 - various changes to fix missing prototype warnings
3970 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3971 - added AES, WHIRLPOOL and CPUID assembler code to build files
3972 - added missing AES assembler make rules to mk1mf.pl
3973 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3974 [Guenter Knauf <eflash@gmx.net>]
3976 *) Implement certificate status request TLS extension defined in RFC3546.
3977 A client can set the appropriate parameters and receive the encoded
3978 OCSP response via a callback. A server can query the supplied parameters
3979 and set the encoded OCSP response in the callback. Add simplified examples
3980 to s_client and s_server.
3983 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3985 *) Fix various bugs:
3986 + Binary incompatibility of ssl_ctx_st structure
3987 + DTLS interoperation with non-compliant servers
3988 + Don't call get_session_cb() without proposed session
3989 + Fix ia64 assembler code
3990 [Andy Polyakov, Steve Henson]
3992 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3994 *) DTLS Handshake overhaul. There were longstanding issues with
3995 OpenSSL DTLS implementation, which were making it impossible for
3996 RFC 4347 compliant client to communicate with OpenSSL server.
3997 Unfortunately just fixing these incompatibilities would "cut off"
3998 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3999 server keeps tolerating non RFC compliant syntax. The opposite is
4000 not true, 0.9.8f client can not communicate with earlier server.
4001 This update even addresses CVE-2007-4995.
4004 *) Changes to avoid need for function casts in OpenSSL: some compilers
4005 (gcc 4.2 and later) reject their use.
4006 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4009 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4010 RFC4507bis. The encrypted ticket format is an encrypted encoded
4011 SSL_SESSION structure, that way new session features are automatically
4014 If a client application caches session in an SSL_SESSION structure
4015 support is transparent because tickets are now stored in the encoded
4018 The SSL_CTX structure automatically generates keys for ticket
4019 protection in servers so again support should be possible
4020 with no application modification.
4022 If a client or server wishes to disable RFC4507 support then the option
4023 SSL_OP_NO_TICKET can be set.
4025 Add a TLS extension debugging callback to allow the contents of any client
4026 or server extensions to be examined.
4028 This work was sponsored by Google.
4031 *) Add initial support for TLS extensions, specifically for the server_name
4032 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4033 have new members for a host name. The SSL data structure has an
4034 additional member SSL_CTX *initial_ctx so that new sessions can be
4035 stored in that context to allow for session resumption, even after the
4036 SSL has been switched to a new SSL_CTX in reaction to a client's
4037 server_name extension.
4039 New functions (subject to change):
4041 SSL_get_servername()
4042 SSL_get_servername_type()
4045 New CTRL codes and macros (subject to change):
4047 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4048 - SSL_CTX_set_tlsext_servername_callback()
4049 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4050 - SSL_CTX_set_tlsext_servername_arg()
4051 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4053 openssl s_client has a new '-servername ...' option.
4055 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4056 '-key2 ...', '-servername_fatal' (subject to change). This allows
4057 testing the HostName extension for a specific single host name ('-cert'
4058 and '-key' remain fallbacks for handshakes without HostName
4059 negotiation). If the unrecognized_name alert has to be sent, this by
4060 default is a warning; it becomes fatal with the '-servername_fatal'
4063 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4065 *) Add AES and SSE2 assembly language support to VC++ build.
4068 *) Mitigate attack on final subtraction in Montgomery reduction.
4071 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4072 (which previously caused an internal error).
4075 *) Squeeze another 10% out of IGE mode when in != out.
4078 *) AES IGE mode speedup.
4079 [Dean Gaudet (Google)]
4081 *) Add the Korean symmetric 128-bit cipher SEED (see
4082 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4083 add SEED ciphersuites from RFC 4162:
4085 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4086 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4087 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4088 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4090 To minimize changes between patchlevels in the OpenSSL 0.9.8
4091 series, SEED remains excluded from compilation unless OpenSSL
4092 is configured with 'enable-seed'.
4093 [KISA, Bodo Moeller]
4095 *) Mitigate branch prediction attacks, which can be practical if a
4096 single processor is shared, allowing a spy process to extract
4097 information. For detailed background information, see
4098 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4099 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4100 and Necessary Software Countermeasures"). The core of the change
4101 are new versions BN_div_no_branch() and
4102 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4103 respectively, which are slower, but avoid the security-relevant
4104 conditional branches. These are automatically called by BN_div()
4105 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4106 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4107 remove a conditional branch.
4109 BN_FLG_CONSTTIME is the new name for the previous
4110 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4111 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4112 in the exponent causes BN_mod_exp_mont() to use the alternative
4113 implementation in BN_mod_exp_mont_consttime().) The old name
4114 remains as a deprecated alias.
4116 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4117 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4118 constant-time implementations for more than just exponentiation.
4119 Here too the old name is kept as a deprecated alias.
4121 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4122 the BN_BLINDING structure gets an independent copy of the
4123 modulus. This means that the previous "BIGNUM *m" argument to
4124 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4125 essentially becomes "const BIGNUM *m", although we can't actually
4126 change this in the header file before 0.9.9. It allows
4127 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4128 enable BN_FLG_CONSTTIME.
4130 [Matthew D Wood (Intel Corp)]
4132 *) In the SSL/TLS server implementation, be strict about session ID
4133 context matching (which matters if an application uses a single
4134 external cache for different purposes). Previously,
4135 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4136 set. This did ensure strict client verification, but meant that,
4137 with applications using a single external cache for quite
4138 different requirements, clients could circumvent ciphersuite
4139 restrictions for a given session ID context by starting a session
4140 in a different context.
4143 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4144 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4145 authentication-only ciphersuites.
4148 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4149 not complete and could lead to a possible single byte overflow
4150 (CVE-2007-5135) [Ben Laurie]
4152 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4154 *) Since AES128 and AES256 (and similarly Camellia128 and
4155 Camellia256) share a single mask bit in the logic of
4156 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4157 kludge to work properly if AES128 is available and AES256 isn't
4158 (or if Camellia128 is available and Camellia256 isn't).
4161 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4162 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4163 When a point or a seed is encoded in a BIT STRING, we need to
4164 prevent the removal of trailing zero bits to get the proper DER
4165 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4166 of a NamedBitList, for which trailing 0 bits need to be removed.)
4169 *) Have SSL/TLS server implementation tolerate "mismatched" record
4170 protocol version while receiving ClientHello even if the
4171 ClientHello is fragmented. (The server can't insist on the
4172 particular protocol version it has chosen before the ServerHello
4173 message has informed the client about his choice.)
4176 *) Add RFC 3779 support.
4177 [Rob Austein for ARIN, Ben Laurie]
4179 *) Load error codes if they are not already present instead of using a
4180 static variable. This allows them to be cleanly unloaded and reloaded.
4181 Improve header file function name parsing.
4184 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4185 or CAPABILITY handshake as required by RFCs.
4188 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4190 *) Introduce limits to prevent malicious keys being able to
4191 cause a denial of service. (CVE-2006-2940)
4192 [Steve Henson, Bodo Moeller]
4194 *) Fix ASN.1 parsing of certain invalid structures that can result
4195 in a denial of service. (CVE-2006-2937) [Steve Henson]
4197 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4198 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4200 *) Fix SSL client code which could crash if connecting to a
4201 malicious SSLv2 server. (CVE-2006-4343)
4202 [Tavis Ormandy and Will Drewry, Google Security Team]
4204 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4205 match only those. Before that, "AES256-SHA" would be interpreted
4206 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4207 the same strength classification in 0.9.7h) as we currently only
4208 have a single AES bit in the ciphersuite description bitmap.
4209 That change, however, also applied to ciphersuite strings such as
4210 "RC4-MD5" that intentionally matched multiple ciphersuites --
4211 namely, SSL 2.0 ciphersuites in addition to the more common ones
4212 from SSL 3.0/TLS 1.0.
4214 So we change the selection algorithm again: Naming an explicit
4215 ciphersuite selects this one ciphersuite, and any other similar
4216 ciphersuite (same bitmap) from *other* protocol versions.
4217 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4218 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4220 Since SSL 2.0 does not have any ciphersuites for which the
4221 128/256 bit distinction would be relevant, this works for now.
4222 The proper fix will be to use different bits for AES128 and
4223 AES256, which would have avoided the problems from the beginning;
4224 however, bits are scarce, so we can only do this in a new release
4225 (not just a patchlevel) when we can change the SSL_CIPHER
4226 definition to split the single 'unsigned long mask' bitmap into
4227 multiple values to extend the available space.
4231 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4233 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4234 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4236 *) Add AES IGE and biIGE modes.
4239 *) Change the Unix randomness entropy gathering to use poll() when
4240 possible instead of select(), since the latter has some
4241 undesirable limitations.
4242 [Darryl Miles via Richard Levitte and Bodo Moeller]
4244 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4245 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4246 cannot be implicitly activated as part of, e.g., the "AES" alias.
4247 However, please upgrade to OpenSSL 0.9.9[-dev] for
4248 non-experimental use of the ECC ciphersuites to get TLS extension
4249 support, which is required for curve and point format negotiation
4250 to avoid potential handshake problems.
4253 *) Disable rogue ciphersuites:
4255 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4256 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4257 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4259 The latter two were purportedly from
4260 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4263 Also deactivate the remaining ciphersuites from
4264 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4265 unofficial, and the ID has long expired.
4268 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4269 dual-core machines) and other potential thread-safety issues.
4272 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4273 versions), which is now available for royalty-free use
4274 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4275 Also, add Camellia TLS ciphersuites from RFC 4132.
4277 To minimize changes between patchlevels in the OpenSSL 0.9.8
4278 series, Camellia remains excluded from compilation unless OpenSSL
4279 is configured with 'enable-camellia'.
4282 *) Disable the padding bug check when compression is in use. The padding
4283 bug check assumes the first packet is of even length, this is not
4284 necessarily true if compresssion is enabled and can result in false
4285 positives causing handshake failure. The actual bug test is ancient
4286 code so it is hoped that implementations will either have fixed it by
4287 now or any which still have the bug do not support compression.
4290 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4292 *) When applying a cipher rule check to see if string match is an explicit
4293 cipher suite and only match that one cipher suite if it is.
4296 *) Link in manifests for VC++ if needed.
4297 [Austin Ziegler <halostatue@gmail.com>]
4299 *) Update support for ECC-based TLS ciphersuites according to
4300 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4301 TLS extensions, which are supported starting with the 0.9.9
4302 branch, not in the OpenSSL 0.9.8 branch).
4305 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4306 opaque EVP_CIPHER_CTX handling.
4309 *) Fixes and enhancements to zlib compression code. We now only use
4310 "zlib1.dll" and use the default __cdecl calling convention on Win32
4311 to conform with the standards mentioned here:
4312 http://www.zlib.net/DLL_FAQ.txt
4313 Static zlib linking now works on Windows and the new --with-zlib-include
4314 --with-zlib-lib options to Configure can be used to supply the location
4315 of the headers and library. Gracefully handle case where zlib library
4319 *) Several fixes and enhancements to the OID generation code. The old code
4320 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4321 handle numbers larger than ULONG_MAX, truncated printing and had a
4322 non standard OBJ_obj2txt() behaviour.
4325 *) Add support for building of engines under engine/ as shared libraries
4326 under VC++ build system.
4329 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4330 Hopefully, we will not see any false combination of paths any more.
4333 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4335 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4336 (part of SSL_OP_ALL). This option used to disable the
4337 countermeasure against man-in-the-middle protocol-version
4338 rollback in the SSL 2.0 server implementation, which is a bad
4339 idea. (CVE-2005-2969)
4341 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4342 for Information Security, National Institute of Advanced Industrial
4343 Science and Technology [AIST], Japan)]
4345 *) Add two function to clear and return the verify parameter flags.
4348 *) Keep cipherlists sorted in the source instead of sorting them at
4349 runtime, thus removing the need for a lock.
4352 *) Avoid some small subgroup attacks in Diffie-Hellman.
4353 [Nick Mathewson and Ben Laurie]
4355 *) Add functions for well-known primes.
4358 *) Extended Windows CE support.
4359 [Satoshi Nakamura and Andy Polyakov]
4361 *) Initialize SSL_METHOD structures at compile time instead of during
4362 runtime, thus removing the need for a lock.
4365 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4366 attempting to decrypt each encrypted key in turn. Add support to
4370 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4372 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4375 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4378 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4379 key into the same file any more.
4382 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4385 *) Add -utf8 command line and config file option to 'ca'.
4386 [Stefan <stf@udoma.org]
4388 *) Removed the macro des_crypt(), as it seems to conflict with some
4389 libraries. Use DES_crypt().
4392 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4393 involves renaming the source and generated shared-libs for
4394 both. The engines will accept the corrected or legacy ids
4395 ('ncipher' and '4758_cca' respectively) when binding. NB,
4396 this only applies when building 'shared'.
4397 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4399 *) Add attribute functions to EVP_PKEY structure. Modify
4400 PKCS12_create() to recognize a CSP name attribute and
4401 use it. Make -CSP option work again in pkcs12 utility.
4404 *) Add new functionality to the bn blinding code:
4405 - automatic re-creation of the BN_BLINDING parameters after
4406 a fixed number of uses (currently 32)
4407 - add new function for parameter creation
4408 - introduce flags to control the update behaviour of the
4409 BN_BLINDING parameters
4410 - hide BN_BLINDING structure
4411 Add a second BN_BLINDING slot to the RSA structure to improve
4412 performance when a single RSA object is shared among several
4416 *) Add support for DTLS.
4417 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4419 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4420 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4423 *) Remove buggy and incomplete DH cert support from
4424 ssl/ssl_rsa.c and ssl/s3_both.c
4427 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4428 the apps/openssl applications.
4431 *) Compile clean with "-Wall -Wmissing-prototypes
4432 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4433 DEBUG_SAFESTACK must also be set.
4436 *) Change ./Configure so that certain algorithms can be disabled by default.
4437 The new counterpiece to "no-xxx" is "enable-xxx".
4439 The patented RC5 and MDC2 algorithms will now be disabled unless
4440 "enable-rc5" and "enable-mdc2", respectively, are specified.
4442 (IDEA remains enabled despite being patented. This is because IDEA
4443 is frequently required for interoperability, and there is no license
4444 fee for non-commercial use. As before, "no-idea" can be used to
4445 avoid this algorithm.)
4449 *) Add processing of proxy certificates (see RFC 3820). This work was
4450 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4451 EGEE (Enabling Grids for E-science in Europe).
4454 *) RC4 performance overhaul on modern architectures/implementations, such
4455 as Intel P4, IA-64 and AMD64.
4458 *) New utility extract-section.pl. This can be used specify an alternative
4459 section number in a pod file instead of having to treat each file as
4460 a separate case in Makefile. This can be done by adding two lines to the
4463 =for comment openssl_section:XXX
4465 The blank line is mandatory.
4469 *) New arguments -certform, -keyform and -pass for s_client and s_server
4470 to allow alternative format key and certificate files and passphrase
4474 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4475 update associated structures and add various utility functions.
4477 Add new policy related verify parameters, include policy checking in
4478 standard verify code. Enhance 'smime' application with extra parameters
4479 to support policy checking and print out.
4482 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4483 Nehemiah processors. These extensions support AES encryption in hardware
4484 as well as RNG (though RNG support is currently disabled).
4485 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4487 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4490 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4491 [Andy Polyakov and a number of other people]
4493 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4494 implementation contributed by IBM.
4495 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4497 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4498 exponent rather than 'unsigned long'. There is a corresponding change to
4499 the new 'rsa_keygen' element of the RSA_METHOD structure.
4500 [Jelte Jansen, Geoff Thorpe]
4502 *) Functionality for creating the initial serial number file is now
4503 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4505 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4506 number file to 1, which is bound to cause problems. To avoid
4507 the problems while respecting compatibility between different 0.9.7
4508 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4509 CA.pl for serial number initialization. With the new release 0.9.8,
4510 we can fix the problem directly in the 'ca' utility.)
4513 *) Reduced header interdepencies by declaring more opaque objects in
4514 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4515 give fewer recursive includes, which could break lazy source code - so
4516 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4517 developers should define this symbol when building and using openssl to
4518 ensure they track the recommended behaviour, interfaces, [etc], but
4519 backwards-compatible behaviour prevails when this isn't defined.
4522 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4525 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4526 This will generate a random key of the appropriate length based on the
4527 cipher context. The EVP_CIPHER can provide its own random key generation
4528 routine to support keys of a specific form. This is used in the des and
4529 3des routines to generate a key of the correct parity. Update S/MIME
4530 code to use new functions and hence generate correct parity DES keys.
4531 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4532 valid (weak or incorrect parity).
4535 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4536 as looking them up. This is useful when the verified structure may contain
4537 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4538 present unless the new PKCS7_NO_CRL flag is asserted.
4541 *) Extend ASN1 oid configuration module. It now additionally accepts the
4544 shortName = some long name, 1.2.3.4
4547 *) Reimplemented the BN_CTX implementation. There is now no more static
4548 limitation on the number of variables it can handle nor the depth of the
4549 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4550 information can now expand as required, and rather than having a single
4551 static array of bignums, BN_CTX now uses a linked-list of such arrays
4552 allowing it to expand on demand whilst maintaining the usefulness of
4553 BN_CTX's "bundling".
4556 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4557 to allow all RSA operations to function using a single BN_CTX.
4560 *) Preliminary support for certificate policy evaluation and checking. This
4561 is initially intended to pass the tests outlined in "Conformance Testing
4562 of Relying Party Client Certificate Path Processing Logic" v1.07.
4565 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4566 remained unused and not that useful. A variety of other little bignum
4567 tweaks and fixes have also been made continuing on from the audit (see
4571 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4572 associated ASN1, EVP and SSL functions and old ASN1 macros.
4575 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4576 and this should never fail. So the return value from the use of
4577 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4578 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4581 *) BN_CTX_get() should return zero-valued bignums, providing the same
4582 initialised value as BN_new().
4583 [Geoff Thorpe, suggested by Ulf Möller]
4585 *) Support for inhibitAnyPolicy certificate extension.
4588 *) An audit of the BIGNUM code is underway, for which debugging code is
4589 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4590 is considered valid when processing BIGNUMs, and causes execution to
4591 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4592 further steps are taken to deliberately pollute unused data in BIGNUM
4593 structures to try and expose faulty code further on. For now, openssl will
4594 (in its default mode of operation) continue to tolerate the inconsistent
4595 forms that it has tolerated in the past, but authors and packagers should
4596 consider trying openssl and their own applications when compiled with
4597 these debugging symbols defined. It will help highlight potential bugs in
4598 their own code, and will improve the test coverage for OpenSSL itself. At
4599 some point, these tighter rules will become openssl's default to improve
4600 maintainability, though the assert()s and other overheads will remain only
4601 in debugging configurations. See bn.h for more details.
4602 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4604 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4605 that can only be obtained through BN_CTX_new() (which implicitly
4606 initialises it). The presence of this function only made it possible
4607 to overwrite an existing structure (and cause memory leaks).
4610 *) Because of the callback-based approach for implementing LHASH as a
4611 template type, lh_insert() adds opaque objects to hash-tables and
4612 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4613 to clean up those corresponding objects before destroying the hash table
4614 (and losing the object pointers). So some over-zealous constifications in
4615 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4616 objects as "const" and the lh_doall[_arg] callback wrappers are not
4617 prototyped to have "const" restrictions on the object pointers they are
4618 given (and so aren't required to cast them away any more).
4621 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4622 (speed) prefers to use its own implementation. The two implementations
4623 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4624 its object type properly exposed (MS_TM) instead of casting to/from "char
4625 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4626 aren't necessarily the greatest nomenclatures - but this is what was used
4627 internally to the implementation so I've used that for now.
4630 *) Ensure that deprecated functions do not get compiled when
4631 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4632 the self-tests were still using deprecated key-generation functions so
4633 these have been updated also.
4636 *) Reorganise PKCS#7 code to separate the digest location functionality
4637 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4638 New function PKCS7_set_digest() to set the digest type for PKCS#7
4639 digestedData type. Add additional code to correctly generate the
4640 digestedData type and add support for this type in PKCS7 initialization
4644 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4645 structure of type "other".
4648 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4649 sure the loop does correctly stop and breaking ("division by zero")
4650 modulus operations are not performed. The (pre-generated) prime
4651 table crypto/bn/bn_prime.h was already correct, but it could not be
4652 re-generated on some platforms because of the "division by zero"
4653 situation in the script.
4654 [Ralf S. Engelschall]
4656 *) Update support for ECC-based TLS ciphersuites according to
4657 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4658 SHA-1 now is only used for "small" curves (where the
4659 representation of a field element takes up to 24 bytes); for
4660 larger curves, the field element resulting from ECDH is directly
4661 used as premaster secret.
4662 [Douglas Stebila (Sun Microsystems Laboratories)]
4664 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4665 curve secp160r1 to the tests.
4666 [Douglas Stebila (Sun Microsystems Laboratories)]
4668 *) Add the possibility to load symbols globally with DSO.
4669 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4671 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4672 control of the error stack.
4675 *) Add support for STORE in ENGINE.
4678 *) Add the STORE type. The intention is to provide a common interface
4679 to certificate and key stores, be they simple file-based stores, or
4680 HSM-type store, or LDAP stores, or...
4681 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4684 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4685 pass a list of arguments to any function as well as provide a way
4686 for a function to pass data back to the caller.
4689 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4690 works like BUF_strdup() but can be used to duplicate a portion of
4691 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4695 *) Add the function sk_find_ex() which works like sk_find(), but will
4696 return an index to an element even if an exact match couldn't be
4697 found. The index is guaranteed to point at the element where the
4698 searched-for key would be inserted to preserve sorting order.
4701 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4702 takes an extra flags argument for optional functionality. Currently,
4703 the following flags are defined:
4705 OBJ_BSEARCH_VALUE_ON_NOMATCH
4706 This one gets OBJ_bsearch_ex() to return a pointer to the first
4707 element where the comparing function returns a negative or zero
4710 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4711 This one gets OBJ_bsearch_ex() to return a pointer to the first
4712 element where the comparing function returns zero. This is useful
4713 if there are more than one element where the comparing function
4717 *) Make it possible to create self-signed certificates with 'openssl ca'
4718 in such a way that the self-signed certificate becomes part of the
4719 CA database and uses the same mechanisms for serial number generation
4720 as all other certificate signing. The new flag '-selfsign' enables
4721 this functionality. Adapt CA.sh and CA.pl.in.
4724 *) Add functionality to check the public key of a certificate request
4725 against a given private. This is useful to check that a certificate
4726 request can be signed by that key (self-signing).
4729 *) Make it possible to have multiple active certificates with the same
4730 subject in the CA index file. This is done only if the keyword
4731 'unique_subject' is set to 'no' in the main CA section (default
4732 if 'CA_default') of the configuration file. The value is saved
4733 with the database itself in a separate index attribute file,
4734 named like the index file with '.attr' appended to the name.
4737 *) Generate muti valued AVAs using '+' notation in config files for
4741 *) Support for nameConstraints certificate extension.
4744 *) Support for policyConstraints certificate extension.
4747 *) Support for policyMappings certificate extension.
4750 *) Make sure the default DSA_METHOD implementation only uses its
4751 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4752 and change its own handlers to be NULL so as to remove unnecessary
4753 indirection. This lets alternative implementations fallback to the
4754 default implementation more easily.
4757 *) Support for directoryName in GeneralName related extensions
4761 *) Make it possible to link applications using Makefile.shared.
4762 Make that possible even when linking against static libraries!
4765 *) Support for single pass processing for S/MIME signing. This now
4766 means that S/MIME signing can be done from a pipe, in addition
4767 cleartext signing (multipart/signed type) is effectively streaming
4768 and the signed data does not need to be all held in memory.
4770 This is done with a new flag PKCS7_STREAM. When this flag is set
4771 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4772 is done after the data is output (and digests calculated) in
4773 SMIME_write_PKCS7().
4776 *) Add full support for -rpath/-R, both in shared libraries and
4777 applications, at least on the platforms where it's known how
4781 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4782 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4783 will now compute a table of multiples of the generator that
4784 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4785 faster (notably in the case of a single point multiplication,
4786 scalar * generator).
4787 [Nils Larsch, Bodo Moeller]
4789 *) IPv6 support for certificate extensions. The various extensions
4790 which use the IP:a.b.c.d can now take IPv6 addresses using the
4791 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4795 *) Added an ENGINE that implements RSA by performing private key
4796 exponentiations with the GMP library. The conversions to and from
4797 GMP's mpz_t format aren't optimised nor are any montgomery forms
4798 cached, and on x86 it appears OpenSSL's own performance has caught up.
4799 However there are likely to be other architectures where GMP could
4800 provide a boost. This ENGINE is not built in by default, but it can be
4801 specified at Configure time and should be accompanied by the necessary
4802 linker additions, eg;
4803 ./config -DOPENSSL_USE_GMP -lgmp
4806 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4807 testing availability of engines with "-t" - the old behaviour is
4808 produced by increasing the feature's verbosity with "-tt".
4811 *) ECDSA routines: under certain error conditions uninitialized BN objects
4812 could be freed. Solution: make sure initialization is performed early
4813 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4817 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4818 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4819 software implementations. For DSA and DH, parameter generation can
4820 also be overriden by providing the appropriate method callbacks.
4823 *) Change the "progress" mechanism used in key-generation and
4824 primality testing to functions that take a new BN_GENCB pointer in
4825 place of callback/argument pairs. The new API functions have "_ex"
4826 postfixes and the older functions are reimplemented as wrappers for
4827 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4828 declarations of the old functions to help (graceful) attempts to
4829 migrate to the new functions. Also, the new key-generation API
4830 functions operate on a caller-supplied key-structure and return
4831 success/failure rather than returning a key or NULL - this is to
4832 help make "keygen" another member function of RSA_METHOD etc.
4834 Example for using the new callback interface:
4836 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4840 BN_GENCB_set(&my_cb, my_callback, my_arg);
4842 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4843 /* For the meaning of a, b in calls to my_callback(), see the
4844 * documentation of the function that calls the callback.
4845 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4846 * my_callback should return 1 if it wants BN_is_prime_ex()
4847 * to continue, or 0 to stop.
4852 *) Change the ZLIB compression method to be stateful, and make it
4853 available to TLS with the number defined in
4854 draft-ietf-tls-compression-04.txt.
4857 *) Add the ASN.1 structures and functions for CertificatePair, which
4858 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4860 CertificatePair ::= SEQUENCE {
4861 forward [0] Certificate OPTIONAL,
4862 reverse [1] Certificate OPTIONAL,
4863 -- at least one of the pair shall be present -- }
4865 Also implement the PEM functions to read and write certificate
4866 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4868 This needed to be defined, mostly for the sake of the LDAP
4869 attribute crossCertificatePair, but may prove useful elsewhere as
4873 *) Make it possible to inhibit symlinking of shared libraries in
4874 Makefile.shared, for Cygwin's sake.
4877 *) Extend the BIGNUM API by creating a function
4878 void BN_set_negative(BIGNUM *a, int neg);
4879 and a macro that behave like
4880 int BN_is_negative(const BIGNUM *a);
4882 to avoid the need to access 'a->neg' directly in applications.
4885 *) Implement fast modular reduction for pseudo-Mersenne primes
4886 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4887 EC_GROUP_new_curve_GFp() will now automatically use this
4889 [Nils Larsch <nla@trustcenter.de>]
4891 *) Add new lock type (CRYPTO_LOCK_BN).
4894 *) Change the ENGINE framework to automatically load engines
4895 dynamically from specific directories unless they could be
4896 found to already be built in or loaded. Move all the
4897 current engines except for the cryptodev one to a new
4899 The engines in engines/ are built as shared libraries if
4900 the "shared" options was given to ./Configure or ./config.
4901 Otherwise, they are inserted in libcrypto.a.
4902 /usr/local/ssl/engines is the default directory for dynamic
4903 engines, but that can be overriden at configure time through
4904 the usual use of --prefix and/or --openssldir, and at run
4905 time with the environment variable OPENSSL_ENGINES.
4906 [Geoff Thorpe and Richard Levitte]
4908 *) Add Makefile.shared, a helper makefile to build shared
4909 libraries. Adapt Makefile.org.
4912 *) Add version info to Win32 DLLs.
4913 [Peter 'Luna' Runestig" <peter@runestig.com>]
4915 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4916 can be added using this API to created arbitrary PKCS#12
4917 files while avoiding the low level API.
4919 New options to PKCS12_create(), key or cert can be NULL and
4920 will then be omitted from the output file. The encryption
4921 algorithm NIDs can be set to -1 for no encryption, the mac
4922 iteration count can be set to 0 to omit the mac.
4924 Enhance pkcs12 utility by making the -nokeys and -nocerts
4925 options work when creating a PKCS#12 file. New option -nomac
4926 to omit the mac, NONE can be set for an encryption algorithm.
4927 New code is modified to use the enhanced PKCS12_create()
4928 instead of the low level API.
4931 *) Extend ASN1 encoder to support indefinite length constructed
4932 encoding. This can output sequences tags and octet strings in
4933 this form. Modify pk7_asn1.c to support indefinite length
4934 encoding. This is experimental and needs additional code to
4935 be useful, such as an ASN1 bio and some enhanced streaming
4938 Extend template encode functionality so that tagging is passed
4939 down to the template encoder.
4942 *) Let 'openssl req' fail if an argument to '-newkey' is not
4943 recognized instead of using RSA as a default.
4946 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4947 As these are not official, they are not included in "ALL";
4948 the "ECCdraft" ciphersuite group alias can be used to select them.
4949 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4951 *) Add ECDH engine support.
4952 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4954 *) Add ECDH in new directory crypto/ecdh/.
4955 [Douglas Stebila (Sun Microsystems Laboratories)]
4957 *) Let BN_rand_range() abort with an error after 100 iterations
4958 without success (which indicates a broken PRNG).
4961 *) Change BN_mod_sqrt() so that it verifies that the input value
4962 is really the square of the return value. (Previously,
4963 BN_mod_sqrt would show GIGO behaviour.)
4966 *) Add named elliptic curves over binary fields from X9.62, SECG,
4967 and WAP/WTLS; add OIDs that were still missing.
4969 [Sheueling Chang Shantz and Douglas Stebila
4970 (Sun Microsystems Laboratories)]
4972 *) Extend the EC library for elliptic curves over binary fields
4973 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4976 EC_GF2m_simple_method
4980 EC_GROUP_new_curve_GF2m
4981 EC_GROUP_set_curve_GF2m
4982 EC_GROUP_get_curve_GF2m
4983 EC_POINT_set_affine_coordinates_GF2m
4984 EC_POINT_get_affine_coordinates_GF2m
4985 EC_POINT_set_compressed_coordinates_GF2m
4987 Point compression for binary fields is disabled by default for
4988 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4991 As binary polynomials are represented as BIGNUMs, various members
4992 of the EC_GROUP and EC_POINT data structures can be shared
4993 between the implementations for prime fields and binary fields;
4994 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4995 are essentially identical to their ..._GFp counterparts.
4996 (For simplicity, the '..._GFp' prefix has been dropped from
4997 various internal method names.)
4999 An internal 'field_div' method (similar to 'field_mul' and
5000 'field_sqr') has been added; this is used only for binary fields.
5002 [Sheueling Chang Shantz and Douglas Stebila
5003 (Sun Microsystems Laboratories)]
5005 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5006 through methods ('mul', 'precompute_mult').
5008 The generic implementations (now internally called 'ec_wNAF_mul'
5009 and 'ec_wNAF_precomputed_mult') remain the default if these
5010 methods are undefined.
5012 [Sheueling Chang Shantz and Douglas Stebila
5013 (Sun Microsystems Laboratories)]
5015 *) New function EC_GROUP_get_degree, which is defined through
5016 EC_METHOD. For curves over prime fields, this returns the bit
5017 length of the modulus.
5019 [Sheueling Chang Shantz and Douglas Stebila
5020 (Sun Microsystems Laboratories)]
5022 *) New functions EC_GROUP_dup, EC_POINT_dup.
5023 (These simply call ..._new and ..._copy).
5025 [Sheueling Chang Shantz and Douglas Stebila
5026 (Sun Microsystems Laboratories)]
5028 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5029 Polynomials are represented as BIGNUMs (where the sign bit is not
5030 used) in the following functions [macros]:
5033 BN_GF2m_sub [= BN_GF2m_add]
5034 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5035 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5036 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5038 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5039 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5040 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5041 BN_GF2m_cmp [= BN_ucmp]
5043 (Note that only the 'mod' functions are actually for fields GF(2^m).
5044 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5046 For some functions, an the irreducible polynomial defining a
5047 field can be given as an 'unsigned int[]' with strictly
5048 decreasing elements giving the indices of those bits that are set;
5049 i.e., p[] represents the polynomial
5050 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5052 p[0] > p[1] > ... > p[k] = 0.
5053 This applies to the following functions:
5058 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5059 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5061 BN_GF2m_mod_sqrt_arr
5062 BN_GF2m_mod_solve_quad_arr
5066 Conversion can be performed by the following functions:
5071 bntest.c has additional tests for binary polynomial arithmetic.
5073 Two implementations for BN_GF2m_mod_div() are available.
5074 The default algorithm simply uses BN_GF2m_mod_inv() and
5075 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5076 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5077 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5079 [Sheueling Chang Shantz and Douglas Stebila
5080 (Sun Microsystems Laboratories)]
5082 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5083 functionality is disabled at compile-time.
5084 [Douglas Stebila <douglas.stebila@sun.com>]
5086 *) Change default behaviour of 'openssl asn1parse' so that more
5087 information is visible when viewing, e.g., a certificate:
5089 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5090 mode the content of non-printable OCTET STRINGs is output in a
5091 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5092 avoid the appearance of a printable string.
5093 [Nils Larsch <nla@trustcenter.de>]
5095 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5097 EC_GROUP_set_asn1_flag()
5098 EC_GROUP_get_asn1_flag()
5099 EC_GROUP_set_point_conversion_form()
5100 EC_GROUP_get_point_conversion_form()
5101 These control ASN1 encoding details:
5102 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5103 has been set to OPENSSL_EC_NAMED_CURVE.
5104 - Points are encoded in uncompressed form by default; options for
5105 asn1_for are as for point2oct, namely
5106 POINT_CONVERSION_COMPRESSED
5107 POINT_CONVERSION_UNCOMPRESSED
5108 POINT_CONVERSION_HYBRID
5110 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5113 EC_GROUP_get0_seed()
5114 EC_GROUP_get_seed_len()
5115 This is used only for ASN1 purposes (so far).
5116 [Nils Larsch <nla@trustcenter.de>]
5118 *) Add 'field_type' member to EC_METHOD, which holds the NID
5119 of the appropriate field type OID. The new function
5120 EC_METHOD_get_field_type() returns this value.
5121 [Nils Larsch <nla@trustcenter.de>]
5126 EC_POINT_point2hex()
5127 EC_POINT_hex2point()
5128 providing useful interfaces to EC_POINT_point2oct() and
5129 EC_POINT_oct2point().
5130 [Nils Larsch <nla@trustcenter.de>]
5132 *) Change internals of the EC library so that the functions
5133 EC_GROUP_set_generator()
5134 EC_GROUP_get_generator()
5135 EC_GROUP_get_order()
5136 EC_GROUP_get_cofactor()
5137 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5138 to methods, which would lead to unnecessary code duplication when
5139 adding different types of curves.
5140 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5142 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5143 arithmetic, and such that modified wNAFs are generated
5144 (which avoid length expansion in many cases).
5147 *) Add a function EC_GROUP_check_discriminant() (defined via
5148 EC_METHOD) that verifies that the curve discriminant is non-zero.
5150 Add a function EC_GROUP_check() that makes some sanity tests
5151 on a EC_GROUP, its generator and order. This includes
5152 EC_GROUP_check_discriminant().
5153 [Nils Larsch <nla@trustcenter.de>]
5155 *) Add ECDSA in new directory crypto/ecdsa/.
5157 Add applications 'openssl ecparam' and 'openssl ecdsa'
5158 (these are based on 'openssl dsaparam' and 'openssl dsa').
5160 ECDSA support is also included in various other files across the
5161 library. Most notably,
5162 - 'openssl req' now has a '-newkey ecdsa:file' option;
5163 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5164 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5165 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5166 them suitable for ECDSA where domain parameters must be
5167 extracted before the specific public key;
5168 - ECDSA engine support has been added.
5169 [Nils Larsch <nla@trustcenter.de>]
5171 *) Include some named elliptic curves, and add OIDs from X9.62,
5172 SECG, and WAP/WTLS. Each curve can be obtained from the new
5174 EC_GROUP_new_by_curve_name(),
5175 and the list of available named curves can be obtained with
5176 EC_get_builtin_curves().
5177 Also add a 'curve_name' member to EC_GROUP objects, which can be
5179 EC_GROUP_set_curve_name()
5180 EC_GROUP_get_curve_name()
5181 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5183 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5184 was actually never needed) and in BN_mul(). The removal in BN_mul()
5185 required a small change in bn_mul_part_recursive() and the addition
5186 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5187 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5188 bn_sub_words() and bn_add_words() except they take arrays with
5192 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5194 *) Cleanse PEM buffers before freeing them since they may contain
5196 [Benjamin Bennett <ben@psc.edu>]
5198 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5199 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5200 authentication-only ciphersuites.
5203 *) Since AES128 and AES256 share a single mask bit in the logic of
5204 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5205 kludge to work properly if AES128 is available and AES256 isn't.
5208 *) Expand security boundary to match 1.1.1 module.
5211 *) Remove redundant features: hash file source, editing of test vectors
5212 modify fipsld to use external fips_premain.c signature.
5215 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5216 run algorithm test programs.
5219 *) Make algorithm test programs more tolerant of whitespace.
5222 *) Have SSL/TLS server implementation tolerate "mismatched" record
5223 protocol version while receiving ClientHello even if the
5224 ClientHello is fragmented. (The server can't insist on the
5225 particular protocol version it has chosen before the ServerHello
5226 message has informed the client about his choice.)
5229 *) Load error codes if they are not already present instead of using a
5230 static variable. This allows them to be cleanly unloaded and reloaded.
5233 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5235 *) Introduce limits to prevent malicious keys being able to
5236 cause a denial of service. (CVE-2006-2940)
5237 [Steve Henson, Bodo Moeller]
5239 *) Fix ASN.1 parsing of certain invalid structures that can result
5240 in a denial of service. (CVE-2006-2937) [Steve Henson]
5242 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5243 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5245 *) Fix SSL client code which could crash if connecting to a
5246 malicious SSLv2 server. (CVE-2006-4343)
5247 [Tavis Ormandy and Will Drewry, Google Security Team]
5249 *) Change ciphersuite string processing so that an explicit
5250 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5251 will no longer include "AES128-SHA"), and any other similar
5252 ciphersuite (same bitmap) from *other* protocol versions (so that
5253 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5254 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5255 changes from 0.9.8b and 0.9.8d.
5258 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5260 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5261 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5263 *) Change the Unix randomness entropy gathering to use poll() when
5264 possible instead of select(), since the latter has some
5265 undesirable limitations.
5266 [Darryl Miles via Richard Levitte and Bodo Moeller]
5268 *) Disable rogue ciphersuites:
5270 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5271 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5272 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5274 The latter two were purportedly from
5275 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5278 Also deactive the remaining ciphersuites from
5279 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5280 unofficial, and the ID has long expired.
5283 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5284 dual-core machines) and other potential thread-safety issues.
5287 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5289 *) Adapt fipsld and the build system to link against the validated FIPS
5290 module in FIPS mode.
5293 *) Fixes for VC++ 2005 build under Windows.
5296 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5297 from a Windows bash shell such as MSYS. It is autodetected from the
5298 "config" script when run from a VC++ environment. Modify standard VC++
5299 build to use fipscanister.o from the GNU make build.
5302 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5304 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5305 The value now differs depending on if you build for FIPS or not.
5306 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5307 safely run with a non-FIPSed libcrypto, as it may crash because of
5308 the difference induced by this change.
5311 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5313 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5314 (part of SSL_OP_ALL). This option used to disable the
5315 countermeasure against man-in-the-middle protocol-version
5316 rollback in the SSL 2.0 server implementation, which is a bad
5317 idea. (CVE-2005-2969)
5319 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5320 for Information Security, National Institute of Advanced Industrial
5321 Science and Technology [AIST], Japan)]
5323 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5324 mainly for FIPS compliance and not fully integrated at this stage.
5327 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5328 the exponentiation using a fixed-length exponent. (Otherwise,
5329 the information leaked through timing could expose the secret key
5330 after many signatures; cf. Bleichenbacher's attack on DSA with
5334 *) Make a new fixed-window mod_exp implementation the default for
5335 RSA, DSA, and DH private-key operations so that the sequence of
5336 squares and multiplies and the memory access pattern are
5337 independent of the particular secret key. This will mitigate
5338 cache-timing and potential related attacks.
5340 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5341 and this is automatically used by BN_mod_exp_mont() if the new flag
5342 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5343 will use this BN flag for private exponents unless the flag
5344 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5345 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5347 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5349 *) Change the client implementation for SSLv23_method() and
5350 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5351 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5352 (Previously, the SSL 2.0 backwards compatible Client Hello
5353 message format would be used even with SSL_OP_NO_SSLv2.)
5356 *) Add support for smime-type MIME parameter in S/MIME messages which some
5360 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5361 a threadsafe manner. Modify rsa code to use new function and add calls
5362 to dsa and dh code (which had race conditions before).
5365 *) Include the fixed error library code in the C error file definitions
5366 instead of fixing them up at runtime. This keeps the error code
5367 structures constant.
5370 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5372 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5375 *) Fixes for newer kerberos headers. NB: the casts are needed because
5376 the 'length' field is signed on one version and unsigned on another
5377 with no (?) obvious way to tell the difference, without these VC++
5378 complains. Also the "definition" of FAR (blank) is no longer included
5379 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5380 some needed definitions.
5383 *) Undo Cygwin change.
5386 *) Added support for proxy certificates according to RFC 3820.
5387 Because they may be a security thread to unaware applications,
5388 they must be explicitly allowed in run-time. See
5389 docs/HOWTO/proxy_certificates.txt for further information.
5392 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5394 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5395 server and client random values. Previously
5396 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5397 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5399 This change has negligible security impact because:
5401 1. Server and client random values still have 24 bytes of pseudo random
5404 2. Server and client random values are sent in the clear in the initial
5407 3. The master secret is derived using the premaster secret (48 bytes in
5408 size for static RSA ciphersuites) as well as client server and random
5411 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5414 [Stephen Henson, reported by UK NISCC]
5416 *) Use Windows randomness collection on Cygwin.
5419 *) Fix hang in EGD/PRNGD query when communication socket is closed
5420 prematurely by EGD/PRNGD.
5421 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5423 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5426 *) Back-port of selected performance improvements from development
5427 branch, as well as improved support for PowerPC platforms.
5430 *) Add lots of checks for memory allocation failure, error codes to indicate
5431 failure and freeing up memory if a failure occurs.
5432 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5434 *) Add new -passin argument to dgst.
5437 *) Perform some character comparisons of different types in X509_NAME_cmp:
5438 this is needed for some certificates that reencode DNs into UTF8Strings
5439 (in violation of RFC3280) and can't or wont issue name rollover
5443 *) Make an explicit check during certificate validation to see that
5444 the CA setting in each certificate on the chain is correct. As a
5445 side effect always do the following basic checks on extensions,
5446 not just when there's an associated purpose to the check:
5448 - if there is an unhandled critical extension (unless the user
5449 has chosen to ignore this fault)
5450 - if the path length has been exceeded (if one is set at all)
5451 - that certain extensions fit the associated purpose (if one has
5455 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5457 *) Avoid a race condition when CRLs are checked in a multi threaded
5458 environment. This would happen due to the reordering of the revoked
5459 entries during signature checking and serial number lookup. Now the
5460 encoding is cached and the serial number sort performed under a lock.
5461 Add new STACK function sk_is_sorted().
5464 *) Add Delta CRL to the extension code.
5467 *) Various fixes to s3_pkt.c so alerts are sent properly.
5468 [David Holmes <d.holmes@f5.com>]
5470 *) Reduce the chances of duplicate issuer name and serial numbers (in
5471 violation of RFC3280) using the OpenSSL certificate creation utilities.
5472 This is done by creating a random 64 bit value for the initial serial
5473 number when a serial number file is created or when a self signed
5474 certificate is created using 'openssl req -x509'. The initial serial
5475 number file is created using 'openssl x509 -next_serial' in CA.pl
5476 rather than being initialized to 1.
5479 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5481 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5482 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5483 [Joe Orton, Steve Henson]
5485 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5487 [Joe Orton, Steve Henson]
5489 *) Make it possible to have multiple active certificates with the same
5490 subject in the CA index file. This is done only if the keyword
5491 'unique_subject' is set to 'no' in the main CA section (default
5492 if 'CA_default') of the configuration file. The value is saved
5493 with the database itself in a separate index attribute file,
5494 named like the index file with '.attr' appended to the name.
5497 *) X509 verify fixes. Disable broken certificate workarounds when
5498 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5499 keyUsage extension present. Don't accept CRLs with unhandled critical
5500 extensions: since verify currently doesn't process CRL extensions this
5501 rejects a CRL with *any* critical extensions. Add new verify error codes
5505 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5506 A clarification of RFC2560 will require the use of OCTET STRINGs and
5507 some implementations cannot handle the current raw format. Since OpenSSL
5508 copies and compares OCSP nonces as opaque blobs without any attempt at
5509 parsing them this should not create any compatibility issues.
5512 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5513 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5514 this HMAC (and other) operations are several times slower than OpenSSL
5518 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5519 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5521 *) Use the correct content when signing type "other".
5524 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5526 *) Fix various bugs revealed by running the NISCC test suite:
5528 Stop out of bounds reads in the ASN1 code when presented with
5529 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5531 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5533 If verify callback ignores invalid public key errors don't try to check
5534 certificate signature with the NULL public key.
5538 *) New -ignore_err option in ocsp application to stop the server
5539 exiting on the first error in a request.
5542 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5543 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5547 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5548 extra data after the compression methods not only for TLS 1.0
5549 but also for SSL 3.0 (as required by the specification).
5550 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5552 *) Change X509_certificate_type() to mark the key as exported/exportable
5553 when it's 512 *bits* long, not 512 bytes.
5556 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5557 blocks during encryption.
5560 *) Various fixes to base64 BIO and non blocking I/O. On write
5561 flushes were not handled properly if the BIO retried. On read
5562 data was not being buffered properly and had various logic bugs.
5563 This also affects blocking I/O when the data being decoded is a
5567 *) Various S/MIME bugfixes and compatibility changes:
5568 output correct application/pkcs7 MIME type if
5569 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5570 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5571 of files as .eml work). Correctly handle very long lines in MIME
5575 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5577 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5578 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5579 a protocol version number mismatch like a decryption error
5580 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5583 *) Turn on RSA blinding by default in the default implementation
5584 to avoid a timing attack. Applications that don't want it can call
5585 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5586 They would be ill-advised to do so in most cases.
5587 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5589 *) Change RSA blinding code so that it works when the PRNG is not
5590 seeded (in this case, the secret RSA exponent is abused as
5591 an unpredictable seed -- if it is not unpredictable, there
5592 is no point in blinding anyway). Make RSA blinding thread-safe
5593 by remembering the creator's thread ID in rsa->blinding and
5594 having all other threads use local one-time blinding factors
5595 (this requires more computation than sharing rsa->blinding, but
5596 avoids excessive locking; and if an RSA object is not shared
5597 between threads, blinding will still be very fast).
5600 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5601 ENGINE as defaults for all supported algorithms irrespective of
5602 the 'flags' parameter. 'flags' is now honoured, so applications
5603 should make sure they are passing it correctly.
5606 *) Target "mingw" now allows native Windows code to be generated in
5607 the Cygwin environment as well as with the MinGW compiler.
5610 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5612 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5613 via timing by performing a MAC computation even if incorrect
5614 block cipher padding has been found. This is a countermeasure
5615 against active attacks where the attacker has to distinguish
5616 between bad padding and a MAC verification error. (CVE-2003-0078)
5618 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5619 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5620 Martin Vuagnoux (EPFL, Ilion)]
5622 *) Make the no-err option work as intended. The intention with no-err
5623 is not to have the whole error stack handling routines removed from
5624 libcrypto, it's only intended to remove all the function name and
5625 reason texts, thereby removing some of the footprint that may not
5626 be interesting if those errors aren't displayed anyway.
5628 NOTE: it's still possible for any application or module to have it's
5629 own set of error texts inserted. The routines are there, just not
5630 used by default when no-err is given.
5633 *) Add support for FreeBSD on IA64.
5634 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5636 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5637 Kerberos function mit_des_cbc_cksum(). Before this change,
5638 the value returned by DES_cbc_cksum() was like the one from
5639 mit_des_cbc_cksum(), except the bytes were swapped.
5640 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5642 *) Allow an application to disable the automatic SSL chain building.
5643 Before this a rather primitive chain build was always performed in
5644 ssl3_output_cert_chain(): an application had no way to send the
5645 correct chain if the automatic operation produced an incorrect result.
5647 Now the chain builder is disabled if either:
5649 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5651 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5653 The reasoning behind this is that an application would not want the
5654 auto chain building to take place if extra chain certificates are
5655 present and it might also want a means of sending no additional
5656 certificates (for example the chain has two certificates and the
5660 *) Add the possibility to build without the ENGINE framework.
5661 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5663 *) Under Win32 gmtime() can return NULL: check return value in
5664 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5667 *) DSA routines: under certain error conditions uninitialized BN objects
5668 could be freed. Solution: make sure initialization is performed early
5669 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5670 Nils Larsch <nla@trustcenter.de> via PR#459)
5673 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5674 checked on reconnect on the client side, therefore session resumption
5675 could still fail with a "ssl session id is different" error. This
5676 behaviour is masked when SSL_OP_ALL is used due to
5677 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5678 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5679 followup to PR #377.
5682 *) IA-32 assembler support enhancements: unified ELF targets, support
5683 for SCO/Caldera platforms, fix for Cygwin shared build.
5686 *) Add support for FreeBSD on sparc64. As a consequence, support for
5687 FreeBSD on non-x86 processors is separate from x86 processors on
5688 the config script, much like the NetBSD support.
5689 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5691 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5693 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5696 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5697 code (06) was taken as the first octet of the session ID and the last
5698 octet was ignored consequently. As a result SSLv2 client side session
5699 caching could not have worked due to the session ID mismatch between
5701 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5705 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5706 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5710 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5711 seems that in spite of existing for more than a year, many application
5712 author have done nothing to provide the necessary callbacks, which
5713 means that this particular engine will not work properly anywhere.
5714 This is a very unfortunate situation which forces us, in the name
5715 of usability, to give the hw_ncipher.c a static lock, which is part
5717 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5718 appear in 0.9.8 or later. We EXPECT application authors to have
5719 dealt properly with this when 0.9.8 is released (unless we actually
5720 make such changes in the libcrypto locking code that changes will
5721 have to be made anyway).
5724 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5725 octets have been read, EOF or an error occurs. Without this change
5726 some truncated ASN1 structures will not produce an error.
5729 *) Disable Heimdal support, since it hasn't been fully implemented.
5730 Still give the possibility to force the use of Heimdal, but with
5731 warnings and a request that patches get sent to openssl-dev.
5734 *) Add the VC-CE target, introduce the WINCE sysname, and add
5735 INSTALL.WCE and appropriate conditionals to make it build.
5736 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5738 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5739 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5740 edit numbers of the version.
5741 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5743 *) Introduce safe string copy and catenation functions
5744 (BUF_strlcpy() and BUF_strlcat()).
5745 [Ben Laurie (CHATS) and Richard Levitte]
5747 *) Avoid using fixed-size buffers for one-line DNs.
5748 [Ben Laurie (CHATS)]
5750 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5751 resizing buffers containing secrets, and use where appropriate.
5752 [Ben Laurie (CHATS)]
5754 *) Avoid using fixed size buffers for configuration file location.
5755 [Ben Laurie (CHATS)]
5757 *) Avoid filename truncation for various CA files.
5758 [Ben Laurie (CHATS)]
5760 *) Use sizeof in preference to magic numbers.
5761 [Ben Laurie (CHATS)]
5763 *) Avoid filename truncation in cert requests.
5764 [Ben Laurie (CHATS)]
5766 *) Add assertions to check for (supposedly impossible) buffer
5768 [Ben Laurie (CHATS)]
5770 *) Don't cache truncated DNS entries in the local cache (this could
5771 potentially lead to a spoofing attack).
5772 [Ben Laurie (CHATS)]
5774 *) Fix various buffers to be large enough for hex/decimal
5775 representations in a platform independent manner.
5776 [Ben Laurie (CHATS)]
5778 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5779 resizing buffers containing secrets, and use where appropriate.
5780 [Ben Laurie (CHATS)]
5782 *) Add BIO_indent() to avoid much slightly worrying code to do
5784 [Ben Laurie (CHATS)]
5786 *) Convert sprintf()/BIO_puts() to BIO_printf().
5787 [Ben Laurie (CHATS)]
5789 *) buffer_gets() could terminate with the buffer only half
5791 [Ben Laurie (CHATS)]
5793 *) Add assertions to prevent user-supplied crypto functions from
5794 overflowing internal buffers by having large block sizes, etc.
5795 [Ben Laurie (CHATS)]
5797 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5799 [Ben Laurie (CHATS)]
5801 *) Eliminate unused copy of key in RC4.
5802 [Ben Laurie (CHATS)]
5804 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5805 [Ben Laurie (CHATS)]
5807 *) Fix off-by-one error in EGD path.
5808 [Ben Laurie (CHATS)]
5810 *) If RANDFILE path is too long, ignore instead of truncating.
5811 [Ben Laurie (CHATS)]
5813 *) Eliminate unused and incorrectly sized X.509 structure
5815 [Ben Laurie (CHATS)]
5817 *) Eliminate unused and dangerous function knumber().
5818 [Ben Laurie (CHATS)]
5820 *) Eliminate unused and dangerous structure, KSSL_ERR.
5821 [Ben Laurie (CHATS)]
5823 *) Protect against overlong session ID context length in an encoded
5824 session object. Since these are local, this does not appear to be
5826 [Ben Laurie (CHATS)]
5828 *) Change from security patch (see 0.9.6e below) that did not affect
5829 the 0.9.6 release series:
5831 Remote buffer overflow in SSL3 protocol - an attacker could
5832 supply an oversized master key in Kerberos-enabled versions.
5834 [Ben Laurie (CHATS)]
5836 *) Change the SSL kerb5 codes to match RFC 2712.
5839 *) Make -nameopt work fully for req and add -reqopt switch.
5840 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5842 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5843 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5845 *) Make sure tests can be performed even if the corresponding algorithms
5846 have been removed entirely. This was also the last step to make
5847 OpenSSL compilable with DJGPP under all reasonable conditions.
5848 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5850 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5851 to allow version independent disabling of normally unselected ciphers,
5852 which may be activated as a side-effect of selecting a single cipher.
5854 (E.g., cipher list string "RSA" enables ciphersuites that are left
5855 out of "ALL" because they do not provide symmetric encryption.
5856 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5857 [Lutz Jaenicke, Bodo Moeller]
5859 *) Add appropriate support for separate platform-dependent build
5860 directories. The recommended way to make a platform-dependent
5861 build directory is the following (tested on Linux), maybe with
5864 # Place yourself outside of the OpenSSL source tree. In
5865 # this example, the environment variable OPENSSL_SOURCE
5866 # is assumed to contain the absolute OpenSSL source directory.
5867 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5868 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5869 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5870 mkdir -p `dirname $F`
5871 ln -s $OPENSSL_SOURCE/$F $F
5874 To be absolutely sure not to disturb the source tree, a "make clean"
5875 is a good thing. If it isn't successful, don't worry about it,
5876 it probably means the source directory is very clean.
5879 *) Make sure any ENGINE control commands make local copies of string
5880 pointers passed to them whenever necessary. Otherwise it is possible
5881 the caller may have overwritten (or deallocated) the original string
5882 data when a later ENGINE operation tries to use the stored values.
5883 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5885 *) Improve diagnostics in file reading and command-line digests.
5886 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5888 *) Add AES modes CFB and OFB to the object database. Correct an
5889 error in AES-CFB decryption.
5892 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5893 allows existing EVP_CIPHER_CTX structures to be reused after
5894 calling EVP_*Final(). This behaviour is used by encryption
5895 BIOs and some applications. This has the side effect that
5896 applications must explicitly clean up cipher contexts with
5897 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5900 *) Check the values of dna and dnb in bn_mul_recursive before calling
5901 bn_mul_comba (a non zero value means the a or b arrays do not contain
5902 n2 elements) and fallback to bn_mul_normal if either is not zero.
5905 *) Fix escaping of non-ASCII characters when using the -subj option
5906 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5909 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5910 form for "surname", serialNumber has no short form.
5911 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5912 therefore remove "mail" short name for "internet 7".
5913 The OID for unique identifiers in X509 certificates is
5914 x500UniqueIdentifier, not uniqueIdentifier.
5915 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5918 *) Add an "init" command to the ENGINE config module and auto initialize
5919 ENGINEs. Without any "init" command the ENGINE will be initialized
5920 after all ctrl commands have been executed on it. If init=1 the
5921 ENGINE is initailized at that point (ctrls before that point are run
5922 on the uninitialized ENGINE and after on the initialized one). If
5923 init=0 then the ENGINE will not be iniatialized at all.
5926 *) Fix the 'app_verify_callback' interface so that the user-defined
5927 argument is actually passed to the callback: In the
5928 SSL_CTX_set_cert_verify_callback() prototype, the callback
5929 declaration has been changed from
5932 int (*cb)(X509_STORE_CTX *,void *);
5933 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5934 i=s->ctx->app_verify_callback(&ctx)
5935 has been changed into
5936 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5938 To update applications using SSL_CTX_set_cert_verify_callback(),
5939 a dummy argument can be added to their callback functions.
5940 [D. K. Smetters <smetters@parc.xerox.com>]
5942 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5943 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5945 *) Add and OPENSSL_LOAD_CONF define which will cause
5946 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5947 This allows older applications to transparently support certain
5948 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5949 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5950 load the config file and OPENSSL_add_all_algorithms_conf() which will
5951 always load it have also been added.
5954 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5955 Adjust NIDs and EVP layer.
5956 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5958 *) Config modules support in openssl utility.
5960 Most commands now load modules from the config file,
5961 though in a few (such as version) this isn't done
5962 because it couldn't be used for anything.
5964 In the case of ca and req the config file used is
5965 the same as the utility itself: that is the -config
5966 command line option can be used to specify an
5970 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5971 use "openssl_conf" if filename is NULL use default openssl config file.
5974 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5975 config section name. Add a new flag to tolerate a missing config file
5976 and move code to CONF_modules_load_file().
5979 *) Support for crypto accelerator cards from Accelerated Encryption
5980 Processing, www.aep.ie. (Use engine 'aep')
5981 The support was copied from 0.9.6c [engine] and adapted/corrected
5982 to work with the new engine framework.
5983 [AEP Inc. and Richard Levitte]
5985 *) Support for SureWare crypto accelerator cards from Baltimore
5986 Technologies. (Use engine 'sureware')
5987 The support was copied from 0.9.6c [engine] and adapted
5988 to work with the new engine framework.
5991 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5992 make the newer ENGINE framework commands for the CHIL engine work.
5993 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5995 *) Make it possible to produce shared libraries on ReliantUNIX.
5996 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5998 *) Add the configuration target debug-linux-ppro.
5999 Make 'openssl rsa' use the general key loading routines
6000 implemented in apps.c, and make those routines able to
6001 handle the key format FORMAT_NETSCAPE and the variant
6003 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6005 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6006 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6008 *) Add -keyform to rsautl, and document -engine.
6009 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6011 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6012 BIO_R_NO_SUCH_FILE error code rather than the generic
6013 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6016 *) Add new functions
6018 ERR_peek_last_error_line
6019 ERR_peek_last_error_line_data.
6020 These are similar to
6023 ERR_peek_error_line_data,
6024 but report on the latest error recorded rather than the first one
6025 still in the error queue.
6026 [Ben Laurie, Bodo Moeller]
6028 *) default_algorithms option in ENGINE config module. This allows things
6030 default_algorithms = ALL
6031 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6034 *) Preliminary ENGINE config module.
6037 *) New experimental application configuration code.
6040 *) Change the AES code to follow the same name structure as all other
6041 symmetric ciphers, and behave the same way. Move everything to
6042 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6043 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6045 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6046 [Ben Laurie and Theo de Raadt]
6048 *) Add option to output public keys in req command.
6049 [Massimiliano Pala madwolf@openca.org]
6051 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6052 (up to about 10% better than before for P-192 and P-224).
6055 *) New functions/macros
6057 SSL_CTX_set_msg_callback(ctx, cb)
6058 SSL_CTX_set_msg_callback_arg(ctx, arg)
6059 SSL_set_msg_callback(ssl, cb)
6060 SSL_set_msg_callback_arg(ssl, arg)
6062 to request calling a callback function
6064 void cb(int write_p, int version, int content_type,
6065 const void *buf, size_t len, SSL *ssl, void *arg)
6067 whenever a protocol message has been completely received
6068 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6069 protocol version according to which the SSL library interprets
6070 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6071 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6072 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6073 specification (change_cipher_spec(20), alert(21), handshake(22)).
6074 'buf' and 'len' point to the actual message, 'ssl' to the
6075 SSL object, and 'arg' is the application-defined value set by
6076 SSL[_CTX]_set_msg_callback_arg().
6078 'openssl s_client' and 'openssl s_server' have new '-msg' options
6079 to enable a callback that displays all protocol messages.
6082 *) Change the shared library support so shared libraries are built as
6083 soon as the corresponding static library is finished, and thereby get
6084 openssl and the test programs linked against the shared library.
6085 This still only happens when the keyword "shard" has been given to
6086 the configuration scripts.
6088 NOTE: shared library support is still an experimental thing, and
6089 backward binary compatibility is still not guaranteed.
6090 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6092 *) Add support for Subject Information Access extension.
6093 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6095 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6096 additional bytes when new memory had to be allocated, not just
6097 when reusing an existing buffer.
6100 *) New command line and configuration option 'utf8' for the req command.
6101 This allows field values to be specified as UTF8 strings.
6104 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6105 runs for the former and machine-readable output for the latter.
6108 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6109 of the e-mail address in the DN (i.e., it will go into a certificate
6110 extension only). The new configuration file option 'email_in_dn = no'
6111 has the same effect.
6112 [Massimiliano Pala madwolf@openca.org]
6114 *) Change all functions with names starting with des_ to be starting
6115 with DES_ instead. Add wrappers that are compatible with libdes,
6116 but are named _ossl_old_des_*. Finally, add macros that map the
6117 des_* symbols to the corresponding _ossl_old_des_* if libdes
6118 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6119 desired, the des_* symbols will be mapped to DES_*, with one
6122 Since we provide two compatibility mappings, the user needs to
6123 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6124 compatibility is desired. The default (i.e., when that macro
6125 isn't defined) is OpenSSL 0.9.6c compatibility.
6127 There are also macros that enable and disable the support of old
6128 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6129 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6130 are defined, the default will apply: to support the old des routines.
6132 In either case, one must include openssl/des.h to get the correct
6133 definitions. Do not try to just include openssl/des_old.h, that
6136 NOTE: This is a major break of an old API into a new one. Software
6137 authors are encouraged to switch to the DES_ style functions. Some
6138 time in the future, des_old.h and the libdes compatibility functions
6139 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6140 default), and then completely removed.
6143 *) Test for certificates which contain unsupported critical extensions.
6144 If such a certificate is found during a verify operation it is
6145 rejected by default: this behaviour can be overridden by either
6146 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6147 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6148 X509_supported_extension() has also been added which returns 1 if a
6149 particular extension is supported.
6152 *) Modify the behaviour of EVP cipher functions in similar way to digests
6153 to retain compatibility with existing code.
6156 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6157 compatibility with existing code. In particular the 'ctx' parameter does
6158 not have to be to be initialized before the call to EVP_DigestInit() and
6159 it is tidied up after a call to EVP_DigestFinal(). New function
6160 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6161 EVP_MD_CTX_copy() changed to not require the destination to be
6162 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6163 requires the destination to be valid.
6165 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6166 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6169 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6170 so that complete 'Handshake' protocol structures are kept in memory
6171 instead of overwriting 'msg_type' and 'length' with 'body' data.
6174 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6175 [Massimo Santin via Richard Levitte]
6177 *) Major restructuring to the underlying ENGINE code. This includes
6178 reduction of linker bloat, separation of pure "ENGINE" manipulation
6179 (initialisation, etc) from functionality dealing with implementations
6180 of specific crypto iterfaces. This change also introduces integrated
6181 support for symmetric ciphers and digest implementations - so ENGINEs
6182 can now accelerate these by providing EVP_CIPHER and EVP_MD
6183 implementations of their own. This is detailed in crypto/engine/README
6184 as it couldn't be adequately described here. However, there are a few
6185 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6186 were changed in the original introduction of ENGINE code have now
6187 reverted back - the hooking from this code to ENGINE is now a good
6188 deal more passive and at run-time, operations deal directly with
6189 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6190 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6191 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6192 they were not being used by the framework as there is no concept of a
6193 BIGNUM_METHOD and they could not be generalised to the new
6194 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6195 ENGINE_cpy() has been removed as it cannot be consistently defined in
6199 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6202 *) Change mkdef.pl to sort symbols that get the same entry number,
6203 and make sure the automatically generated functions ERR_load_*
6204 become part of libeay.num as well.
6207 *) New function SSL_renegotiate_pending(). This returns true once
6208 renegotiation has been requested (either SSL_renegotiate() call
6209 or HelloRequest/ClientHello received from the peer) and becomes
6210 false once a handshake has been completed.
6211 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6212 sends a HelloRequest, but does not ensure that a handshake takes
6213 place. SSL_renegotiate_pending() is useful for checking if the
6214 client has followed the request.)
6217 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6218 By default, clients may request session resumption even during
6219 renegotiation (if session ID contexts permit); with this option,
6220 session resumption is possible only in the first handshake.
6222 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6223 more bits available for options that should not be part of
6224 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6227 *) Add some demos for certificate and certificate request creation.
6230 *) Make maximum certificate chain size accepted from the peer application
6231 settable (SSL*_get/set_max_cert_list()), as proposed by
6232 "Douglas E. Engert" <deengert@anl.gov>.
6235 *) Add support for shared libraries for Unixware-7
6236 (Boyd Lynn Gerber <gerberb@zenez.com>).
6239 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6240 be done prior to destruction. Use this to unload error strings from
6241 ENGINEs that load their own error strings. NB: This adds two new API
6242 functions to "get" and "set" this destroy handler in an ENGINE.
6245 *) Alter all existing ENGINE implementations (except "openssl" and
6246 "openbsd") to dynamically instantiate their own error strings. This
6247 makes them more flexible to be built both as statically-linked ENGINEs
6248 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6249 Also, add stub code to each that makes building them as self-contained
6250 shared-libraries easier (see README.ENGINE).
6253 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6254 implementations into applications that are completely implemented in
6255 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6256 commands that can be used to configure what shared-library to load and
6257 to control aspects of the way it is handled. Also, made an update to
6258 the README.ENGINE file that brings its information up-to-date and
6259 provides some information and instructions on the "dynamic" ENGINE
6260 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6263 *) Make it possible to unload ranges of ERR strings with a new
6264 "ERR_unload_strings" function.
6267 *) Add a copy() function to EVP_MD.
6270 *) Make EVP_MD routines take a context pointer instead of just the
6271 md_data void pointer.
6274 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6275 that the digest can only process a single chunk of data
6276 (typically because it is provided by a piece of
6277 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6278 is only going to provide a single chunk of data, and hence the
6279 framework needn't accumulate the data for oneshot drivers.
6282 *) As with "ERR", make it possible to replace the underlying "ex_data"
6283 functions. This change also alters the storage and management of global
6284 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6285 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6286 index counters. The API functions that use this state have been changed
6287 to take a "class_index" rather than pointers to the class's local STACK
6288 and counter, and there is now an API function to dynamically create new
6289 classes. This centralisation allows us to (a) plug a lot of the
6290 thread-safety problems that existed, and (b) makes it possible to clean
6291 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6292 such data would previously have always leaked in application code and
6293 workarounds were in place to make the memory debugging turn a blind eye
6294 to it. Application code that doesn't use this new function will still
6295 leak as before, but their memory debugging output will announce it now
6296 rather than letting it slide.
6298 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6299 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6300 has a return value to indicate success or failure.
6303 *) Make it possible to replace the underlying "ERR" functions such that the
6304 global state (2 LHASH tables and 2 locks) is only used by the "default"
6305 implementation. This change also adds two functions to "get" and "set"
6306 the implementation prior to it being automatically set the first time
6307 any other ERR function takes place. Ie. an application can call "get",
6308 pass the return value to a module it has just loaded, and that module
6309 can call its own "set" function using that value. This means the
6310 module's "ERR" operations will use (and modify) the error state in the
6311 application and not in its own statically linked copy of OpenSSL code.
6314 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6315 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6316 the operation, and provides a more encapsulated way for external code
6317 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6318 to use these functions rather than manually incrementing the counts.
6320 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6323 *) Add EVP test program.
6326 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6329 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6330 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6331 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6332 These allow a CRL to be built without having to access X509_CRL fields
6333 directly. Modify 'ca' application to use new functions.
6336 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6337 bug workarounds. Rollback attack detection is a security feature.
6338 The problem will only arise on OpenSSL servers when TLSv1 is not
6339 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6340 Software authors not wanting to support TLSv1 will have special reasons
6341 for their choice and can explicitly enable this option.
6342 [Bodo Moeller, Lutz Jaenicke]
6344 *) Rationalise EVP so it can be extended: don't include a union of
6345 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6346 (similar to those existing for EVP_CIPHER_CTX).
6351 EVP_MD_CTX_init(&md); /* new function call */
6352 EVP_DigestInit(&md, EVP_sha1());
6353 EVP_DigestUpdate(&md, in, len);
6354 EVP_DigestFinal(&md, out, NULL);
6355 EVP_MD_CTX_cleanup(&md); /* new function call */
6359 *) Make DES key schedule conform to the usual scheme, as well as
6360 correcting its structure. This means that calls to DES functions
6361 now have to pass a pointer to a des_key_schedule instead of a
6362 plain des_key_schedule (which was actually always a pointer
6365 des_key_schedule ks;
6367 des_set_key_checked(..., &ks);
6368 des_ncbc_encrypt(..., &ks, ...);
6370 (Note that a later change renames 'des_...' into 'DES_...'.)
6373 *) Initial reduction of linker bloat: the use of some functions, such as
6374 PEM causes large amounts of unused functions to be linked in due to
6375 poor organisation. For example pem_all.c contains every PEM function
6376 which has a knock on effect of linking in large amounts of (unused)
6377 ASN1 code. Grouping together similar functions and splitting unrelated
6378 functions prevents this.
6381 *) Cleanup of EVP macros.
6384 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6385 correct _ecb suffix.
6388 *) Add initial OCSP responder support to ocsp application. The
6389 revocation information is handled using the text based index
6390 use by the ca application. The responder can either handle
6391 requests generated internally, supplied in files (for example
6392 via a CGI script) or using an internal minimal server.
6395 *) Add configuration choices to get zlib compression for TLS.
6398 *) Changes to Kerberos SSL for RFC 2712 compliance:
6399 1. Implemented real KerberosWrapper, instead of just using
6400 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6401 2. Implemented optional authenticator field of KerberosWrapper.
6403 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6404 and authenticator structs; see crypto/krb5/.
6406 Generalized Kerberos calls to support multiple Kerberos libraries.
6407 [Vern Staats <staatsvr@asc.hpc.mil>,
6408 Jeffrey Altman <jaltman@columbia.edu>
6409 via Richard Levitte]
6411 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6412 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6413 values for each of the key sizes rather than having just
6414 parameters (and 'speed' generating keys each time).
6417 *) Speed up EVP routines.
6420 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6421 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6422 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6423 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6425 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6426 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6427 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6430 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6432 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6435 *) Added the OS2-EMX target.
6436 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6438 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6439 to support NCONF routines in extension code. New function CONF_set_nconf()
6440 to allow functions which take an NCONF to also handle the old LHASH
6441 structure: this means that the old CONF compatible routines can be
6442 retained (in particular wrt extensions) without having to duplicate the
6443 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6446 *) Enhance the general user interface with mechanisms for inner control
6447 and with possibilities to have yes/no kind of prompts.
6450 *) Change all calls to low level digest routines in the library and
6451 applications to use EVP. Add missing calls to HMAC_cleanup() and
6452 don't assume HMAC_CTX can be copied using memcpy().
6453 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6455 *) Add the possibility to control engines through control names but with
6456 arbitrary arguments instead of just a string.
6457 Change the key loaders to take a UI_METHOD instead of a callback
6458 function pointer. NOTE: this breaks binary compatibility with earlier
6459 versions of OpenSSL [engine].
6460 Adapt the nCipher code for these new conditions and add a card insertion
6464 *) Enhance the general user interface with mechanisms to better support
6465 dialog box interfaces, application-defined prompts, the possibility
6466 to use defaults (for example default passwords from somewhere else)
6467 and interrupts/cancellations.
6470 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6471 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6474 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6475 tidy up some unnecessarily weird code in 'sk_new()').
6476 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6478 *) Change the key loading routines for ENGINEs to use the same kind
6479 callback (pem_password_cb) as all other routines that need this
6483 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6484 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6485 than this minimum value is recommended.
6488 *) New random seeder for OpenVMS, using the system process statistics
6489 that are easily reachable.
6492 *) Windows apparently can't transparently handle global
6493 variables defined in DLLs. Initialisations such as:
6495 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6497 wont compile. This is used by the any applications that need to
6498 declare their own ASN1 modules. This was fixed by adding the option
6499 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6500 needed for static libraries under Win32.
6503 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6504 setting of purpose and trust fields. New X509_STORE trust and
6505 purpose functions and tidy up setting in other SSL functions.
6508 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6509 structure. These are inherited by X509_STORE_CTX when it is
6510 initialised. This allows various defaults to be set in the
6511 X509_STORE structure (such as flags for CRL checking and custom
6512 purpose or trust settings) for functions which only use X509_STORE_CTX
6513 internally such as S/MIME.
6515 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6516 trust settings if they are not set in X509_STORE. This allows X509_STORE
6517 purposes and trust (in S/MIME for example) to override any set by default.
6519 Add command line options for CRL checking to smime, s_client and s_server
6523 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6524 are set then the CRL is looked up in the X509_STORE structure and
6525 its validity and signature checked, then if the certificate is found
6526 in the CRL the verify fails with a revoked error.
6528 Various new CRL related callbacks added to X509_STORE_CTX structure.
6530 Command line options added to 'verify' application to support this.
6532 This needs some additional work, such as being able to handle multiple
6533 CRLs with different times, extension based lookup (rather than just
6534 by subject name) and ultimately more complete V2 CRL extension
6538 *) Add a general user interface API (crypto/ui/). This is designed
6539 to replace things like des_read_password and friends (backward
6540 compatibility functions using this new API are provided).
6541 The purpose is to remove prompting functions from the DES code
6542 section as well as provide for prompting through dialog boxes in
6543 a window system and the like.
6546 *) Add "ex_data" support to ENGINE so implementations can add state at a
6547 per-structure level rather than having to store it globally.
6550 *) Make it possible for ENGINE structures to be copied when retrieved by
6551 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6552 This causes the "original" ENGINE structure to act like a template,
6553 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6554 operational state can be localised to each ENGINE structure, despite the
6555 fact they all share the same "methods". New ENGINE structures returned in
6556 this case have no functional references and the return value is the single
6557 structural reference. This matches the single structural reference returned
6558 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6562 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6563 needs to match any other type at all we need to manually clear the
6567 *) Changes to the "openssl engine" utility to include;
6568 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6569 about an ENGINE's available control commands.
6570 - executing control commands from command line arguments using the
6571 '-pre' and '-post' switches. '-post' is only used if '-t' is
6572 specified and the ENGINE is successfully initialised. The syntax for
6573 the individual commands are colon-separated, for example;
6574 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6577 *) New dynamic control command support for ENGINEs. ENGINEs can now
6578 declare their own commands (numbers), names (strings), descriptions,
6579 and input types for run-time discovery by calling applications. A
6580 subset of these commands are implicitly classed as "executable"
6581 depending on their input type, and only these can be invoked through
6582 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6583 can be based on user input, config files, etc). The distinction is
6584 that "executable" commands cannot return anything other than a boolean
6585 result and can only support numeric or string input, whereas some
6586 discoverable commands may only be for direct use through
6587 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6588 pointers, or other custom uses. The "executable" commands are to
6589 support parameterisations of ENGINE behaviour that can be
6590 unambiguously defined by ENGINEs and used consistently across any
6591 OpenSSL-based application. Commands have been added to all the
6592 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6593 control over shared-library paths without source code alterations.
6596 *) Changed all ENGINE implementations to dynamically allocate their
6597 ENGINEs rather than declaring them statically. Apart from this being
6598 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6599 this also allows the implementations to compile without using the
6600 internal engine_int.h header.
6603 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6604 'const' value. Any code that should be able to modify a RAND_METHOD
6605 should already have non-const pointers to it (ie. they should only
6606 modify their own ones).
6609 *) Made a variety of little tweaks to the ENGINE code.
6610 - "atalla" and "ubsec" string definitions were moved from header files
6611 to C code. "nuron" string definitions were placed in variables
6612 rather than hard-coded - allowing parameterisation of these values
6613 later on via ctrl() commands.
6614 - Removed unused "#if 0"'d code.
6615 - Fixed engine list iteration code so it uses ENGINE_free() to release
6616 structural references.
6617 - Constified the RAND_METHOD element of ENGINE structures.
6618 - Constified various get/set functions as appropriate and added
6619 missing functions (including a catch-all ENGINE_cpy that duplicates
6620 all ENGINE values onto a new ENGINE except reference counts/state).
6621 - Removed NULL parameter checks in get/set functions. Setting a method
6622 or function to NULL is a way of cancelling out a previously set
6623 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6624 and doesn't justify the extra error symbols and code.
6625 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6626 flags from engine_int.h to engine.h.
6627 - Changed prototypes for ENGINE handler functions (init(), finish(),
6628 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6631 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6632 to the algorithm using long division. The binary algorithm can be
6633 used only if the modulus is odd. On 32-bit systems, it is faster
6634 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6635 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6636 up to 450 bits. In 64-bit environments, the binary algorithm
6637 appears to be advantageous for much longer moduli; here we use it
6638 for moduli up to 2048 bits.
6641 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6642 could not support the combine flag in choice fields.
6645 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6646 extensions from a certificate request to the certificate.
6649 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6650 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6651 file: this allows the display of the certificate about to be
6652 signed to be customised, to allow certain fields to be included
6653 or excluded and extension details. The old system didn't display
6654 multicharacter strings properly, omitted fields not in the policy
6655 and couldn't display additional details such as extensions.
6658 *) Function EC_POINTs_mul for multiple scalar multiplication
6659 of an arbitrary number of elliptic curve points
6660 \sum scalars[i]*points[i],
6661 optionally including the generator defined for the EC_GROUP:
6662 scalar*generator + \sum scalars[i]*points[i].
6664 EC_POINT_mul is a simple wrapper function for the typical case
6665 that the point list has just one item (besides the optional
6669 *) First EC_METHODs for curves over GF(p):
6671 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6672 operations and provides various method functions that can also
6673 operate with faster implementations of modular arithmetic.
6675 EC_GFp_mont_method() reuses most functions that are part of
6676 EC_GFp_simple_method, but uses Montgomery arithmetic.
6678 [Bodo Moeller; point addition and point doubling
6679 implementation directly derived from source code provided by
6680 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6682 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6683 crypto/ec/ec_lib.c):
6685 Curves are EC_GROUP objects (with an optional group generator)
6686 based on EC_METHODs that are built into the library.
6688 Points are EC_POINT objects based on EC_GROUP objects.
6690 Most of the framework would be able to handle curves over arbitrary
6691 finite fields, but as there are no obvious types for fields other
6692 than GF(p), some functions are limited to that for now.
6695 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6696 that the file contains a complete HTTP response.
6699 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6700 change the def and num file printf format specifier from "%-40sXXX"
6701 to "%-39s XXX". The latter will always guarantee a space after the
6702 field while the former will cause them to run together if the field
6703 is 40 of more characters long.
6706 *) Constify the cipher and digest 'method' functions and structures
6707 and modify related functions to take constant EVP_MD and EVP_CIPHER
6711 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6712 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6715 *) Modify EVP_Digest*() routines so they now return values. Although the
6716 internal software routines can never fail additional hardware versions
6720 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6722 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6723 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6726 ERR_R_NESTED_ASN1_ERROR
6728 ERR_R_MISSING_ASN1_EOS
6729 were 4 .. 9, conflicting with
6730 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6732 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6733 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6735 Add new error code 'ERR_R_INTERNAL_ERROR'.
6738 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6742 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6743 sets the subject name for a new request or supersedes the
6744 subject name in a given request. Formats that can be parsed are
6745 'CN=Some Name, OU=myOU, C=IT'
6747 'CN=Some Name/OU=myOU/C=IT'.
6749 Add options '-batch' and '-verbose' to 'openssl req'.
6750 [Massimiliano Pala <madwolf@hackmasters.net>]
6752 *) Introduce the possibility to access global variables through
6753 functions on platform were that's the best way to handle exporting
6754 global variables in shared libraries. To enable this functionality,
6755 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6756 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6757 is normally done by Configure or something similar).
6759 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6760 in the source file (foo.c) like this:
6762 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6763 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6765 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6766 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6768 OPENSSL_DECLARE_GLOBAL(int,foo);
6769 #define foo OPENSSL_GLOBAL_REF(foo)
6770 OPENSSL_DECLARE_GLOBAL(double,bar);
6771 #define bar OPENSSL_GLOBAL_REF(bar)
6773 The #defines are very important, and therefore so is including the
6774 header file everywhere where the defined globals are used.
6776 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6777 of ASN.1 items, but that structure is a bit different.
6779 The largest change is in util/mkdef.pl which has been enhanced with
6780 better and easier to understand logic to choose which symbols should
6781 go into the Windows .def files as well as a number of fixes and code
6782 cleanup (among others, algorithm keywords are now sorted
6783 lexicographically to avoid constant rewrites).
6786 *) In BN_div() keep a copy of the sign of 'num' before writing the
6787 result to 'rm' because if rm==num the value will be overwritten
6788 and produce the wrong result if 'num' is negative: this caused
6789 problems with BN_mod() and BN_nnmod().
6792 *) Function OCSP_request_verify(). This checks the signature on an
6793 OCSP request and verifies the signer certificate. The signer
6794 certificate is just checked for a generic purpose and OCSP request
6798 *) Add OCSP_check_validity() function to check the validity of OCSP
6799 responses. OCSP responses are prepared in real time and may only
6800 be a few seconds old. Simply checking that the current time lies
6801 between thisUpdate and nextUpdate max reject otherwise valid responses
6802 caused by either OCSP responder or client clock inaccuracy. Instead
6803 we allow thisUpdate and nextUpdate to fall within a certain period of
6804 the current time. The age of the response can also optionally be
6805 checked. Two new options -validity_period and -status_age added to
6809 *) If signature or public key algorithm is unrecognized print out its
6810 OID rather that just UNKNOWN.
6813 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6814 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6815 ID to be generated from the issuer certificate alone which can then be
6816 passed to OCSP_id_issuer_cmp().
6819 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6820 ASN1 modules to export functions returning ASN1_ITEM pointers
6821 instead of the ASN1_ITEM structures themselves. This adds several
6822 new macros which allow the underlying ASN1 function/structure to
6823 be accessed transparently. As a result code should not use ASN1_ITEM
6824 references directly (such as &X509_it) but instead use the relevant
6825 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6826 use of the new ASN1 code on platforms where exporting structures
6827 is problematical (for example in shared libraries) but exporting
6828 functions returning pointers to structures is not.
6831 *) Add support for overriding the generation of SSL/TLS session IDs.
6832 These callbacks can be registered either in an SSL_CTX or per SSL.
6833 The purpose of this is to allow applications to control, if they wish,
6834 the arbitrary values chosen for use as session IDs, particularly as it
6835 can be useful for session caching in multiple-server environments. A
6836 command-line switch for testing this (and any client code that wishes
6837 to use such a feature) has been added to "s_server".
6838 [Geoff Thorpe, Lutz Jaenicke]
6840 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6841 of the form '#if defined(...) || defined(...) || ...' and
6842 '#if !defined(...) && !defined(...) && ...'. This also avoids
6843 the growing number of special cases it was previously handling.
6846 *) Make all configuration macros available for application by making
6847 sure they are available in opensslconf.h, by giving them names starting
6848 with "OPENSSL_" to avoid conflicts with other packages and by making
6849 sure e_os2.h will cover all platform-specific cases together with
6851 Additionally, it is now possible to define configuration/platform-
6852 specific names (called "system identities"). In the C code, these
6853 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6854 macro with the name beginning with "OPENSSL_SYS_", which is determined
6855 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6859 *) New option -set_serial to 'req' and 'x509' this allows the serial
6860 number to use to be specified on the command line. Previously self
6861 signed certificates were hard coded with serial number 0 and the
6862 CA options of 'x509' had to use a serial number in a file which was
6866 *) New options to 'ca' utility to support V2 CRL entry extensions.
6867 Currently CRL reason, invalidity date and hold instruction are
6868 supported. Add new CRL extensions to V3 code and some new objects.
6871 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6872 disable standard block padding (aka PKCS#5 padding) in the EVP
6873 API, which was previously mandatory. This means that the data is
6874 not padded in any way and so the total length much be a multiple
6875 of the block size, otherwise an error occurs.
6878 *) Initial (incomplete) OCSP SSL support.
6881 *) New function OCSP_parse_url(). This splits up a URL into its host,
6882 port and path components: primarily to parse OCSP URLs. New -url
6883 option to ocsp utility.
6886 *) New nonce behavior. The return value of OCSP_check_nonce() now
6887 reflects the various checks performed. Applications can decide
6888 whether to tolerate certain situations such as an absent nonce
6889 in a response when one was present in a request: the ocsp application
6890 just prints out a warning. New function OCSP_add1_basic_nonce()
6891 this is to allow responders to include a nonce in a response even if
6892 the request is nonce-less.
6895 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6896 skipped when using openssl x509 multiple times on a single input file,
6897 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6900 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6901 set string type: to handle setting ASN1_TIME structures. Fix ca
6902 utility to correctly initialize revocation date of CRLs.
6905 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6906 the clients preferred ciphersuites and rather use its own preferences.
6907 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6908 Internet Explorer by ensuring unchanged hash method during stepup.
6909 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6912 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6913 to aes and add a new 'exist' option to print out symbols that don't
6917 *) Additional options to ocsp utility to allow flags to be set and
6918 additional certificates supplied.
6921 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6922 OCSP client a number of certificate to only verify the response
6926 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6927 handle the new API. Currently only ECB, CBC modes supported. Add new
6930 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6931 Encryption Standard (AES) Ciphersuites for Transport Layer
6932 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6933 not enabled by default and were not part of the "ALL" ciphersuite
6934 alias because they were not yet official; they could be
6935 explicitly requested by specifying the "AESdraft" ciphersuite
6936 group alias. In the final release of OpenSSL 0.9.7, the group
6937 alias is called "AES" and is part of "ALL".)
6938 [Ben Laurie, Steve Henson, Bodo Moeller]
6940 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6941 request to response.
6944 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6945 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6946 extract information from a certificate request. OCSP_response_create()
6947 creates a response and optionally adds a basic response structure.
6948 OCSP_basic_add1_status() adds a complete single response to a basic
6949 response and returns the OCSP_SINGLERESP structure just added (to allow
6950 extensions to be included for example). OCSP_basic_add1_cert() adds a
6951 certificate to a basic response and OCSP_basic_sign() signs a basic
6952 response with various flags. New helper functions ASN1_TIME_check()
6953 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6954 (converts ASN1_TIME to GeneralizedTime).
6957 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6958 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6959 structure from a certificate. X509_pubkey_digest() digests the public_key
6960 contents: this is used in various key identifiers.
6963 *) Make sk_sort() tolerate a NULL argument.
6964 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6966 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6967 passed by the function are trusted implicitly. If any of them signed the
6968 response then it is assumed to be valid and is not verified.
6971 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6972 to data. This was previously part of the PKCS7 ASN1 code. This
6973 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6974 [Steve Henson, reported by Kenneth R. Robinette
6975 <support@securenetterm.com>]
6977 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6978 routines: without these tracing memory leaks is very painful.
6979 Fix leaks in PKCS12 and PKCS7 routines.
6982 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6983 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6984 effectively meant GeneralizedTime would never be used. Now it
6985 is initialised to -1 but X509_time_adj() now has to check the value
6986 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6987 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6988 [Steve Henson, reported by Kenneth R. Robinette
6989 <support@securenetterm.com>]
6991 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6992 result in a zero length in the ASN1_INTEGER structure which was
6993 not consistent with the structure when d2i_ASN1_INTEGER() was used
6994 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6995 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6996 where it did not print out a minus for negative ASN1_INTEGER.
6999 *) Add summary printout to ocsp utility. The various functions which
7000 convert status values to strings have been renamed to:
7001 OCSP_response_status_str(), OCSP_cert_status_str() and
7002 OCSP_crl_reason_str() and are no longer static. New options
7003 to verify nonce values and to disable verification. OCSP response
7004 printout format cleaned up.
7007 *) Add additional OCSP certificate checks. These are those specified
7008 in RFC2560. This consists of two separate checks: the CA of the
7009 certificate being checked must either be the OCSP signer certificate
7010 or the issuer of the OCSP signer certificate. In the latter case the
7011 OCSP signer certificate must contain the OCSP signing extended key
7012 usage. This check is performed by attempting to match the OCSP
7013 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7014 in the OCSP_CERTID structures of the response.
7017 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7018 and related routines. This uses the standard OpenSSL certificate
7019 verify routines to perform initial checks (just CA validity) and
7020 to obtain the certificate chain. Then additional checks will be
7021 performed on the chain. Currently the root CA is checked to see
7022 if it is explicitly trusted for OCSP signing. This is used to set
7023 a root CA as a global signing root: that is any certificate that
7024 chains to that CA is an acceptable OCSP signing certificate.
7027 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7028 extensions from a separate configuration file.
7029 As when reading extensions from the main configuration file,
7030 the '-extensions ...' option may be used for specifying the
7032 [Massimiliano Pala <madwolf@comune.modena.it>]
7034 *) New OCSP utility. Allows OCSP requests to be generated or
7035 read. The request can be sent to a responder and the output
7036 parsed, outputed or printed in text form. Not complete yet:
7037 still needs to check the OCSP response validity.
7040 *) New subcommands for 'openssl ca':
7041 'openssl ca -status <serial>' prints the status of the cert with
7042 the given serial number (according to the index file).
7043 'openssl ca -updatedb' updates the expiry status of certificates
7045 [Massimiliano Pala <madwolf@comune.modena.it>]
7047 *) New '-newreq-nodes' command option to CA.pl. This is like
7048 '-newreq', but calls 'openssl req' with the '-nodes' option
7049 so that the resulting key is not encrypted.
7050 [Damien Miller <djm@mindrot.org>]
7052 *) New configuration for the GNU Hurd.
7053 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7055 *) Initial code to implement OCSP basic response verify. This
7056 is currently incomplete. Currently just finds the signer's
7057 certificate and verifies the signature on the response.
7060 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7061 value of OPENSSLDIR. This is available via the new '-d' option
7062 to 'openssl version', and is also included in 'openssl version -a'.
7065 *) Allowing defining memory allocation callbacks that will be given
7066 file name and line number information in additional arguments
7067 (a const char* and an int). The basic functionality remains, as
7068 well as the original possibility to just replace malloc(),
7069 realloc() and free() by functions that do not know about these
7070 additional arguments. To register and find out the current
7071 settings for extended allocation functions, the following
7072 functions are provided:
7074 CRYPTO_set_mem_ex_functions
7075 CRYPTO_set_locked_mem_ex_functions
7076 CRYPTO_get_mem_ex_functions
7077 CRYPTO_get_locked_mem_ex_functions
7079 These work the same way as CRYPTO_set_mem_functions and friends.
7080 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7081 extended allocation function is enabled.
7082 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7083 a conventional allocation function is enabled.
7084 [Richard Levitte, Bodo Moeller]
7086 *) Finish off removing the remaining LHASH function pointer casts.
7087 There should no longer be any prototype-casting required when using
7088 the LHASH abstraction, and any casts that remain are "bugs". See
7089 the callback types and macros at the head of lhash.h for details
7090 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7093 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7094 If /dev/[u]random devices are not available or do not return enough
7095 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7097 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7098 /etc/entropy will be queried once each in this sequence, quering stops
7099 when enough entropy was collected without querying more sockets.
7102 *) Change the Unix RAND_poll() variant to be able to poll several
7103 random devices, as specified by DEVRANDOM, until a sufficient amount
7104 of data has been collected. We spend at most 10 ms on each file
7105 (select timeout) and read in non-blocking mode. DEVRANDOM now
7106 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7107 (previously it was just the string "/dev/urandom"), so on typical
7108 platforms the 10 ms delay will never occur.
7109 Also separate out the Unix variant to its own file, rand_unix.c.
7110 For VMS, there's a currently-empty rand_vms.c.
7113 *) Move OCSP client related routines to ocsp_cl.c. These
7114 provide utility functions which an application needing
7115 to issue a request to an OCSP responder and analyse the
7116 response will typically need: as opposed to those which an
7117 OCSP responder itself would need which will be added later.
7119 OCSP_request_sign() signs an OCSP request with an API similar
7120 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7121 response. OCSP_response_get1_basic() extracts basic response
7122 from response. OCSP_resp_find_status(): finds and extracts status
7123 information from an OCSP_CERTID structure (which will be created
7124 when the request structure is built). These are built from lower
7125 level functions which work on OCSP_SINGLERESP structures but
7126 wont normally be used unless the application wishes to examine
7127 extensions in the OCSP response for example.
7129 Replace nonce routines with a pair of functions.
7130 OCSP_request_add1_nonce() adds a nonce value and optionally
7131 generates a random value. OCSP_check_nonce() checks the
7132 validity of the nonce in an OCSP response.
7135 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7136 This doesn't copy the supplied OCSP_CERTID and avoids the
7137 need to free up the newly created id. Change return type
7138 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7139 This can then be used to add extensions to the request.
7140 Deleted OCSP_request_new(), since most of its functionality
7141 is now in OCSP_REQUEST_new() (and the case insensitive name
7142 clash) apart from the ability to set the request name which
7143 will be added elsewhere.
7146 *) Update OCSP API. Remove obsolete extensions argument from
7147 various functions. Extensions are now handled using the new
7148 OCSP extension code. New simple OCSP HTTP function which
7149 can be used to send requests and parse the response.
7152 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7153 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7154 uses the special reorder version of SET OF to sort the attributes
7155 and reorder them to match the encoded order. This resolves a long
7156 standing problem: a verify on a PKCS7 structure just after signing
7157 it used to fail because the attribute order did not match the
7158 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7159 it uses the received order. This is necessary to tolerate some broken
7160 software that does not order SET OF. This is handled by encoding
7161 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7162 to produce the required SET OF.
7165 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7166 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7167 files to get correct declarations of the ASN.1 item variables.
7170 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7171 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7172 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7173 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7174 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7175 ASN1_ITEM and no wrapper functions.
7178 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7179 replace the old function pointer based I/O routines. Change most of
7180 the *_d2i_bio() and *_d2i_fp() functions to use these.
7183 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7184 lines, recognice more "algorithms" that can be deselected, and make
7185 it complain about algorithm deselection that isn't recognised.
7188 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7189 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7190 to use new functions. Add NO_ASN1_OLD which can be set to remove
7191 some old style ASN1 functions: this can be used to determine if old
7192 code will still work when these eventually go away.
7195 *) New extension functions for OCSP structures, these follow the
7196 same conventions as certificates and CRLs.
7199 *) New function X509V3_add1_i2d(). This automatically encodes and
7200 adds an extension. Its behaviour can be customised with various
7201 flags to append, replace or delete. Various wrappers added for
7202 certifcates and CRLs.
7205 *) Fix to avoid calling the underlying ASN1 print routine when
7206 an extension cannot be parsed. Correct a typo in the
7207 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7210 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7211 entries for variables.
7214 *) Add functionality to apps/openssl.c for detecting locking
7215 problems: As the program is single-threaded, all we have
7216 to do is register a locking callback using an array for
7217 storing which locks are currently held by the program.
7220 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7221 SSL_get_ex_data_X509_STORE_idx(), which is used in
7222 ssl_verify_cert_chain() and thus can be called at any time
7223 during TLS/SSL handshakes so that thread-safety is essential.
7224 Unfortunately, the ex_data design is not at all suited
7225 for multi-threaded use, so it probably should be abolished.
7228 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7229 [Broadcom, tweaked and integrated by Geoff Thorpe]
7231 *) Move common extension printing code to new function
7232 X509V3_print_extensions(). Reorganise OCSP print routines and
7233 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7236 *) New function X509_signature_print() to remove duplication in some
7240 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7241 set (this was treated exactly the same as SET OF previously). This
7242 is used to reorder the STACK representing the structure to match the
7243 encoding. This will be used to get round a problem where a PKCS7
7244 structure which was signed could not be verified because the STACK
7245 order did not reflect the encoded order.
7248 *) Reimplement the OCSP ASN1 module using the new code.
7251 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7252 for its ASN1 operations. The old style function pointers still exist
7253 for now but they will eventually go away.
7256 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7257 completely replaces the old ASN1 functionality with a table driven
7258 encoder and decoder which interprets an ASN1_ITEM structure describing
7259 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7260 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7261 has also been converted to the new form.
7264 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7265 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7266 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7267 for negative moduli.
7270 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7271 of not touching the result's sign bit.
7274 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7278 *) Changed the LHASH code to use prototypes for callbacks, and created
7279 macros to declare and implement thin (optionally static) functions
7280 that provide type-safety and avoid function pointer casting for the
7281 type-specific callbacks.
7284 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7286 [Veers Staats <staatsvr@asc.hpc.mil>,
7287 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7289 *) Reformat the FAQ so the different questions and answers can be divided
7290 in sections depending on the subject.
7293 *) Have the zlib compression code load ZLIB.DLL dynamically under
7297 *) New function BN_mod_sqrt for computing square roots modulo a prime
7298 (using the probabilistic Tonelli-Shanks algorithm unless
7299 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7300 be handled deterministically).
7301 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7303 *) Make BN_mod_inverse faster by explicitly handling small quotients
7304 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7305 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7308 *) New function BN_kronecker.
7311 *) Fix BN_gcd so that it works on negative inputs; the result is
7312 positive unless both parameters are zero.
7313 Previously something reasonably close to an infinite loop was
7314 possible because numbers could be growing instead of shrinking
7315 in the implementation of Euclid's algorithm.
7318 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7319 sign of the number in question.
7321 Fix BN_is_word(a,w) to work correctly for w == 0.
7323 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7324 because its test if the absolute value of 'a' equals 'w'.
7325 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7326 it exists mostly for use in the implementations of BN_is_zero(),
7327 BN_is_one(), and BN_is_word().
7330 *) New function BN_swap.
7333 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7334 the exponentiation functions are more likely to produce reasonable
7335 results on negative inputs.
7338 *) Change BN_mod_mul so that the result is always non-negative.
7339 Previously, it could be negative if one of the factors was negative;
7340 I don't think anyone really wanted that behaviour.
7343 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7344 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7345 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7346 and add new functions:
7355 BN_mod_lshift1_quick
7359 These functions always generate non-negative results.
7361 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7362 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7364 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7365 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7366 be reduced modulo m.
7367 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7370 The following entry accidentally appeared in the CHANGES file
7371 distributed with OpenSSL 0.9.7. The modifications described in
7372 it do *not* apply to OpenSSL 0.9.7.
7374 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7375 was actually never needed) and in BN_mul(). The removal in BN_mul()
7376 required a small change in bn_mul_part_recursive() and the addition
7377 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7378 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7379 bn_sub_words() and bn_add_words() except they take arrays with
7384 *) In 'openssl passwd', verify passwords read from the terminal
7385 unless the '-salt' option is used (which usually means that
7386 verification would just waste user's time since the resulting
7387 hash is going to be compared with some given password hash)
7388 or the new '-noverify' option is used.
7390 This is an incompatible change, but it does not affect
7391 non-interactive use of 'openssl passwd' (passwords on the command
7392 line, '-stdin' option, '-in ...' option) and thus should not
7396 *) Remove all references to RSAref, since there's no more need for it.
7399 *) Make DSO load along a path given through an environment variable
7400 (SHLIB_PATH) with shl_load().
7403 *) Constify the ENGINE code as a result of BIGNUM constification.
7404 Also constify the RSA code and most things related to it. In a
7405 few places, most notable in the depth of the ASN.1 code, ugly
7406 casts back to non-const were required (to be solved at a later
7410 *) Make it so the openssl application has all engines loaded by default.
7413 *) Constify the BIGNUM routines a little more.
7416 *) Add the following functions:
7418 ENGINE_load_cswift()
7420 ENGINE_load_atalla()
7422 ENGINE_load_builtin_engines()
7424 That way, an application can itself choose if external engines that
7425 are built-in in OpenSSL shall ever be used or not. The benefit is
7426 that applications won't have to be linked with libdl or other dso
7427 libraries unless it's really needed.
7429 Changed 'openssl engine' to load all engines on demand.
7430 Changed the engine header files to avoid the duplication of some
7431 declarations (they differed!).
7434 *) 'openssl engine' can now list capabilities.
7437 *) Better error reporting in 'openssl engine'.
7440 *) Never call load_dh_param(NULL) in s_server.
7443 *) Add engine application. It can currently list engines by name and
7444 identity, and test if they are actually available.
7447 *) Improve RPM specification file by forcing symbolic linking and making
7448 sure the installed documentation is also owned by root.root.
7449 [Damien Miller <djm@mindrot.org>]
7451 *) Give the OpenSSL applications more possibilities to make use of
7452 keys (public as well as private) handled by engines.
7455 *) Add OCSP code that comes from CertCo.
7458 *) Add VMS support for the Rijndael code.
7461 *) Added untested support for Nuron crypto accelerator.
7464 *) Add support for external cryptographic devices. This code was
7465 previously distributed separately as the "engine" branch.
7466 [Geoff Thorpe, Richard Levitte]
7468 *) Rework the filename-translation in the DSO code. It is now possible to
7469 have far greater control over how a "name" is turned into a filename
7470 depending on the operating environment and any oddities about the
7471 different shared library filenames on each system.
7474 *) Support threads on FreeBSD-elf in Configure.
7477 *) Fix for SHA1 assembly problem with MASM: it produces
7478 warnings about corrupt line number information when assembling
7479 with debugging information. This is caused by the overlapping
7481 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7484 NCONF_get_number() has no error checking at all. As a replacement,
7485 NCONF_get_number_e() is defined (_e for "error checking") and is
7486 promoted strongly. The old NCONF_get_number is kept around for
7487 binary backward compatibility.
7488 Make it possible for methods to load from something other than a BIO,
7489 by providing a function pointer that is given a name instead of a BIO.
7490 For example, this could be used to load configuration data from an
7494 *) Fix for non blocking accept BIOs. Added new I/O special reason
7495 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7496 with non blocking I/O was not possible because no retry code was
7497 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7501 *) Added the beginnings of Rijndael support.
7504 *) Fix for bug in DirectoryString mask setting. Add support for
7505 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7506 to allow certificate printing to more controllable, additional
7507 'certopt' option to 'x509' to allow new printing options to be
7511 *) Clean old EAY MD5 hack from e_os.h.
7514 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7516 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7517 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7518 [Joe Orton, Steve Henson]
7520 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7522 *) Fix additional bug revealed by the NISCC test suite:
7524 Stop bug triggering large recursion when presented with
7525 certain ASN.1 tags (CVE-2003-0851)
7528 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7530 *) Fix various bugs revealed by running the NISCC test suite:
7532 Stop out of bounds reads in the ASN1 code when presented with
7533 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7535 If verify callback ignores invalid public key errors don't try to check
7536 certificate signature with the NULL public key.
7540 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7541 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7545 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7546 extra data after the compression methods not only for TLS 1.0
7547 but also for SSL 3.0 (as required by the specification).
7548 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7550 *) Change X509_certificate_type() to mark the key as exported/exportable
7551 when it's 512 *bits* long, not 512 bytes.
7554 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7556 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7557 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7558 a protocol version number mismatch like a decryption error
7559 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7562 *) Turn on RSA blinding by default in the default implementation
7563 to avoid a timing attack. Applications that don't want it can call
7564 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7565 They would be ill-advised to do so in most cases.
7566 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7568 *) Change RSA blinding code so that it works when the PRNG is not
7569 seeded (in this case, the secret RSA exponent is abused as
7570 an unpredictable seed -- if it is not unpredictable, there
7571 is no point in blinding anyway). Make RSA blinding thread-safe
7572 by remembering the creator's thread ID in rsa->blinding and
7573 having all other threads use local one-time blinding factors
7574 (this requires more computation than sharing rsa->blinding, but
7575 avoids excessive locking; and if an RSA object is not shared
7576 between threads, blinding will still be very fast).
7579 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7581 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7582 via timing by performing a MAC computation even if incorrrect
7583 block cipher padding has been found. This is a countermeasure
7584 against active attacks where the attacker has to distinguish
7585 between bad padding and a MAC verification error. (CVE-2003-0078)
7587 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7588 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7589 Martin Vuagnoux (EPFL, Ilion)]
7591 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7593 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7594 memory from it's contents. This is done with a counter that will
7595 place alternating values in each byte. This can be used to solve
7596 two issues: 1) the removal of calls to memset() by highly optimizing
7597 compilers, and 2) cleansing with other values than 0, since those can
7598 be read through on certain media, for example a swap space on disk.
7601 *) Bugfix: client side session caching did not work with external caching,
7602 because the session->cipher setting was not restored when reloading
7603 from the external cache. This problem was masked, when
7604 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7605 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7608 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7609 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7610 [Zeev Lieber <zeev-l@yahoo.com>]
7612 *) Undo an undocumented change introduced in 0.9.6e which caused
7613 repeated calls to OpenSSL_add_all_ciphers() and
7614 OpenSSL_add_all_digests() to be ignored, even after calling
7618 *) Change the default configuration reader to deal with last line not
7619 being properly terminated.
7622 *) Change X509_NAME_cmp() so it applies the special rules on handling
7623 DN values that are of type PrintableString, as well as RDNs of type
7624 emailAddress where the value has the type ia5String.
7625 [stefank@valicert.com via Richard Levitte]
7627 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7628 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7629 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7630 the bitwise-OR of the two for use by the majority of applications
7631 wanting this behaviour, and update the docs. The documented
7632 behaviour and actual behaviour were inconsistent and had been
7633 changing anyway, so this is more a bug-fix than a behavioural
7635 [Geoff Thorpe, diagnosed by Nadav Har'El]
7637 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7638 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7641 *) Fix initialization code race conditions in
7642 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7643 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7644 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7645 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7646 ssl2_get_cipher_by_char(),
7647 ssl3_get_cipher_by_char().
7648 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7650 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7651 the cached sessions are flushed, as the remove_cb() might use ex_data
7652 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7653 (see [openssl.org #212]).
7654 [Geoff Thorpe, Lutz Jaenicke]
7656 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7657 length, instead of the encoding length to d2i_ASN1_OBJECT.
7660 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7662 *) [In 0.9.6g-engine release:]
7663 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7664 [Lynn Gazis <lgazis@rainbow.com>]
7666 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7668 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7669 and get fix the header length calculation.
7670 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7671 Alon Kantor <alonk@checkpoint.com> (and others),
7674 *) Use proper error handling instead of 'assertions' in buffer
7675 overflow checks added in 0.9.6e. This prevents DoS (the
7676 assertions could call abort()).
7677 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7679 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7681 *) Add various sanity checks to asn1_get_length() to reject
7682 the ASN1 length bytes if they exceed sizeof(long), will appear
7683 negative or the content length exceeds the length of the
7685 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7687 *) Fix cipher selection routines: ciphers without encryption had no flags
7688 for the cipher strength set and where therefore not handled correctly
7689 by the selection routines (PR #130).
7692 *) Fix EVP_dsa_sha macro.
7696 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7697 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7698 that was added in OpenSSL 0.9.6d.
7700 As the countermeasure turned out to be incompatible with some
7701 broken SSL implementations, the new option is part of SSL_OP_ALL.
7702 SSL_OP_ALL is usually employed when compatibility with weird SSL
7703 implementations is desired (e.g. '-bugs' option to 's_client' and
7704 's_server'), so the new option is automatically set in many
7708 *) Changes in security patch:
7710 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7711 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7712 Air Force Materiel Command, USAF, under agreement number
7715 *) Add various sanity checks to asn1_get_length() to reject
7716 the ASN1 length bytes if they exceed sizeof(long), will appear
7717 negative or the content length exceeds the length of the
7718 supplied buffer. (CVE-2002-0659)
7719 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7721 *) Assertions for various potential buffer overflows, not known to
7723 [Ben Laurie (CHATS)]
7725 *) Various temporary buffers to hold ASCII versions of integers were
7726 too small for 64 bit platforms. (CVE-2002-0655)
7727 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7729 *) Remote buffer overflow in SSL3 protocol - an attacker could
7730 supply an oversized session ID to a client. (CVE-2002-0656)
7731 [Ben Laurie (CHATS)]
7733 *) Remote buffer overflow in SSL2 protocol - an attacker could
7734 supply an oversized client master key. (CVE-2002-0656)
7735 [Ben Laurie (CHATS)]
7737 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7739 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7740 encoded as NULL) with id-dsa-with-sha1.
7741 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7743 *) Check various X509_...() return values in apps/req.c.
7744 [Nils Larsch <nla@trustcenter.de>]
7746 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7747 an end-of-file condition would erronously be flagged, when the CRLF
7748 was just at the end of a processed block. The bug was discovered when
7749 processing data through a buffering memory BIO handing the data to a
7750 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7751 <ptsekov@syntrex.com> and Nedelcho Stanev.
7754 *) Implement a countermeasure against a vulnerability recently found
7755 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7756 before application data chunks to avoid the use of known IVs
7757 with data potentially chosen by the attacker.
7760 *) Fix length checks in ssl3_get_client_hello().
7763 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7764 to prevent ssl3_read_internal() from incorrectly assuming that
7765 ssl3_read_bytes() found application data while handshake
7766 processing was enabled when in fact s->s3->in_read_app_data was
7767 merely automatically cleared during the initial handshake.
7768 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7770 *) Fix object definitions for Private and Enterprise: they were not
7771 recognized in their shortname (=lowercase) representation. Extend
7772 obj_dat.pl to issue an error when using undefined keywords instead
7773 of silently ignoring the problem (Svenning Sorensen
7774 <sss@sss.dnsalias.net>).
7777 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7778 generators, i.e. generators other than 2 and 5. (Previously, the
7779 code did not properly initialise the 'add' and 'rem' values to
7780 BN_generate_prime().)
7782 In the new general case, we do not insist that 'generator' is
7783 actually a primitive root: This requirement is rather pointless;
7784 a generator of the order-q subgroup is just as good, if not
7788 *) Map new X509 verification errors to alerts. Discovered and submitted by
7789 Tom Wu <tom@arcot.com>.
7792 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7793 returning non-zero before the data has been completely received
7794 when using non-blocking I/O.
7795 [Bodo Moeller; problem pointed out by John Hughes]
7797 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7798 [Ben Laurie, Lutz Jaenicke]
7800 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7801 Yoram Zahavi <YoramZ@gilian.com>).
7804 *) Add information about CygWin 1.3 and on, and preserve proper
7805 configuration for the versions before that.
7806 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7808 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7809 check whether we deal with a copy of a session and do not delete from
7810 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7811 <izhar@checkpoint.com>.
7814 *) Do not store session data into the internal session cache, if it
7815 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7816 flag is set). Proposed by Aslam <aslam@funk.com>.
7819 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7823 *) [In 0.9.6d-engine release:]
7824 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7825 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7827 *) Add the configuration target linux-s390x.
7828 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7830 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7831 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7832 variable as an indication that a ClientHello message has been
7833 received. As the flag value will be lost between multiple
7834 invocations of ssl3_accept when using non-blocking I/O, the
7835 function may not be aware that a handshake has actually taken
7836 place, thus preventing a new session from being added to the
7839 To avoid this problem, we now set s->new_session to 2 instead of
7840 using a local variable.
7841 [Lutz Jaenicke, Bodo Moeller]
7843 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7844 if the SSL_R_LENGTH_MISMATCH error is detected.
7845 [Geoff Thorpe, Bodo Moeller]
7847 *) New 'shared_ldflag' column in Configure platform table.
7850 *) Fix EVP_CIPHER_mode macro.
7851 ["Dan S. Camper" <dan@bti.net>]
7853 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7854 type, we must throw them away by setting rr->length to 0.
7855 [D P Chang <dpc@qualys.com>]
7857 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7859 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7860 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7861 worked incorrectly for those cases where range = 10..._2 and
7862 3*range is two bits longer than range.)
7865 *) Only add signing time to PKCS7 structures if it is not already
7869 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7870 OBJ_ld_ce should be OBJ_id_ce.
7871 Also some ip-pda OIDs in crypto/objects/objects.txt were
7872 incorrect (cf. RFC 3039).
7873 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7875 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7876 returns early because it has nothing to do.
7877 [Andy Schneider <andy.schneider@bjss.co.uk>]
7879 *) [In 0.9.6c-engine release:]
7880 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7881 [Andy Schneider <andy.schneider@bjss.co.uk>]
7883 *) [In 0.9.6c-engine release:]
7884 Add support for Cryptographic Appliance's keyserver technology.
7885 (Use engine 'keyclient')
7886 [Cryptographic Appliances and Geoff Thorpe]
7888 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7889 is called via tools/c89.sh because arguments have to be
7890 rearranged (all '-L' options must appear before the first object
7892 [Richard Shapiro <rshapiro@abinitio.com>]
7894 *) [In 0.9.6c-engine release:]
7895 Add support for Broadcom crypto accelerator cards, backported
7897 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7899 *) [In 0.9.6c-engine release:]
7900 Add support for SureWare crypto accelerator cards from
7901 Baltimore Technologies. (Use engine 'sureware')
7902 [Baltimore Technologies and Mark Cox]
7904 *) [In 0.9.6c-engine release:]
7905 Add support for crypto accelerator cards from Accelerated
7906 Encryption Processing, www.aep.ie. (Use engine 'aep')
7907 [AEP Inc. and Mark Cox]
7909 *) Add a configuration entry for gcc on UnixWare.
7910 [Gary Benson <gbenson@redhat.com>]
7912 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7913 messages are stored in a single piece (fixed-length part and
7914 variable-length part combined) and fix various bugs found on the way.
7917 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7918 instead. BIO_gethostbyname() does not know what timeouts are
7919 appropriate, so entries would stay in cache even when they have
7921 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7923 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7924 faced with a pathologically small ClientHello fragment that does
7925 not contain client_version: Instead of aborting with an error,
7926 simply choose the highest available protocol version (i.e.,
7927 TLS 1.0 unless it is disabled). In practice, ClientHello
7928 messages are never sent like this, but this change gives us
7929 strictly correct behaviour at least for TLS.
7932 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7933 never resets s->method to s->ctx->method when called from within
7934 one of the SSL handshake functions.
7935 [Bodo Moeller; problem pointed out by Niko Baric]
7937 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7938 (sent using the client's version number) if client_version is
7939 smaller than the protocol version in use. Also change
7940 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7941 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7942 the client will at least see that alert.
7945 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7949 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7950 client receives HelloRequest while in a handshake.
7951 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7953 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7954 should end in 'break', not 'goto end' which circumvents various
7955 cleanups done in state SSL_ST_OK. But session related stuff
7956 must be disabled for SSL_ST_OK in the case that we just sent a
7959 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7960 before just sending a HelloRequest.
7961 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7963 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7964 reveal whether illegal block cipher padding was found or a MAC
7965 verification error occurred. (Neither SSLerr() codes nor alerts
7966 are directly visible to potential attackers, but the information
7967 may leak via logfiles.)
7969 Similar changes are not required for the SSL 2.0 implementation
7970 because the number of padding bytes is sent in clear for SSL 2.0,
7971 and the extra bytes are just ignored. However ssl/s2_pkt.c
7972 failed to verify that the purported number of padding bytes is in
7976 *) Add OpenUNIX-8 support including shared libraries
7977 (Boyd Lynn Gerber <gerberb@zenez.com>).
7980 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7981 'wristwatch attack' using huge encoding parameters (cf.
7982 James H. Manger's CRYPTO 2001 paper). Note that the
7983 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7984 encoding parameters and hence was not vulnerable.
7987 *) BN_sqr() bug fix.
7988 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7990 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7991 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7992 followed by modular reduction.
7993 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7995 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7996 equivalent based on BN_pseudo_rand() instead of BN_rand().
7999 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8000 This function was broken, as the check for a new client hello message
8001 to handle SGC did not allow these large messages.
8002 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8005 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8008 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8009 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8012 *) Rework the configuration and shared library support for Tru64 Unix.
8013 The configuration part makes use of modern compiler features and
8014 still retains old compiler behavior for those that run older versions
8015 of the OS. The shared library support part includes a variant that
8016 uses the RPATH feature, and is available through the special
8017 configuration target "alpha-cc-rpath", which will never be selected
8019 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8021 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8022 with the same message size as in ssl3_get_certificate_request().
8023 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8024 messages might inadvertently be reject as too long.
8025 [Petr Lampa <lampa@fee.vutbr.cz>]
8027 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8030 *) Modified SSL library such that the verify_callback that has been set
8031 specificly for an SSL object with SSL_set_verify() is actually being
8032 used. Before the change, a verify_callback set with this function was
8033 ignored and the verify_callback() set in the SSL_CTX at the time of
8034 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8035 to allow the necessary settings.
8038 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8039 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8040 done automatically (in contradiction to the requirements of the C
8041 standard). This made problems when used from OpenSSH.
8044 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8045 dh->length and always used
8047 BN_rand_range(priv_key, dh->p).
8049 BN_rand_range() is not necessary for Diffie-Hellman, and this
8050 specific range makes Diffie-Hellman unnecessarily inefficient if
8051 dh->length (recommended exponent length) is much smaller than the
8052 length of dh->p. We could use BN_rand_range() if the order of
8053 the subgroup was stored in the DH structure, but we only have
8058 BN_rand(priv_key, l, ...)
8060 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8066 RSA_eay_public_encrypt
8067 RSA_eay_private_decrypt
8068 RSA_eay_private_encrypt (signing)
8069 RSA_eay_public_decrypt (signature verification)
8071 (default implementations for RSA_public_encrypt,
8072 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8073 always reject numbers >= n.
8076 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8077 to synchronize access to 'locking_thread'. This is necessary on
8078 systems where access to 'locking_thread' (an 'unsigned long'
8079 variable) is not atomic.
8082 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8083 *before* setting the 'crypto_lock_rand' flag. The previous code had
8084 a race condition if 0 is a valid thread ID.
8085 [Travis Vitek <vitek@roguewave.com>]
8087 *) Add support for shared libraries under Irix.
8088 [Albert Chin-A-Young <china@thewrittenword.com>]
8090 *) Add configuration option to build on Linux on both big-endian and
8092 [Ralf Baechle <ralf@uni-koblenz.de>]
8094 *) Add the possibility to create shared libraries on HP-UX.
8097 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8099 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8100 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8101 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8102 PRNG state recovery was possible based on the output of
8103 one PRNG request appropriately sized to gain knowledge on
8104 'md' followed by enough consecutive 1-byte PRNG requests
8105 to traverse all of 'state'.
8107 1. When updating 'md_local' (the current thread's copy of 'md')
8108 during PRNG output generation, hash all of the previous
8109 'md_local' value, not just the half used for PRNG output.
8111 2. Make the number of bytes from 'state' included into the hash
8112 independent from the number of PRNG bytes requested.
8114 The first measure alone would be sufficient to avoid
8115 Markku-Juhani's attack. (Actually it had never occurred
8116 to me that the half of 'md_local' used for chaining was the
8117 half from which PRNG output bytes were taken -- I had always
8118 assumed that the secret half would be used.) The second
8119 measure makes sure that additional data from 'state' is never
8120 mixed into 'md_local' in small portions; this heuristically
8121 further strengthens the PRNG.
8124 *) Fix crypto/bn/asm/mips3.s.
8127 *) When only the key is given to "enc", the IV is undefined. Print out
8128 an error message in this case.
8131 *) Handle special case when X509_NAME is empty in X509 printing routines.
8134 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8135 positive and less than q.
8138 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8139 used: it isn't thread safe and the add_lock_callback should handle
8141 [Paul Rose <Paul.Rose@bridge.com>]
8143 *) Verify that incoming data obeys the block size in
8144 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8148 [Ulf Möller, Bodo Möller]
8150 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8151 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8152 when fixing the server behaviour for backwards-compatible 'client
8153 hello' messages. (Note that the attack is impractical against
8154 SSL 3.0 and TLS 1.0 anyway because length and version checking
8155 means that the probability of guessing a valid ciphertext is
8156 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8159 Before 0.9.5, the countermeasure (hide the error by generating a
8160 random 'decryption result') did not work properly because
8161 ERR_clear_error() was missing, meaning that SSL_get_error() would
8162 detect the supposedly ignored error.
8164 Both problems are now fixed.
8167 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8168 (previously it was 1024).
8171 *) Fix for compatibility mode trust settings: ignore trust settings
8172 unless some valid trust or reject settings are present.
8175 *) Fix for blowfish EVP: its a variable length cipher.
8178 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8179 parameters in DSA public key structures and return an error in the
8180 DSA routines if parameters are absent.
8183 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8184 in the current directory if neither $RANDFILE nor $HOME was set.
8185 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8186 caused some confusion to Windows users who haven't defined $HOME.
8187 Thus RAND_file_name() is changed again: e_os.h can define a
8188 DEFAULT_HOME, which will be used if $HOME is not set.
8189 For Windows, we use "C:"; on other platforms, we still require
8190 environment variables.
8192 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8193 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8194 having multiple threads call RAND_poll() concurrently.
8197 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8198 combination of a flag and a thread ID variable.
8199 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8200 flag), *other* threads can enter ssleay_add_bytes without obeying
8201 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8202 that they do not hold after the first thread unsets add_do_not_lock).
8205 *) Change bctest again: '-x' expressions are not available in all
8209 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8211 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8212 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8214 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8215 the default extension for executables, if any. Also, make the perl
8216 scripts that use symlink() to test if it really exists and use "cp"
8217 if it doesn't. All this made OpenSSL compilable and installable in
8221 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8222 If SEQUENCE is length is indefinite just set c->slen to the total
8223 amount of data available.
8224 [Steve Henson, reported by shige@FreeBSD.org]
8225 [This change does not apply to 0.9.7.]
8227 *) Change bctest to avoid here-documents inside command substitution
8228 (workaround for FreeBSD /bin/sh bug).
8229 For compatibility with Ultrix, avoid shell functions (introduced
8230 in the bctest version that searches along $PATH).
8233 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8234 with des_encrypt() defined on some operating systems, like Solaris
8238 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8239 On the Importance of Eliminating Errors in Cryptographic
8240 Computations, J. Cryptology 14 (2001) 2, 101-119,
8241 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8244 *) MIPS assembler BIGNUM division bug fix.
8247 *) Disabled incorrect Alpha assembler code.
8250 *) Fix PKCS#7 decode routines so they correctly update the length
8251 after reading an EOC for the EXPLICIT tag.
8253 [This change does not apply to 0.9.7.]
8255 *) Fix bug in PKCS#12 key generation routines. This was triggered
8256 if a 3DES key was generated with a 0 initial byte. Include
8257 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8258 (but broken) behaviour.
8261 *) Enhance bctest to search for a working bc along $PATH and print
8263 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8265 *) Fix memory leaks in err.c: free err_data string if necessary;
8266 don't write to the wrong index in ERR_set_error_data.
8269 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8273 *) Replace rdtsc with _emit statements for VC++ version 5.
8274 [Jeremy Cooper <jeremy@baymoo.org>]
8276 *) Make it possible to reuse SSLv2 sessions.
8279 *) In copy_email() check for >= 0 as a return value for
8280 X509_NAME_get_index_by_NID() since 0 is a valid index.
8281 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8283 *) Avoid coredump with unsupported or invalid public keys by checking if
8284 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8285 PKCS7_verify() fails with non detached data.
8288 *) Don't use getenv in library functions when run as setuid/setgid.
8289 New function OPENSSL_issetugid().
8292 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8293 due to incorrect handling of multi-threading:
8295 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8297 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8299 3. Count how many times MemCheck_off() has been called so that
8300 nested use can be treated correctly. This also avoids
8301 inband-signalling in the previous code (which relied on the
8302 assumption that thread ID 0 is impossible).
8305 *) Add "-rand" option also to s_client and s_server.
8308 *) Fix CPU detection on Irix 6.x.
8309 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8310 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8312 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8315 [This change does not apply to 0.9.7.]
8317 *) Use the cached encoding of an X509_NAME structure rather than
8318 copying it. This is apparently the reason for the libsafe "errors"
8319 but the code is actually correct.
8322 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8323 Bleichenbacher's DSA attack.
8324 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8325 to be set and top=0 forces the highest bit to be set; top=-1 is new
8326 and leaves the highest bit random.
8327 [Ulf Moeller, Bodo Moeller]
8329 *) In the NCONF_...-based implementations for CONF_... queries
8330 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8331 a temporary CONF structure with the data component set to NULL
8332 (which gives segmentation faults in lh_retrieve).
8333 Instead, use NULL for the CONF pointer in CONF_get_string and
8334 CONF_get_number (which may use environment variables) and directly
8335 return NULL from CONF_get_section.
8338 *) Fix potential buffer overrun for EBCDIC.
8341 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8342 keyUsage if basicConstraints absent for a CA.
8345 *) Make SMIME_write_PKCS7() write mail header values with a format that
8346 is more generally accepted (no spaces before the semicolon), since
8347 some programs can't parse those values properly otherwise. Also make
8348 sure BIO's that break lines after each write do not create invalid
8352 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8353 macros previously used would not encode an empty SEQUENCE OF
8354 and break the signature.
8356 [This change does not apply to 0.9.7.]
8358 *) Zero the premaster secret after deriving the master secret in
8362 *) Add some EVP_add_digest_alias registrations (as found in
8363 OpenSSL_add_all_digests()) to SSL_library_init()
8364 aka OpenSSL_add_ssl_algorithms(). This provides improved
8365 compatibility with peers using X.509 certificates
8366 with unconventional AlgorithmIdentifier OIDs.
8369 *) Fix for Irix with NO_ASM.
8370 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8372 *) ./config script fixes.
8373 [Ulf Moeller, Richard Levitte]
8375 *) Fix 'openssl passwd -1'.
8378 *) Change PKCS12_key_gen_asc() so it can cope with non null
8379 terminated strings whose length is passed in the passlen
8380 parameter, for example from PEM callbacks. This was done
8381 by adding an extra length parameter to asc2uni().
8382 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8384 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8385 call failed, free the DSA structure.
8388 *) Fix to uni2asc() to cope with zero length Unicode strings.
8389 These are present in some PKCS#12 files.
8392 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8393 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8394 when writing a 32767 byte record.
8395 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8397 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8398 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8400 (RSA objects have a reference count access to which is protected
8401 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8402 so they are meant to be shared between threads.)
8403 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8404 "Reddie, Steven" <Steven.Reddie@ca.com>]
8406 *) Fix a deadlock in CRYPTO_mem_leaks().
8409 *) Use better test patterns in bntest.
8412 *) rand_win.c fix for Borland C.
8415 *) BN_rshift bugfix for n == 0.
8418 *) Add a 'bctest' script that checks for some known 'bc' bugs
8419 so that 'make test' does not abort just because 'bc' is broken.
8422 *) Store verify_result within SSL_SESSION also for client side to
8423 avoid potential security hole. (Re-used sessions on the client side
8424 always resulted in verify_result==X509_V_OK, not using the original
8425 result of the server certificate verification.)
8428 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8429 SSL3_RT_APPLICATION_DATA, return 0.
8430 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8434 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8435 releases, have been re-implemented by renaming the previous
8436 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8437 and ssl3_read_internal, respectively, and adding 'peek' parameters
8438 to them. The new ssl[23]_{read,peek} functions are calls to
8439 ssl[23]_read_internal with the 'peek' flag set appropriately.
8440 A 'peek' parameter has also been added to ssl3_read_bytes, which
8441 does the actual work for ssl3_read_internal.
8444 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8445 the method-specific "init()" handler. Also clean up ex_data after
8446 calling the method-specific "finish()" handler. Previously, this was
8447 happening the other way round.
8450 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8451 The previous value, 12, was not always sufficient for BN_mod_exp().
8454 *) Make sure that shared libraries get the internal name engine with
8455 the full version number and not just 0. This should mark the
8456 shared libraries as not backward compatible. Of course, this should
8457 be changed again when we can guarantee backward binary compatibility.
8460 *) Fix typo in get_cert_by_subject() in by_dir.c
8461 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8463 *) Rework the system to generate shared libraries:
8465 - Make note of the expected extension for the shared libraries and
8466 if there is a need for symbolic links from for example libcrypto.so.0
8467 to libcrypto.so.0.9.7. There is extended info in Configure for
8470 - Make as few rebuilds of the shared libraries as possible.
8472 - Still avoid linking the OpenSSL programs with the shared libraries.
8474 - When installing, install the shared libraries separately from the
8478 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8480 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8481 and not in SSL_clear because the latter is also used by the
8482 accept/connect functions; previously, the settings made by
8483 SSL_set_read_ahead would be lost during the handshake.
8484 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8486 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8487 Previously, it would create entries for disabled algorithms no
8491 *) Added several new manual pages for SSL_* function.
8494 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8496 *) In ssl23_get_client_hello, generate an error message when faced
8497 with an initial SSL 3.0/TLS record that is too small to contain the
8498 first two bytes of the ClientHello message, i.e. client_version.
8499 (Note that this is a pathologic case that probably has never happened
8500 in real life.) The previous approach was to use the version number
8501 from the record header as a substitute; but our protocol choice
8502 should not depend on that one because it is not authenticated
8503 by the Finished messages.
8506 *) More robust randomness gathering functions for Windows.
8507 [Jeffrey Altman <jaltman@columbia.edu>]
8509 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8510 not set then we don't setup the error code for issuer check errors
8511 to avoid possibly overwriting other errors which the callback does
8512 handle. If an application does set the flag then we assume it knows
8513 what it is doing and can handle the new informational codes
8517 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8518 a general "ANY" type, as such it should be able to decode anything
8519 including tagged types. However it didn't check the class so it would
8520 wrongly interpret tagged types in the same way as their universal
8521 counterpart and unknown types were just rejected. Changed so that the
8522 tagged and unknown types are handled in the same way as a SEQUENCE:
8523 that is the encoding is stored intact. There is also a new type
8524 "V_ASN1_OTHER" which is used when the class is not universal, in this
8525 case we have no idea what the actual type is so we just lump them all
8529 *) On VMS, stdout may very well lead to a file that is written to
8530 in a record-oriented fashion. That means that every write() will
8531 write a separate record, which will be read separately by the
8532 programs trying to read from it. This can be very confusing.
8534 The solution is to put a BIO filter in the way that will buffer
8535 text until a linefeed is reached, and then write everything a
8536 line at a time, so every record written will be an actual line,
8537 not chunks of lines and not (usually doesn't happen, but I've
8538 seen it once) several lines in one record. BIO_f_linebuffer() is
8541 Currently, it's a VMS-only method, because that's where it has
8542 been tested well enough.
8545 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8546 it can return incorrect results.
8547 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8548 but it was in 0.9.6-beta[12].)
8551 *) Disable the check for content being present when verifying detached
8552 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8553 include zero length content when signing messages.
8556 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8557 BIO_ctrl (for BIO pairs).
8560 *) Add DSO method for VMS.
8563 *) Bug fix: Montgomery multiplication could produce results with the
8567 *) Add RPM specification openssl.spec and modify it to build three
8568 packages. The default package contains applications, application
8569 documentation and run-time libraries. The devel package contains
8570 include files, static libraries and function documentation. The
8571 doc package contains the contents of the doc directory. The original
8572 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8575 *) Add a large number of documentation files for many SSL routines.
8576 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8578 *) Add a configuration entry for Sony News 4.
8579 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8581 *) Don't set the two most significant bits to one when generating a
8582 random number < q in the DSA library.
8585 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8586 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8587 the underlying transport is blocking) if a handshake took place.
8588 (The default behaviour is needed by applications such as s_client
8589 and s_server that use select() to determine when to use SSL_read;
8590 but for applications that know in advance when to expect data, it
8591 just makes things more complicated.)
8594 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8598 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8599 work better on such systems.
8600 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8602 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8603 Update PKCS12_parse() so it copies the friendlyName and the
8604 keyid to the certificates aux info.
8607 *) Fix bug in PKCS7_verify() which caused an infinite loop
8608 if there was more than one signature.
8609 [Sven Uszpelkat <su@celocom.de>]
8611 *) Major change in util/mkdef.pl to include extra information
8612 about each symbol, as well as presenting variables as well
8613 as functions. This change means that there's n more need
8614 to rebuild the .num files when some algorithms are excluded.
8617 *) Allow the verify time to be set by an application,
8618 rather than always using the current time.
8621 *) Phase 2 verify code reorganisation. The certificate
8622 verify code now looks up an issuer certificate by a
8623 number of criteria: subject name, authority key id
8624 and key usage. It also verifies self signed certificates
8625 by the same criteria. The main comparison function is
8626 X509_check_issued() which performs these checks.
8628 Lot of changes were necessary in order to support this
8629 without completely rewriting the lookup code.
8631 Authority and subject key identifier are now cached.
8633 The LHASH 'certs' is X509_STORE has now been replaced
8634 by a STACK_OF(X509_OBJECT). This is mainly because an
8635 LHASH can't store or retrieve multiple objects with
8636 the same hash value.
8638 As a result various functions (which were all internal
8639 use only) have changed to handle the new X509_STORE
8640 structure. This will break anything that messed round
8641 with X509_STORE internally.
8643 The functions X509_STORE_add_cert() now checks for an
8644 exact match, rather than just subject name.
8646 The X509_STORE API doesn't directly support the retrieval
8647 of multiple certificates matching a given criteria, however
8648 this can be worked round by performing a lookup first
8649 (which will fill the cache with candidate certificates)
8650 and then examining the cache for matches. This is probably
8651 the best we can do without throwing out X509_LOOKUP
8652 entirely (maybe later...).
8654 The X509_VERIFY_CTX structure has been enhanced considerably.
8656 All certificate lookup operations now go via a get_issuer()
8657 callback. Although this currently uses an X509_STORE it
8658 can be replaced by custom lookups. This is a simple way
8659 to bypass the X509_STORE hackery necessary to make this
8660 work and makes it possible to use more efficient techniques
8661 in future. A very simple version which uses a simple
8662 STACK for its trusted certificate store is also provided
8663 using X509_STORE_CTX_trusted_stack().
8665 The verify_cb() and verify() callbacks now have equivalents
8666 in the X509_STORE_CTX structure.
8668 X509_STORE_CTX also has a 'flags' field which can be used
8669 to customise the verify behaviour.
8672 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8673 excludes S/MIME capabilities.
8676 *) When a certificate request is read in keep a copy of the
8677 original encoding of the signed data and use it when outputing
8678 again. Signatures then use the original encoding rather than
8679 a decoded, encoded version which may cause problems if the
8680 request is improperly encoded.
8683 *) For consistency with other BIO_puts implementations, call
8684 buffer_write(b, ...) directly in buffer_puts instead of calling
8687 In BIO_puts, increment b->num_write as in BIO_write.
8688 [Peter.Sylvester@EdelWeb.fr]
8690 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8691 BN_zero, we may not return a BIGNUM with an array consisting of
8695 *) Avoid calling abort() from within the library when problems are
8696 detected, except if preprocessor symbols have been defined
8697 (such as REF_CHECK, BN_DEBUG etc.).
8700 *) New openssl application 'rsautl'. This utility can be
8701 used for low level RSA operations. DER public key
8702 BIO/fp routines also added.
8705 *) New Configure entry and patches for compiling on QNX 4.
8706 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8708 *) A demo state-machine implementation was sponsored by
8709 Nuron (http://www.nuron.com/) and is now available in
8710 demos/state_machine.
8713 *) New options added to the 'dgst' utility for signature
8714 generation and verification.
8717 *) Unrecognized PKCS#7 content types are now handled via a
8718 catch all ASN1_TYPE structure. This allows unsupported
8719 types to be stored as a "blob" and an application can
8720 encode and decode it manually.
8723 *) Fix various signed/unsigned issues to make a_strex.c
8725 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8727 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8728 length if passed a buffer. ASN1_INTEGER_to_BN failed
8729 if passed a NULL BN and its argument was negative.
8730 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8732 *) Modification to PKCS#7 encoding routines to output definite
8733 length encoding. Since currently the whole structures are in
8734 memory there's not real point in using indefinite length
8735 constructed encoding. However if OpenSSL is compiled with
8736 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8739 *) Added BIO_vprintf() and BIO_vsnprintf().
8742 *) Added more prefixes to parse for in the the strings written
8743 through a logging bio, to cover all the levels that are available
8744 through syslog. The prefixes are now:
8746 PANIC, EMERG, EMR => LOG_EMERG
8747 ALERT, ALR => LOG_ALERT
8748 CRIT, CRI => LOG_CRIT
8749 ERROR, ERR => LOG_ERR
8750 WARNING, WARN, WAR => LOG_WARNING
8751 NOTICE, NOTE, NOT => LOG_NOTICE
8752 INFO, INF => LOG_INFO
8753 DEBUG, DBG => LOG_DEBUG
8755 and as before, if none of those prefixes are present at the
8756 beginning of the string, LOG_ERR is chosen.
8758 On Win32, the LOG_* levels are mapped according to this:
8760 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8761 LOG_WARNING => EVENTLOG_WARNING_TYPE
8762 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8766 *) Made it possible to reconfigure with just the configuration
8767 argument "reconf" or "reconfigure". The command line arguments
8768 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8769 and are retrieved from there when reconfiguring.
8773 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8775 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8778 *) The obj_dat.pl script was messing up the sorting of object
8779 names. The reason was that it compared the quoted version
8780 of strings as a result "OCSP" > "OCSP Signing" because
8781 " > SPACE. Changed script to store unquoted versions of
8782 names and add quotes on output. It was also omitting some
8783 names from the lookup table if they were given a default
8784 value (that is if SN is missing it is given the same
8785 value as LN and vice versa), these are now added on the
8786 grounds that if an object has a name we should be able to
8787 look it up. Finally added warning output when duplicate
8788 short or long names are found.
8791 *) Changes needed for Tandem NSK.
8792 [Scott Uroff <scott@xypro.com>]
8794 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8795 RSA_padding_check_SSLv23(), special padding was never detected
8796 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8797 version rollback attacks was not effective.
8799 In s23_clnt.c, don't use special rollback-attack detection padding
8800 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8801 client; similarly, in s23_srvr.c, don't do the rollback check if
8802 SSL 2.0 is the only protocol enabled in the server.
8805 *) Make it possible to get hexdumps of unprintable data with 'openssl
8806 asn1parse'. By implication, the functions ASN1_parse_dump() and
8807 BIO_dump_indent() are added.
8810 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8811 these print out strings and name structures based on various
8812 flags including RFC2253 support and proper handling of
8813 multibyte characters. Added options to the 'x509' utility
8814 to allow the various flags to be set.
8817 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8818 Also change the functions X509_cmp_current_time() and
8819 X509_gmtime_adj() work with an ASN1_TIME structure,
8820 this will enable certificates using GeneralizedTime in validity
8821 dates to be checked.
8824 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8825 negative public key encodings) on by default,
8826 NO_NEG_PUBKEY_BUG can be set to disable it.
8829 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8830 content octets. An i2c_ASN1_OBJECT is unnecessary because
8831 the encoding can be trivially obtained from the structure.
8834 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8835 not read locks (CRYPTO_r_[un]lock).
8838 *) A first attempt at creating official support for shared
8839 libraries through configuration. I've kept it so the
8840 default is static libraries only, and the OpenSSL programs
8841 are always statically linked for now, but there are
8842 preparations for dynamic linking in place.
8843 This has been tested on Linux and Tru64.
8846 *) Randomness polling function for Win9x, as described in:
8847 Peter Gutmann, Software Generation of Practically Strong
8851 *) Fix so PRNG is seeded in req if using an already existing
8855 *) New options to smime application. -inform and -outform
8856 allow alternative formats for the S/MIME message including
8857 PEM and DER. The -content option allows the content to be
8858 specified separately. This should allow things like Netscape
8859 form signing output easier to verify.
8862 *) Fix the ASN1 encoding of tags using the 'long form'.
8865 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8866 STRING types. These convert content octets to and from the
8867 underlying type. The actual tag and length octets are
8868 already assumed to have been read in and checked. These
8869 are needed because all other string types have virtually
8870 identical handling apart from the tag. By having versions
8871 of the ASN1 functions that just operate on content octets
8872 IMPLICIT tagging can be handled properly. It also allows
8873 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8874 and ASN1_INTEGER are identical apart from the tag.
8877 *) Change the handling of OID objects as follows:
8879 - New object identifiers are inserted in objects.txt, following
8880 the syntax given in objects.README.
8881 - objects.pl is used to process obj_mac.num and create a new
8883 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8886 This is currently kind of a hack, and the perl code in objects.pl
8887 isn't very elegant, but it works as I intended. The simplest way
8888 to check that it worked correctly is to look in obj_dat.h and
8889 check the array nid_objs and make sure the objects haven't moved
8890 around (this is important!). Additions are OK, as well as
8891 consistent name changes.
8894 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8897 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8898 The given file adds to whatever has already been seeded into the
8899 random pool through the RANDFILE configuration file option or
8900 environment variable, or the default random state file.
8903 *) mkstack.pl now sorts each macro group into lexical order.
8904 Previously the output order depended on the order the files
8905 appeared in the directory, resulting in needless rewriting
8909 *) Patches to make OpenSSL compile under Win32 again. Mostly
8910 work arounds for the VC++ problem that it treats func() as
8911 func(void). Also stripped out the parts of mkdef.pl that
8912 added extra typesafe functions: these no longer exist.
8915 *) Reorganisation of the stack code. The macros are now all
8916 collected in safestack.h . Each macro is defined in terms of
8917 a "stack macro" of the form SKM_<name>(type, a, b). The
8918 DEBUG_SAFESTACK is now handled in terms of function casts,
8919 this has the advantage of retaining type safety without the
8920 use of additional functions. If DEBUG_SAFESTACK is not defined
8921 then the non typesafe macros are used instead. Also modified the
8922 mkstack.pl script to handle the new form. Needs testing to see
8923 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8924 the default if no major problems. Similar behaviour for ASN1_SET_OF
8925 and PKCS12_STACK_OF.
8928 *) When some versions of IIS use the 'NET' form of private key the
8929 key derivation algorithm is different. Normally MD5(password) is
8930 used as a 128 bit RC4 key. In the modified case
8931 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8932 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8933 as the old Netscape_RSA functions except they have an additional
8934 'sgckey' parameter which uses the modified algorithm. Also added
8935 an -sgckey command line option to the rsa utility. Thanks to
8936 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8937 algorithm to openssl-dev.
8940 *) The evp_local.h macros were using 'c.##kname' which resulted in
8941 invalid expansion on some systems (SCO 5.0.5 for example).
8942 Corrected to 'c.kname'.
8943 [Phillip Porch <root@theporch.com>]
8945 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8946 a STACK of email addresses from a certificate or request, these look
8947 in the subject name and the subject alternative name extensions and
8948 omit any duplicate addresses.
8951 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8952 This makes DSA verification about 2 % faster.
8955 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8956 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8957 plus overhead for 1024 bit moduli).
8958 This makes exponentiations about 0.5 % faster for 1024 bit
8959 exponents (as measured by "openssl speed rsa2048").
8962 *) Rename memory handling macros to avoid conflicts with other
8964 Malloc => OPENSSL_malloc
8965 Malloc_locked => OPENSSL_malloc_locked
8966 Realloc => OPENSSL_realloc
8967 Free => OPENSSL_free
8970 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8971 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8974 *) CygWin32 support.
8975 [John Jarvie <jjarvie@newsguy.com>]
8977 *) The type-safe stack code has been rejigged. It is now only compiled
8978 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8979 by default all type-specific stack functions are "#define"d back to
8980 standard stack functions. This results in more streamlined output
8981 but retains the type-safety checking possibilities of the original
8985 *) The STACK code has been cleaned up, and certain type declarations
8986 that didn't make a lot of sense have been brought in line. This has
8987 also involved a cleanup of sorts in safestack.h to more correctly
8988 map type-safe stack functions onto their plain stack counterparts.
8989 This work has also resulted in a variety of "const"ifications of
8990 lots of the code, especially "_cmp" operations which should normally
8991 be prototyped with "const" parameters anyway.
8994 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8995 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8996 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8997 where all of 'md' is used each time the PRNG is used, but 'state'
8998 is used only indexed by a cyclic counter. As entropy may not be
8999 well distributed from the beginning, 'md' is important as a
9000 chaining variable. However, the output function chains only half
9001 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9002 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9003 in all of 'state' being rewritten, with the new values depending
9004 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9007 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9008 the handshake is continued after ssl_verify_cert_chain();
9009 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9010 can lead to 'unexplainable' connection aborts later.
9011 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9013 *) Major EVP API cipher revision.
9014 Add hooks for extra EVP features. This allows various cipher
9015 parameters to be set in the EVP interface. Support added for variable
9016 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9017 setting of RC2 and RC5 parameters.
9019 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9022 Remove lots of duplicated code from the EVP library. For example *every*
9023 cipher init() function handles the 'iv' in the same way according to the
9024 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9025 for CFB and OFB modes they zero ctx->num.
9027 New functionality allows removal of S/MIME code RC2 hack.
9029 Most of the routines have the same form and so can be declared in terms
9032 By shifting this to the top level EVP_CipherInit() it can be removed from
9033 all individual ciphers. If the cipher wants to handle IVs or keys
9034 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9037 Change lots of functions like EVP_EncryptUpdate() to now return a
9038 value: although software versions of the algorithms cannot fail
9039 any installed hardware versions can.
9042 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9043 this option is set, tolerate broken clients that send the negotiated
9044 protocol version number instead of the requested protocol version
9048 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9049 i.e. non-zero for export ciphersuites, zero otherwise.
9050 Previous versions had this flag inverted, inconsistent with
9051 rsa_tmp_cb (..._TMP_RSA_CB).
9052 [Bodo Moeller; problem reported by Amit Chopra]
9054 *) Add missing DSA library text string. Work around for some IIS
9055 key files with invalid SEQUENCE encoding.
9058 *) Add a document (doc/standards.txt) that list all kinds of standards
9059 and so on that are implemented in OpenSSL.
9062 *) Enhance c_rehash script. Old version would mishandle certificates
9063 with the same subject name hash and wouldn't handle CRLs at all.
9064 Added -fingerprint option to crl utility, to support new c_rehash
9068 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9071 *) Fix for SSL server purpose checking. Server checking was
9072 rejecting certificates which had extended key usage present
9073 but no ssl client purpose.
9074 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9076 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9077 is a little unclear about how a blank password is handled.
9078 Since the password in encoded as a BMPString with terminating
9079 double NULL a zero length password would end up as just the
9080 double NULL. However no password at all is different and is
9081 handled differently in the PKCS#12 key generation code. NS
9082 treats a blank password as zero length. MSIE treats it as no
9083 password on export: but it will try both on import. We now do
9084 the same: PKCS12_parse() tries zero length and no password if
9085 the password is set to "" or NULL (NULL is now a valid password:
9086 it wasn't before) as does the pkcs12 application.
9089 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9090 perror when PEM_read_bio_X509_REQ fails, the error message must
9091 be obtained from the error queue.
9094 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9095 it in ERR_remove_state if appropriate, and change ERR_get_state
9096 accordingly to avoid race conditions (this is necessary because
9097 thread_hash is no longer constant once set).
9100 *) Bugfix for linux-elf makefile.one.
9103 *) RSA_get_default_method() will now cause a default
9104 RSA_METHOD to be chosen if one doesn't exist already.
9105 Previously this was only set during a call to RSA_new()
9106 or RSA_new_method(NULL) meaning it was possible for
9107 RSA_get_default_method() to return NULL.
9110 *) Added native name translation to the existing DSO code
9111 that will convert (if the flag to do so is set) filenames
9112 that are sufficiently small and have no path information
9113 into a canonical native form. Eg. "blah" converted to
9114 "libblah.so" or "blah.dll" etc.
9117 *) New function ERR_error_string_n(e, buf, len) which is like
9118 ERR_error_string(e, buf), but writes at most 'len' bytes
9119 including the 0 terminator. For ERR_error_string_n, 'buf'
9121 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9123 *) CONF library reworked to become more general. A new CONF
9124 configuration file reader "class" is implemented as well as a
9125 new functions (NCONF_*, for "New CONF") to handle it. The now
9126 old CONF_* functions are still there, but are reimplemented to
9127 work in terms of the new functions. Also, a set of functions
9128 to handle the internal storage of the configuration data is
9129 provided to make it easier to write new configuration file
9130 reader "classes" (I can definitely see something reading a
9131 configuration file in XML format, for example), called _CONF_*,
9132 or "the configuration storage API"...
9134 The new configuration file reading functions are:
9136 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9137 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9139 NCONF_default, NCONF_WIN32
9141 NCONF_dump_fp, NCONF_dump_bio
9143 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9144 NCONF_new creates a new CONF object. This works in the same way
9145 as other interfaces in OpenSSL, like the BIO interface.
9146 NCONF_dump_* dump the internal storage of the configuration file,
9147 which is useful for debugging. All other functions take the same
9148 arguments as the old CONF_* functions wth the exception of the
9149 first that must be a `CONF *' instead of a `LHASH *'.
9151 To make it easer to use the new classes with the old CONF_* functions,
9152 the function CONF_set_default_method is provided.
9155 *) Add '-tls1' option to 'openssl ciphers', which was already
9156 mentioned in the documentation but had not been implemented.
9157 (This option is not yet really useful because even the additional
9158 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9161 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9162 OpenSSL-based applications) load shared libraries and bind to
9163 them in a portable way.
9164 [Geoff Thorpe, with contributions from Richard Levitte]
9166 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9168 *) Make sure _lrotl and _lrotr are only used with MSVC.
9170 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9171 (the default implementation of RAND_status).
9173 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9174 to '-clrext' (= clear extensions), as intended and documented.
9175 [Bodo Moeller; inconsistency pointed out by Michael Attili
9178 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9179 was larger than the MD block size.
9180 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9182 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9183 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9184 using the passed key: if the passed key was a private key the result
9185 of X509_print(), for example, would be to print out all the private key
9189 *) des_quad_cksum() byte order bug fix.
9190 [Ulf Möller, using the problem description in krb4-0.9.7, where
9191 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9193 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9195 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9197 *) For easily testing in shell scripts whether some command
9198 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9199 returns with exit code 0 iff no command of the given name is available.
9200 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9201 the output goes to stdout and nothing is printed to stderr.
9202 Additional arguments are always ignored.
9204 Since for each cipher there is a command of the same name,
9205 the 'no-cipher' compilation switches can be tested this way.
9207 ('openssl no-XXX' is not able to detect pseudo-commands such
9208 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9211 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9214 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9215 is set; it will be thrown away anyway because each handshake creates
9217 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9218 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9219 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9220 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9223 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9224 'Q' and 'R' lose their special meanings (quit/renegotiate).
9225 This is part of what -quiet does; unlike -quiet, -ign_eof
9226 does not suppress any output.
9229 *) Add compatibility options to the purpose and trust code. The
9230 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9231 accepts a certificate or CA, this was the previous behaviour,
9232 with all the associated security issues.
9234 X509_TRUST_COMPAT is the old trust behaviour: only and
9235 automatically trust self signed roots in certificate store. A
9236 new trust setting X509_TRUST_DEFAULT is used to specify that
9237 a purpose has no associated trust setting and it should instead
9238 use the value in the default purpose.
9241 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9242 and fix a memory leak.
9245 *) In util/mkerr.pl (which implements 'make errors'), preserve
9246 reason strings from the previous version of the .c file, as
9247 the default to have only downcase letters (and digits) in
9248 automatically generated reasons codes is not always appropriate.
9251 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9252 using strerror. Previously, ERR_reason_error_string() returned
9253 library names as reason strings for SYSerr; but SYSerr is a special
9254 case where small numbers are errno values, not library numbers.
9257 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9258 converts DSA parameters into DH parameters. (When creating parameters,
9259 DSA_generate_parameters is used.)
9262 *) Include 'length' (recommended exponent length) in C code generated
9263 by 'openssl dhparam -C'.
9266 *) The second argument to set_label in perlasm was already being used
9267 so couldn't be used as a "file scope" flag. Moved to third argument
9271 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9272 instead of RAND_bytes for encryption IVs and salts.
9275 *) Include RAND_status() into RAND_METHOD instead of implementing
9276 it only for md_rand.c Otherwise replacing the PRNG by calling
9277 RAND_set_rand_method would be impossible.
9280 *) Don't let DSA_generate_key() enter an infinite loop if the random
9281 number generation fails.
9284 *) New 'rand' application for creating pseudo-random output.
9287 *) Added configuration support for Linux/IA64
9288 [Rolf Haberrecker <rolf@suse.de>]
9290 *) Assembler module support for Mingw32.
9293 *) Shared library support for HPUX (in shlib/).
9294 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9296 *) Shared library support for Solaris gcc.
9297 [Lutz Behnke <behnke@trustcenter.de>]
9299 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9301 *) PKCS7_encrypt() was adding text MIME headers twice because they
9302 were added manually and by SMIME_crlf_copy().
9305 *) In bntest.c don't call BN_rand with zero bits argument.
9306 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9308 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9309 case was implemented. This caused BN_div_recp() to fail occasionally.
9312 *) Add an optional second argument to the set_label() in the perl
9313 assembly language builder. If this argument exists and is set
9314 to 1 it signals that the assembler should use a symbol whose
9315 scope is the entire file, not just the current function. This
9316 is needed with MASM which uses the format label:: for this scope.
9317 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9319 *) Change the ASN1 types so they are typedefs by default. Before
9320 almost all types were #define'd to ASN1_STRING which was causing
9321 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9325 *) Change names of new functions to the new get1/get0 naming
9326 convention: After 'get1', the caller owns a reference count
9327 and has to call ..._free; 'get0' returns a pointer to some
9328 data structure without incrementing reference counters.
9329 (Some of the existing 'get' functions increment a reference
9330 counter, some don't.)
9331 Similarly, 'set1' and 'add1' functions increase reference
9332 counters or duplicate objects.
9335 *) Allow for the possibility of temp RSA key generation failure:
9336 the code used to assume it always worked and crashed on failure.
9339 *) Fix potential buffer overrun problem in BIO_printf().
9340 [Ulf Möller, using public domain code by Patrick Powell; problem
9341 pointed out by David Sacerdote <das33@cornell.edu>]
9343 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9344 RAND_egd() and RAND_status(). In the command line application,
9345 the EGD socket can be specified like a seed file using RANDFILE
9349 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9350 Some CAs (e.g. Verisign) distribute certificates in this form.
9353 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9354 list to exclude them. This means that no special compilation option
9355 is needed to use anonymous DH: it just needs to be included in the
9359 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9360 EVP_MD_type. The old functionality is available in a new macro called
9361 EVP_MD_md(). Change code that uses it and update docs.
9364 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9365 where the 'void *' argument is replaced by a function pointer argument.
9366 Previously 'void *' was abused to point to functions, which works on
9367 many platforms, but is not correct. As these functions are usually
9368 called by macros defined in OpenSSL header files, most source code
9369 should work without changes.
9372 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9373 sections with information on -D... compiler switches used for
9374 compiling the library so that applications can see them. To enable
9375 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9376 must be defined. E.g.,
9377 #define OPENSSL_ALGORITHM_DEFINES
9378 #include <openssl/opensslconf.h>
9379 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9380 [Richard Levitte, Ulf and Bodo Möller]
9382 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9386 *) Change the 'other' type in certificate aux info to a STACK_OF
9387 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9388 the required ASN1 format: arbitrary types determined by an OID.
9391 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9392 argument to 'req'. This is not because the function is newer or
9393 better than others it just uses the work 'NEW' in the certificate
9394 request header lines. Some software needs this.
9397 *) Reorganise password command line arguments: now passwords can be
9398 obtained from various sources. Delete the PEM_cb function and make
9399 it the default behaviour: i.e. if the callback is NULL and the
9400 usrdata argument is not NULL interpret it as a null terminated pass
9401 phrase. If usrdata and the callback are NULL then the pass phrase
9402 is prompted for as usual.
9405 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9406 the support is automatically enabled. The resulting binaries will
9407 autodetect the card and use it if present.
9408 [Ben Laurie and Compaq Inc.]
9410 *) Work around for Netscape hang bug. This sends certificate request
9411 and server done in one record. Since this is perfectly legal in the
9412 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9413 the bugs/SSLv3 entry for more info.
9416 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9419 *) Add -rand argument to smime and pkcs12 applications and read/write
9423 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9426 *) Add command line password options to the remaining applications.
9429 *) Bug fix for BN_div_recp() for numerators with an even number of
9433 *) More tests in bntest.c, and changed test_bn output.
9436 *) ./config recognizes MacOS X now.
9439 *) Bug fix for BN_div() when the first words of num and divsor are
9440 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9443 *) Add support for various broken PKCS#8 formats, and command line
9444 options to produce them.
9447 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9448 get temporary BIGNUMs from a BN_CTX.
9451 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9455 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9456 include a #define from the old name to the new. The original intent
9457 was that statically linked binaries could for example just call
9458 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9459 link with digests. This never worked because SSLeay_add_all_digests()
9460 and SSLeay_add_all_ciphers() were in the same source file so calling
9461 one would link with the other. They are now in separate source files.
9464 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9467 *) Use a less unusual form of the Miller-Rabin primality test (it used
9468 a binary algorithm for exponentiation integrated into the Miller-Rabin
9469 loop, our standard modexp algorithms are faster).
9472 *) Support for the EBCDIC character set completed.
9473 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9475 *) Source code cleanups: use const where appropriate, eliminate casts,
9476 use void * instead of char * in lhash.
9479 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9480 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9481 this the server could overwrite ephemeral keys that the client
9485 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9486 using 50 iterations of the Rabin-Miller test.
9488 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9489 iterations of the Rabin-Miller test as required by the appendix
9490 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9491 As BN_is_prime_fasttest includes trial division, DSA parameter
9492 generation becomes much faster.
9494 This implies a change for the callback functions in DSA_is_prime
9495 and DSA_generate_parameters: The callback function is called once
9496 for each positive witness in the Rabin-Miller test, not just
9497 occasionally in the inner loop; and the parameters to the
9498 callback function now provide an iteration count for the outer
9499 loop rather than for the current invocation of the inner loop.
9500 DSA_generate_parameters additionally can call the callback
9501 function with an 'iteration count' of -1, meaning that a
9502 candidate has passed the trial division test (when q is generated
9503 from an application-provided seed, trial division is skipped).
9506 *) New function BN_is_prime_fasttest that optionally does trial
9507 division before starting the Rabin-Miller test and has
9508 an additional BN_CTX * argument (whereas BN_is_prime always
9509 has to allocate at least one BN_CTX).
9510 'callback(1, -1, cb_arg)' is called when a number has passed the
9511 trial division stage.
9514 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9518 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9521 *) New function BN_pseudo_rand().
9524 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9525 bignum version of BN_from_montgomery() with the working code from
9526 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9530 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9531 made it impossible to use the same SSL_SESSION data structure in
9532 SSL2 clients in multiple threads.
9535 *) The return value of RAND_load_file() no longer counts bytes obtained
9536 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9537 to seed the PRNG (previously an explicit byte count was required).
9538 [Ulf Möller, Bodo Möller]
9540 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9541 used (char *) instead of (void *) and had casts all over the place.
9544 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9547 *) Retain source code compatibility for BN_prime_checks macro:
9548 BN_is_prime(..., BN_prime_checks, ...) now uses
9549 BN_prime_checks_for_size to determine the appropriate number of
9550 Rabin-Miller iterations.
9553 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9554 DH_CHECK_P_NOT_SAFE_PRIME.
9555 (Check if this is true? OpenPGP calls them "strong".)
9558 *) Merge the functionality of "dh" and "gendh" programs into a new program
9559 "dhparam". The old programs are retained for now but will handle DH keys
9560 (instead of parameters) in future.
9563 *) Make the ciphers, s_server and s_client programs check the return values
9564 when a new cipher list is set.
9567 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9568 ciphers. Before when the 56bit ciphers were enabled the sorting was
9571 The syntax for the cipher sorting has been extended to support sorting by
9572 cipher-strength (using the strength_bits hard coded in the tables).
9573 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9575 Fix a bug in the cipher-command parser: when supplying a cipher command
9576 string with an "undefined" symbol (neither command nor alphanumeric
9577 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9578 an error is flagged.
9580 Due to the strength-sorting extension, the code of the
9581 ssl_create_cipher_list() function was completely rearranged. I hope that
9582 the readability was also increased :-)
9583 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9585 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9586 for the first serial number and places 2 in the serial number file. This
9587 avoids problems when the root CA is created with serial number zero and
9588 the first user certificate has the same issuer name and serial number
9592 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9593 the new code. Add documentation for this stuff.
9596 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9597 X509_*() to X509at_*() on the grounds that they don't handle X509
9598 structures and behave in an analogous way to the X509v3 functions:
9599 they shouldn't be called directly but wrapper functions should be used
9602 So we also now have some wrapper functions that call the X509at functions
9603 when passed certificate requests. (TO DO: similar things can be done with
9604 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9605 things. Some of these need some d2i or i2d and print functionality
9606 because they handle more complex structures.)
9609 *) Add missing #ifndefs that caused missing symbols when building libssl
9610 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9611 NO_RSA in ssl/s2*.c.
9612 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9614 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9615 has a return value which indicates the quality of the random data
9616 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9617 error queue. New function RAND_pseudo_bytes() generates output that is
9618 guaranteed to be unique but not unpredictable. RAND_add is like
9619 RAND_seed, but takes an extra argument for an entropy estimate
9620 (RAND_seed always assumes full entropy).
9623 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9624 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9625 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9626 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9627 false-positive rate of at most 2^-80 for random input.
9630 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9633 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9634 in the 0.9.5 release), this returns the chain
9635 from an X509_CTX structure with a dup of the stack and all
9636 the X509 reference counts upped: so the stack will exist
9637 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9640 Also make SSL_SESSION_print() print out the verify return
9644 *) Add manpage for the pkcs12 command. Also change the default
9645 behaviour so MAC iteration counts are used unless the new
9646 -nomaciter option is used. This improves file security and
9647 only older versions of MSIE (4.0 for example) need it.
9650 *) Honor the no-xxx Configure options when creating .DEF files.
9653 *) Add PKCS#10 attributes to field table: challengePassword,
9654 unstructuredName and unstructuredAddress. These are taken from
9655 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9656 international characters are used.
9658 More changes to X509_ATTRIBUTE code: allow the setting of types
9659 based on strings. Remove the 'loc' parameter when adding
9660 attributes because these will be a SET OF encoding which is sorted
9664 *) Initial changes to the 'req' utility to allow request generation
9665 automation. This will allow an application to just generate a template
9666 file containing all the field values and have req construct the
9669 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9670 used all over the place including certificate requests and PKCS#7
9671 structures. They are currently handled manually where necessary with
9672 some primitive wrappers for PKCS#7. The new functions behave in a
9673 manner analogous to the X509 extension functions: they allow
9674 attributes to be looked up by NID and added.
9676 Later something similar to the X509V3 code would be desirable to
9677 automatically handle the encoding, decoding and printing of the
9678 more complex types. The string types like challengePassword can
9679 be handled by the string table functions.
9681 Also modified the multi byte string table handling. Now there is
9682 a 'global mask' which masks out certain types. The table itself
9683 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9684 is useful when for example there is only one permissible type
9685 (as in countryName) and using the mask might result in no valid
9689 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9690 SSL_get_peer_finished to allow applications to obtain the latest
9691 Finished messages sent to the peer or expected from the peer,
9692 respectively. (SSL_get_peer_finished is usually the Finished message
9693 actually received from the peer, otherwise the protocol will be aborted.)
9695 As the Finished message are message digests of the complete handshake
9696 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9697 be used for external authentication procedures when the authentication
9698 provided by SSL/TLS is not desired or is not enough.
9701 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9702 the host supports BWX extension and if Compaq C is present on the
9703 $PATH. Just exploiting of the BWX extension results in 20-30%
9704 performance kick for some algorithms, e.g. DES and RC4 to mention
9705 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9709 *) Add support for MS "fast SGC". This is arguably a violation of the
9710 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9711 weak crypto and after checking the certificate is SGC a second one
9712 with strong crypto. MS SGC stops the first handshake after receiving
9713 the server certificate message and sends a second client hello. Since
9714 a server will typically do all the time consuming operations before
9715 expecting any further messages from the client (server key exchange
9716 is the most expensive) there is little difference between the two.
9718 To get OpenSSL to support MS SGC we have to permit a second client
9719 hello message after we have sent server done. In addition we have to
9720 reset the MAC if we do get this second client hello.
9723 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9724 if a DER encoded private key is RSA or DSA traditional format. Changed
9725 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9726 format DER encoded private key. Newer code should use PKCS#8 format which
9727 has the key type encoded in the ASN1 structure. Added DER private key
9728 support to pkcs8 application.
9731 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9732 ciphersuites has been selected (as required by the SSL 3/TLS 1
9733 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9734 is set, we interpret this as a request to violate the specification
9735 (the worst that can happen is a handshake failure, and 'correct'
9736 behaviour would result in a handshake failure anyway).
9739 *) In SSL_CTX_add_session, take into account that there might be multiple
9740 SSL_SESSION structures with the same session ID (e.g. when two threads
9741 concurrently obtain them from an external cache).
9742 The internal cache can handle only one SSL_SESSION with a given ID,
9743 so if there's a conflict, we now throw out the old one to achieve
9747 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9748 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9749 some routines that use cipher OIDs: some ciphers do not have OIDs
9750 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9754 *) Simplify the trust setting structure and code. Now we just have
9755 two sequences of OIDs for trusted and rejected settings. These will
9756 typically have values the same as the extended key usage extension
9757 and any application specific purposes.
9759 The trust checking code now has a default behaviour: it will just
9760 check for an object with the same NID as the passed id. Functions can
9761 be provided to override either the default behaviour or the behaviour
9762 for a given id. SSL client, server and email already have functions
9763 in place for compatibility: they check the NID and also return "trusted"
9764 if the certificate is self signed.
9767 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9768 traditional format into an EVP_PKEY structure.
9771 *) Add a password callback function PEM_cb() which either prompts for
9772 a password if usr_data is NULL or otherwise assumes it is a null
9773 terminated password. Allow passwords to be passed on command line
9774 environment or config files in a few more utilities.
9777 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9778 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9779 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9780 Update documentation.
9783 *) Support for ASN1 "NULL" type. This could be handled before by using
9784 ASN1_TYPE but there wasn't any function that would try to read a NULL
9785 and produce an error if it couldn't. For compatibility we also have
9786 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9787 don't allocate anything because they don't need to.
9790 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9792 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9794 *) Rebuild of the memory allocation routines used by OpenSSL code and
9795 possibly others as well. The purpose is to make an interface that
9796 provide hooks so anyone can build a separate set of allocation and
9797 deallocation routines to be used by OpenSSL, for example memory
9798 pool implementations, or something else, which was previously hard
9799 since Malloc(), Realloc() and Free() were defined as macros having
9800 the values malloc, realloc and free, respectively (except for Win32
9801 compilations). The same is provided for memory debugging code.
9802 OpenSSL already comes with functionality to find memory leaks, but
9803 this gives people a chance to debug other memory problems.
9805 With these changes, a new set of functions and macros have appeared:
9807 CRYPTO_set_mem_debug_functions() [F]
9808 CRYPTO_get_mem_debug_functions() [F]
9809 CRYPTO_dbg_set_options() [F]
9810 CRYPTO_dbg_get_options() [F]
9811 CRYPTO_malloc_debug_init() [M]
9813 The memory debug functions are NULL by default, unless the library
9814 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9815 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9816 gives the standard debugging functions that come with OpenSSL) or
9817 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9818 provided by the library user) must be used. When the standard
9819 debugging functions are used, CRYPTO_dbg_set_options can be used to
9820 request additional information:
9821 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9822 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9824 Also, things like CRYPTO_set_mem_functions will always give the
9825 expected result (the new set of functions is used for allocation
9826 and deallocation) at all times, regardless of platform and compiler
9829 To finish it up, some functions that were never use in any other
9830 way than through macros have a new API and new semantic:
9833 CRYPTO_dbg_realloc()
9836 All macros of value have retained their old syntax.
9837 [Richard Levitte and Bodo Moeller]
9839 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9840 ordering of SMIMECapabilities wasn't in "strength order" and there
9841 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9845 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9846 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9847 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9849 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9850 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9851 functionality to handle multipart/signed properly) and a utility
9852 called 'smime' to call all this stuff. This is based on code I
9853 originally wrote for Celo who have kindly allowed it to be
9854 included in OpenSSL.
9857 *) Add variants des_set_key_checked and des_set_key_unchecked of
9858 des_set_key (aka des_key_sched). Global variable des_check_key
9859 decides which of these is called by des_set_key; this way
9860 des_check_key behaves as it always did, but applications and
9861 the library itself, which was buggy for des_check_key == 1,
9862 have a cleaner way to pick the version they need.
9865 *) New function PKCS12_newpass() which changes the password of a
9869 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9870 dynamic mix. In both cases the ids can be used as an index into the
9871 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9872 functions so they accept a list of the field values and the
9873 application doesn't need to directly manipulate the X509_TRUST
9877 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9881 *) Modify the way the V3 extension code looks up extensions. This now
9882 works in a similar way to the object code: we have some "standard"
9883 extensions in a static table which is searched with OBJ_bsearch()
9884 and the application can add dynamic ones if needed. The file
9885 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9886 updated whenever a new extension is added to the core code and kept
9887 in ext_nid order. There is a simple program 'tabtest.c' which checks
9888 this. New extensions are not added too often so this file can readily
9889 be maintained manually.
9891 There are two big advantages in doing things this way. The extensions
9892 can be looked up immediately and no longer need to be "added" using
9893 X509V3_add_standard_extensions(): this function now does nothing.
9894 [Side note: I get *lots* of email saying the extension code doesn't
9895 work because people forget to call this function]
9896 Also no dynamic allocation is done unless new extensions are added:
9897 so if we don't add custom extensions there is no need to call
9898 X509V3_EXT_cleanup().
9901 *) Modify enc utility's salting as follows: make salting the default. Add a
9902 magic header, so unsalted files fail gracefully instead of just decrypting
9903 to garbage. This is because not salting is a big security hole, so people
9904 should be discouraged from doing it.
9907 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9908 digest to be passed on the command line but it only used this
9909 parameter when signing a certificate. Modified so all relevant
9910 operations are affected by the digest parameter including the
9911 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9912 DSA key was used because it didn't fix the digest.
9915 *) Initial certificate chain verify code. Currently tests the untrusted
9916 certificates for consistency with the verify purpose (which is set
9917 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9919 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9920 this is because it will reject chains with invalid extensions whereas
9921 every previous version of OpenSSL and SSLeay made no checks at all.
9923 Trust code: checks the root CA for the relevant trust settings. Trust
9924 settings have an initial value consistent with the verify purpose: e.g.
9925 if the verify purpose is for SSL client use it expects the CA to be
9926 trusted for SSL client use. However the default value can be changed to
9927 permit custom trust settings: one example of this would be to only trust
9928 certificates from a specific "secure" set of CAs.
9930 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9931 which should be used for version portability: especially since the
9932 verify structure is likely to change more often now.
9934 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9935 to set them. If not set then assume SSL clients will verify SSL servers
9938 Two new options to the verify program: -untrusted allows a set of
9939 untrusted certificates to be passed in and -purpose which sets the
9940 intended purpose of the certificate. If a purpose is set then the
9941 new chain verify code is used to check extension consistency.
9944 *) Support for the authority information access extension.
9947 *) Modify RSA and DSA PEM read routines to transparently handle
9948 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9949 public keys in a format compatible with certificate
9950 SubjectPublicKeyInfo structures. Unfortunately there were already
9951 functions called *_PublicKey_* which used various odd formats so
9952 these are retained for compatibility: however the DSA variants were
9953 never in a public release so they have been deleted. Changed dsa/rsa
9954 utilities to handle the new format: note no releases ever handled public
9955 keys so we should be OK.
9957 The primary motivation for this change is to avoid the same fiasco
9958 that dogs private keys: there are several incompatible private key
9959 formats some of which are standard and some OpenSSL specific and
9960 require various evil hacks to allow partial transparent handling and
9961 even then it doesn't work with DER formats. Given the option anything
9962 other than PKCS#8 should be dumped: but the other formats have to
9963 stay in the name of compatibility.
9965 With public keys and the benefit of hindsight one standard format
9966 is used which works with EVP_PKEY, RSA or DSA structures: though
9967 it clearly returns an error if you try to read the wrong kind of key.
9969 Added a -pubkey option to the 'x509' utility to output the public key.
9970 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9971 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9972 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9973 that do the same as the EVP_PKEY_assign_*() except they up the
9974 reference count of the added key (they don't "swallow" the
9978 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9979 CRLs would fail if the file contained no certificates or no CRLs:
9980 added a new function to read in both types and return the number
9981 read: this means that if none are read it will be an error. The
9982 DER versions of the certificate and CRL reader would always fail
9983 because it isn't possible to mix certificates and CRLs in DER format
9984 without choking one or the other routine. Changed this to just read
9985 a certificate: this is the best we can do. Also modified the code
9986 in apps/verify.c to take notice of return codes: it was previously
9987 attempting to read in certificates from NULL pointers and ignoring
9988 any errors: this is one reason why the cert and CRL reader seemed
9989 to work. It doesn't check return codes from the default certificate
9990 routines: these may well fail if the certificates aren't installed.
9993 *) Code to support otherName option in GeneralName.
9996 *) First update to verify code. Change the verify utility
9997 so it warns if it is passed a self signed certificate:
9998 for consistency with the normal behaviour. X509_verify
9999 has been modified to it will now verify a self signed
10000 certificate if *exactly* the same certificate appears
10001 in the store: it was previously impossible to trust a
10002 single self signed certificate. This means that:
10003 openssl verify ss.pem
10004 now gives a warning about a self signed certificate but
10005 openssl verify -CAfile ss.pem ss.pem
10009 *) For servers, store verify_result in SSL_SESSION data structure
10010 (and add it to external session representation).
10011 This is needed when client certificate verifications fails,
10012 but an application-provided verification callback (set by
10013 SSL_CTX_set_cert_verify_callback) allows accepting the session
10014 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10015 but returns 1): When the session is reused, we have to set
10016 ssl->verify_result to the appropriate error code to avoid
10018 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10020 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10021 case in PKCS7_dataInit() where the signed PKCS7 structure
10022 didn't contain any existing data because it was being created.
10023 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10025 *) Add a salt to the key derivation routines in enc.c. This
10026 forms the first 8 bytes of the encrypted file. Also add a
10027 -S option to allow a salt to be input on the command line.
10030 *) New function X509_cmp(). Oddly enough there wasn't a function
10031 to compare two certificates. We do this by working out the SHA1
10032 hash and comparing that. X509_cmp() will be needed by the trust
10036 *) SSL_get1_session() is like SSL_get_session(), but increments
10037 the reference count in the SSL_SESSION returned.
10038 [Geoff Thorpe <geoff@eu.c2.net>]
10040 *) Fix for 'req': it was adding a null to request attributes.
10041 Also change the X509_LOOKUP and X509_INFO code to handle
10042 certificate auxiliary information.
10045 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10049 *) Add the possibility to add extra information to the memory leak
10050 detecting output, to form tracebacks, showing from where each
10051 allocation was originated: CRYPTO_push_info("constant string") adds
10052 the string plus current file name and line number to a per-thread
10053 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10054 is like calling CYRPTO_pop_info() until the stack is empty.
10055 Also updated memory leak detection code to be multi-thread-safe.
10058 *) Add options -text and -noout to pkcs7 utility and delete the
10059 encryption options which never did anything. Update docs.
10062 *) Add options to some of the utilities to allow the pass phrase
10063 to be included on either the command line (not recommended on
10064 OSes like Unix) or read from the environment. Update the
10065 manpages and fix a few bugs.
10068 *) Add a few manpages for some of the openssl commands.
10071 *) Fix the -revoke option in ca. It was freeing up memory twice,
10072 leaking and not finding already revoked certificates.
10075 *) Extensive changes to support certificate auxiliary information.
10076 This involves the use of X509_CERT_AUX structure and X509_AUX
10077 functions. An X509_AUX function such as PEM_read_X509_AUX()
10078 can still read in a certificate file in the usual way but it
10079 will also read in any additional "auxiliary information". By
10080 doing things this way a fair degree of compatibility can be
10081 retained: existing certificates can have this information added
10082 using the new 'x509' options.
10084 Current auxiliary information includes an "alias" and some trust
10085 settings. The trust settings will ultimately be used in enhanced
10086 certificate chain verification routines: currently a certificate
10087 can only be trusted if it is self signed and then it is trusted
10091 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10092 The problem was that one of the replacement routines had not been working
10093 since SSLeay releases. For now the offending routine has been replaced
10094 with non-optimised assembler. Even so, this now gives around 95%
10095 performance improvement for 1024 bit RSA signs.
10098 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10099 handling. Most clients have the effective key size in bits equal to
10100 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10101 A few however don't do this and instead use the size of the decrypted key
10102 to determine the RC2 key length and the AlgorithmIdentifier to determine
10103 the effective key length. In this case the effective key length can still
10104 be 40 bits but the key length can be 168 bits for example. This is fixed
10105 by manually forcing an RC2 key into the EVP_PKEY structure because the
10106 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10107 the key length and effective key length are equal.
10110 *) Add a bunch of functions that should simplify the creation of
10111 X509_NAME structures. Now you should be able to do:
10112 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10113 and have it automatically work out the correct field type and fill in
10114 the structures. The more adventurous can try:
10115 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10116 and it will (hopefully) work out the correct multibyte encoding.
10119 *) Change the 'req' utility to use the new field handling and multibyte
10120 copy routines. Before the DN field creation was handled in an ad hoc
10121 way in req, ca, and x509 which was rather broken and didn't support
10122 BMPStrings or UTF8Strings. Since some software doesn't implement
10123 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10124 using the dirstring_type option. See the new comment in the default
10125 openssl.cnf for more info.
10128 *) Make crypto/rand/md_rand.c more robust:
10129 - Assure unique random numbers after fork().
10130 - Make sure that concurrent threads access the global counter and
10131 md serializably so that we never lose entropy in them
10132 or use exactly the same state in multiple threads.
10133 Access to the large state is not always serializable because
10134 the additional locking could be a performance killer, and
10135 md should be large enough anyway.
10138 *) New file apps/app_rand.c with commonly needed functionality
10139 for handling the random seed file.
10141 Use the random seed file in some applications that previously did not:
10143 dsaparam -genkey (which also ignored its '-rand' option),
10146 x509 (when signing).
10147 Except on systems with /dev/urandom, it is crucial to have a random
10148 seed file at least for key creation, DSA signing, and for DH exchanges;
10149 for RSA signatures we could do without one.
10151 gendh and gendsa (unlike genrsa) used to read only the first byte
10152 of each file listed in the '-rand' option. The function as previously
10153 found in genrsa is now in app_rand.c and is used by all programs
10154 that support '-rand'.
10157 *) In RAND_write_file, use mode 0600 for creating files;
10158 don't just chmod when it may be too late.
10161 *) Report an error from X509_STORE_load_locations
10162 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10165 *) New function ASN1_mbstring_copy() this copies a string in either
10166 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10167 into an ASN1_STRING type. A mask of permissible types is passed
10168 and it chooses the "minimal" type to use or an error if not type
10172 *) Add function equivalents to the various macros in asn1.h. The old
10173 macros are retained with an M_ prefix. Code inside the library can
10174 use the M_ macros. External code (including the openssl utility)
10175 should *NOT* in order to be "shared library friendly".
10178 *) Add various functions that can check a certificate's extensions
10179 to see if it usable for various purposes such as SSL client,
10180 server or S/MIME and CAs of these types. This is currently
10181 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10182 verification. Also added a -purpose flag to x509 utility to
10183 print out all the purposes.
10186 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10190 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10191 for, obtain and decode and extension and obtain its critical flag.
10192 This allows all the necessary extension code to be handled in a
10193 single function call.
10196 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10197 platforms. See crypto/rc4/rc4_enc.c for further details.
10200 *) New -noout option to asn1parse. This causes no output to be produced
10201 its main use is when combined with -strparse and -out to extract data
10202 from a file (which may not be in ASN.1 format).
10205 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10206 when producing the local key id.
10207 [Richard Levitte <levitte@stacken.kth.se>]
10209 *) New option -dhparam in s_server. This allows a DH parameter file to be
10210 stated explicitly. If it is not stated then it tries the first server
10211 certificate file. The previous behaviour hard coded the filename
10215 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10216 a public key to be input or output. For example:
10217 openssl rsa -in key.pem -pubout -out pubkey.pem
10218 Also added necessary DSA public key functions to handle this.
10221 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10222 in the message. This was handled by allowing
10223 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10224 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10226 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10227 to the end of the strings whereas this didn't. This would cause problems
10228 if strings read with d2i_ASN1_bytes() were later modified.
10229 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10231 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10232 data and it contains EOF it will end up returning an error. This is
10233 caused by input 46 bytes long. The cause is due to the way base64
10234 BIOs find the start of base64 encoded data. They do this by trying a
10235 trial decode on each line until they find one that works. When they
10236 do a flag is set and it starts again knowing it can pass all the
10237 data directly through the decoder. Unfortunately it doesn't reset
10238 the context it uses. This means that if EOF is reached an attempt
10239 is made to pass two EOFs through the context and this causes the
10240 resulting error. This can also cause other problems as well. As is
10241 usual with these problems it takes *ages* to find and the fix is
10242 trivial: move one line.
10243 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10245 *) Ugly workaround to get s_client and s_server working under Windows. The
10246 old code wouldn't work because it needed to select() on sockets and the
10247 tty (for keypresses and to see if data could be written). Win32 only
10248 supports select() on sockets so we select() with a 1s timeout on the
10249 sockets and then see if any characters are waiting to be read, if none
10250 are present then we retry, we also assume we can always write data to
10251 the tty. This isn't nice because the code then blocks until we've
10252 received a complete line of data and it is effectively polling the
10253 keyboard at 1s intervals: however it's quite a bit better than not
10254 working at all :-) A dedicated Windows application might handle this
10255 with an event loop for example.
10258 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10259 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10260 will be called when RSA_sign() and RSA_verify() are used. This is useful
10261 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10262 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10263 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10264 This necessitated the support of an extra signature type NID_md5_sha1
10265 for SSL signatures and modifications to the SSL library to use it instead
10266 of calling RSA_public_decrypt() and RSA_private_encrypt().
10269 *) Add new -verify -CAfile and -CApath options to the crl program, these
10270 will lookup a CRL issuers certificate and verify the signature in a
10271 similar way to the verify program. Tidy up the crl program so it
10272 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10273 less strict. It will now permit CRL extensions even if it is not
10274 a V2 CRL: this will allow it to tolerate some broken CRLs.
10277 *) Initialize all non-automatic variables each time one of the openssl
10278 sub-programs is started (this is necessary as they may be started
10279 multiple times from the "OpenSSL>" prompt).
10280 [Lennart Bang, Bodo Moeller]
10282 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10283 removing all other RSA functionality (this is what NO_RSA does). This
10284 is so (for example) those in the US can disable those operations covered
10285 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10289 *) Non-copying interface to BIO pairs.
10290 (still largely untested)
10293 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10294 ASCII string. This was handled independently in various places before.
10297 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10298 UTF8 strings a character at a time.
10301 *) Use client_version from client hello to select the protocol
10302 (s23_srvr.c) and for RSA client key exchange verification
10303 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10306 *) Add various utility functions to handle SPKACs, these were previously
10307 handled by poking round in the structure internals. Added new function
10308 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10309 print, verify and generate SPKACs. Based on an original idea from
10310 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10313 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10316 *) Allow the config file extension section to be overwritten on the
10317 command line. Based on an original idea from Massimiliano Pala
10318 <madwolf@comune.modena.it>. The new option is called -extensions
10319 and can be applied to ca, req and x509. Also -reqexts to override
10320 the request extensions in req and -crlexts to override the crl extensions
10324 *) Add new feature to the SPKAC handling in ca. Now you can include
10325 the same field multiple times by preceding it by "XXXX." for example:
10328 this is the same syntax as used in the req config file.
10331 *) Allow certificate extensions to be added to certificate requests. These
10332 are specified in a 'req_extensions' option of the req section of the
10333 config file. They can be printed out with the -text option to req but
10334 are otherwise ignored at present.
10337 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10338 data read consists of only the final block it would not decrypted because
10339 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10340 A misplaced 'break' also meant the decrypted final block might not be
10341 copied until the next read.
10344 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10345 a few extra parameters to the DH structure: these will be useful if
10346 for example we want the value of 'q' or implement X9.42 DH.
10349 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10350 provides hooks that allow the default DSA functions or functions on a
10351 "per key" basis to be replaced. This allows hardware acceleration and
10352 hardware key storage to be handled without major modification to the
10353 library. Also added low level modexp hooks and CRYPTO_EX structure and
10354 associated functions.
10357 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10358 as "read only": it can't be written to and the buffer it points to will
10359 not be freed. Reading from a read only BIO is much more efficient than
10360 a normal memory BIO. This was added because there are several times when
10361 an area of memory needs to be read from a BIO. The previous method was
10362 to create a memory BIO and write the data to it, this results in two
10363 copies of the data and an O(n^2) reading algorithm. There is a new
10364 function BIO_new_mem_buf() which creates a read only memory BIO from
10365 an area of memory. Also modified the PKCS#7 routines to use read only
10369 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10370 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10371 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10372 but a retry condition occurred while trying to read the rest.
10375 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10376 NID_pkcs7_encrypted by default: this was wrong since this should almost
10377 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10378 the encrypted data type: this is a more sensible place to put it and it
10379 allows the PKCS#12 code to be tidied up that duplicated this
10383 *) Changed obj_dat.pl script so it takes its input and output files on
10384 the command line. This should avoid shell escape redirection problems
10388 *) Initial support for certificate extension requests, these are included
10389 in things like Xenroll certificate requests. Included functions to allow
10390 extensions to be obtained and added.
10393 *) -crlf option to s_client and s_server for sending newlines as
10394 CRLF (as required by many protocols).
10397 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10399 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10400 [Ralf S. Engelschall]
10402 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10403 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10405 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10409 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10410 DH parameters/keys (q is lost during that conversion, but the resulting
10411 DH parameters contain its length).
10413 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10414 much faster than DH_generate_parameters (which creates parameters
10415 where p = 2*q + 1), and also the smaller q makes DH computations
10416 much more efficient (160-bit exponentiation instead of 1024-bit
10417 exponentiation); so this provides a convenient way to support DHE
10418 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10419 utter importance to use
10420 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10422 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10423 when such DH parameters are used, because otherwise small subgroup
10424 attacks may become possible!
10427 *) Avoid memory leak in i2d_DHparams.
10430 *) Allow the -k option to be used more than once in the enc program:
10431 this allows the same encrypted message to be read by multiple recipients.
10434 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10435 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10436 it will always use the numerical form of the OID, even if it has a short
10440 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10441 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10442 otherwise bn_mod_exp was called. In the case of hardware keys for example
10443 no private key components need be present and it might store extra data
10444 in the RSA structure, which cannot be accessed from bn_mod_exp.
10445 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10446 private key operations.
10449 *) Added support for SPARC Linux.
10452 *) pem_password_cb function type incompatibly changed from
10453 typedef int pem_password_cb(char *buf, int size, int rwflag);
10455 ....(char *buf, int size, int rwflag, void *userdata);
10456 so that applications can pass data to their callbacks:
10457 The PEM[_ASN1]_{read,write}... functions and macros now take an
10458 additional void * argument, which is just handed through whenever
10459 the password callback is called.
10460 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10462 New function SSL_CTX_set_default_passwd_cb_userdata.
10464 Compatibility note: As many C implementations push function arguments
10465 onto the stack in reverse order, the new library version is likely to
10466 interoperate with programs that have been compiled with the old
10467 pem_password_cb definition (PEM_whatever takes some data that
10468 happens to be on the stack as its last argument, and the callback
10469 just ignores this garbage); but there is no guarantee whatsoever that
10472 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10473 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10474 problems not only on Windows, but also on some Unix platforms.
10475 To avoid problematic command lines, these definitions are now in an
10476 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10477 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10480 *) MIPS III/IV assembler module is reimplemented.
10483 *) More DES library cleanups: remove references to srand/rand and
10484 delete an unused file.
10487 *) Add support for the the free Netwide assembler (NASM) under Win32,
10488 since not many people have MASM (ml) and it can be hard to obtain.
10489 This is currently experimental but it seems to work OK and pass all
10490 the tests. Check out INSTALL.W32 for info.
10493 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10494 without temporary keys kept an extra copy of the server key,
10495 and connections with temporary keys did not free everything in case
10499 *) New function RSA_check_key and new openssl rsa option -check
10500 for verifying the consistency of RSA keys.
10501 [Ulf Moeller, Bodo Moeller]
10503 *) Various changes to make Win32 compile work:
10504 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10505 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10506 comparison" warnings.
10507 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10510 *) Add a debugging option to PKCS#5 v2 key generation function: when
10511 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10512 derived keys are printed to stderr.
10515 *) Copy the flags in ASN1_STRING_dup().
10516 [Roman E. Pavlov <pre@mo.msk.ru>]
10518 *) The x509 application mishandled signing requests containing DSA
10519 keys when the signing key was also DSA and the parameters didn't match.
10521 It was supposed to omit the parameters when they matched the signing key:
10522 the verifying software was then supposed to automatically use the CA's
10523 parameters if they were absent from the end user certificate.
10525 Omitting parameters is no longer recommended. The test was also
10526 the wrong way round! This was probably due to unusual behaviour in
10527 EVP_cmp_parameters() which returns 1 if the parameters match.
10528 This meant that parameters were omitted when they *didn't* match and
10529 the certificate was useless. Certificates signed with 'ca' didn't have
10531 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10533 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10534 The interface is as follows:
10535 Applications can use
10536 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10537 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10538 "off" is now the default.
10539 The library internally uses
10540 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10541 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10542 to disable memory-checking temporarily.
10544 Some inconsistent states that previously were possible (and were
10545 even the default) are now avoided.
10547 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10548 with each memory chunk allocated; this is occasionally more helpful
10549 than just having a counter.
10551 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10553 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10557 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10558 which largely parallels "options", but is for changing API behaviour,
10559 whereas "options" are about protocol behaviour.
10560 Initial "mode" flags are:
10562 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10563 a single record has been written.
10564 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10565 retries use the same buffer location.
10566 (But all of the contents must be
10570 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10573 *) Fix problems with no-hmac etc.
10574 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10576 *) New functions RSA_get_default_method(), RSA_set_method() and
10577 RSA_get_method(). These allows replacement of RSA_METHODs without having
10578 to mess around with the internals of an RSA structure.
10581 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10582 Also really enable memory leak checks in openssl.c and in some
10584 [Chad C. Mulligan, Bodo Moeller]
10586 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10587 up the length of negative integers. This has now been simplified to just
10588 store the length when it is first determined and use it later, rather
10589 than trying to keep track of where data is copied and updating it to
10591 [Steve Henson, reported by Brien Wheeler
10592 <bwheeler@authentica-security.com>]
10594 *) Add a new function PKCS7_signatureVerify. This allows the verification
10595 of a PKCS#7 signature but with the signing certificate passed to the
10596 function itself. This contrasts with PKCS7_dataVerify which assumes the
10597 certificate is present in the PKCS#7 structure. This isn't always the
10598 case: certificates can be omitted from a PKCS#7 structure and be
10599 distributed by "out of band" means (such as a certificate database).
10602 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10603 function prototypes in pem.h, also change util/mkdef.pl to add the
10604 necessary function names.
10607 *) mk1mf.pl (used by Windows builds) did not properly read the
10608 options set by Configure in the top level Makefile, and Configure
10609 was not even able to write more than one option correctly.
10610 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10613 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10614 file to be loaded from a BIO or FILE pointer. The BIO version will
10615 for example allow memory BIOs to contain config info.
10618 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10619 Whoever hopes to achieve shared-library compatibility across versions
10620 must use this, not the compile-time macro.
10621 (Exercise 0.9.4: Which is the minimum library version required by
10623 Note: All this applies only to multi-threaded programs, others don't
10627 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10628 through a BIO pair triggered the default case, i.e.
10629 SSLerr(...,SSL_R_UNKNOWN_STATE).
10632 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10633 can use the SSL library even if none of the specific BIOs is
10637 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10638 for the encoded length.
10639 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10641 *) Add initial documentation of the X509V3 functions.
10644 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10645 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10646 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10647 secure PKCS#8 private key format with a high iteration count.
10650 *) Fix determination of Perl interpreter: A perl or perl5
10651 _directory_ in $PATH was also accepted as the interpreter.
10652 [Ralf S. Engelschall]
10654 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10655 wrong with it but it was very old and did things like calling
10656 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10657 unusual formatting.
10660 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10661 to use the new extension code.
10664 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10665 with macros. This should make it easier to change their form, add extra
10666 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10670 *) Add to configuration table a new entry that can specify an alternative
10671 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10672 according to Mark Crispin <MRC@Panda.COM>.
10676 *) DES CBC did not update the IV. Weird.
10679 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10680 Changing the behaviour of the former might break existing programs --
10681 where IV updating is needed, des_ncbc_encrypt can be used.
10684 *) When bntest is run from "make test" it drives bc to check its
10685 calculations, as well as internally checking them. If an internal check
10686 fails, it needs to cause bc to give a non-zero result or make test carries
10687 on without noticing the failure. Fixed.
10690 *) DES library cleanups.
10693 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10694 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10695 ciphers. NOTE: although the key derivation function has been verified
10696 against some published test vectors it has not been extensively tested
10697 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10701 *) Instead of "mkdir -p", which is not fully portable, use new
10702 Perl script "util/mkdir-p.pl".
10705 *) Rewrite the way password based encryption (PBE) is handled. It used to
10706 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10707 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10708 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10709 the 'parameter' field of the AlgorithmIdentifier is passed to the
10710 underlying key generation function so it must do its own ASN1 parsing.
10711 This has also changed the EVP_PBE_CipherInit() function which now has a
10712 'parameter' argument instead of literal salt and iteration count values
10713 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10716 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10717 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10718 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10719 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10720 value was just used as a "magic string" and not used directly its
10721 value doesn't matter.
10724 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10728 *) "linux-sparc64" configuration (ultrapenguin).
10729 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10730 "linux-sparc" configuration.
10731 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10733 *) config now generates no-xxx options for missing ciphers.
10736 *) Support the EBCDIC character set (work in progress).
10737 File ebcdic.c not yet included because it has a different license.
10738 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10740 *) Support BS2000/OSD-POSIX.
10741 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10743 *) Make callbacks for key generation use void * instead of char *.
10746 *) Make S/MIME samples compile (not yet tested).
10749 *) Additional typesafe stacks.
10752 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10756 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10758 *) New configuration variant "sco5-gcc".
10760 *) Updated some demos.
10761 [Sean O Riordain, Wade Scholine]
10763 *) Add missing BIO_free at exit of pkcs12 application.
10766 *) Fix memory leak in conf.c.
10769 *) Updates for Win32 to assembler version of MD5.
10772 *) Set #! path to perl in apps/der_chop to where we found it
10773 instead of using a fixed path.
10776 *) SHA library changes for irix64-mips4-cc.
10779 *) Improvements for VMS support.
10783 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10785 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10786 This also avoids the problems with SC4.2 and unpatched SC5.
10787 [Andy Polyakov <appro@fy.chalmers.se>]
10789 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10790 These are required because of the typesafe stack would otherwise break
10791 existing code. If old code used a structure member which used to be STACK
10792 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10793 sk_num or sk_value it would produce an error because the num, data members
10794 are not present in STACK_OF. Now it just produces a warning. sk_set
10795 replaces the old method of assigning a value to sk_value
10796 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10797 that does this will no longer work (and should use sk_set instead) but
10798 this could be regarded as a "questionable" behaviour anyway.
10801 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10802 correctly handle encrypted S/MIME data.
10805 *) Change type of various DES function arguments from des_cblock
10806 (which means, in function argument declarations, pointer to char)
10807 to des_cblock * (meaning pointer to array with 8 char elements),
10808 which allows the compiler to do more typechecking; it was like
10809 that back in SSLeay, but with lots of ugly casts.
10811 Introduce new type const_des_cblock.
10814 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10815 problems: find RecipientInfo structure that matches recipient certificate
10816 and initialise the ASN1 structures properly based on passed cipher.
10819 *) Belatedly make the BN tests actually check the results.
10822 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10823 to and from BNs: it was completely broken. New compilation option
10824 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10825 key elements as negative integers.
10828 *) Reorganize and speed up MD5.
10829 [Andy Polyakov <appro@fy.chalmers.se>]
10832 [Richard Levitte <richard@levitte.org>]
10834 *) New option -out to asn1parse to allow the parsed structure to be
10835 output to a file. This is most useful when combined with the -strparse
10836 option to examine the output of things like OCTET STRINGS.
10839 *) Make SSL library a little more fool-proof by not requiring any longer
10840 that SSL_set_{accept,connect}_state be called before
10841 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10842 in many applications because usually everything *appeared* to work as
10843 intended anyway -- now it really works as intended).
10846 *) Move openssl.cnf out of lib/.
10849 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10850 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10851 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10852 [Ralf S. Engelschall]
10854 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10855 handle PKCS#7 enveloped data properly.
10856 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10858 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10859 copying pointers. The cert_st handling is changed by this in
10860 various ways (and thus what used to be known as ctx->default_cert
10861 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10862 any longer when s->cert does not give us what we need).
10863 ssl_cert_instantiate becomes obsolete by this change.
10864 As soon as we've got the new code right (possibly it already is?),
10865 we have solved a couple of bugs of the earlier code where s->cert
10866 was used as if it could not have been shared with other SSL structures.
10868 Note that using the SSL API in certain dirty ways now will result
10869 in different behaviour than observed with earlier library versions:
10870 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10871 does not influence s as it used to.
10873 In order to clean up things more thoroughly, inside SSL_SESSION
10874 we don't use CERT any longer, but a new structure SESS_CERT
10875 that holds per-session data (if available); currently, this is
10876 the peer's certificate chain and, for clients, the server's certificate
10877 and temporary key. CERT holds only those values that can have
10878 meaningful defaults in an SSL_CTX.
10881 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10882 from the internal representation. Various PKCS#7 fixes: remove some
10883 evil casts and set the enc_dig_alg field properly based on the signing
10887 *) Allow PKCS#12 password to be set from the command line or the
10888 environment. Let 'ca' get its config file name from the environment
10889 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10893 *) Allow certificate policies extension to use an IA5STRING for the
10894 organization field. This is contrary to the PKIX definition but
10895 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10899 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10900 without disallowing inline assembler and the like for non-pedantic builds.
10903 *) Support Borland C++ builder.
10904 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10906 *) Support Mingw32.
10909 *) SHA-1 cleanups and performance enhancements.
10910 [Andy Polyakov <appro@fy.chalmers.se>]
10912 *) Sparc v8plus assembler for the bignum library.
10913 [Andy Polyakov <appro@fy.chalmers.se>]
10915 *) Accept any -xxx and +xxx compiler options in Configure.
10918 *) Update HPUX configuration.
10921 *) Add missing sk_<type>_unshift() function to safestack.h
10922 [Ralf S. Engelschall]
10924 *) New function SSL_CTX_use_certificate_chain_file that sets the
10925 "extra_cert"s in addition to the certificate. (This makes sense
10926 only for "PEM" format files, as chains as a whole are not
10930 *) Support verify_depth from the SSL API.
10931 x509_vfy.c had what can be considered an off-by-one-error:
10932 Its depth (which was not part of the external interface)
10933 was actually counting the number of certificates in a chain;
10934 now it really counts the depth.
10937 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10938 instead of X509err, which often resulted in confusing error
10939 messages since the error codes are not globally unique
10940 (e.g. an alleged error in ssl3_accept when a certificate
10941 didn't match the private key).
10943 *) New function SSL_CTX_set_session_id_context that allows to set a default
10944 value (so that you don't need SSL_set_session_id_context for each
10945 connection using the SSL_CTX).
10948 *) OAEP decoding bug fix.
10951 *) Support INSTALL_PREFIX for package builders, as proposed by
10955 *) New Configure options "threads" and "no-threads". For systems
10956 where the proper compiler options are known (currently Solaris
10957 and Linux), "threads" is the default.
10960 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10963 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10964 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10965 such as /usr/local/bin.
10968 *) "make linux-shared" to build shared libraries.
10969 [Niels Poppe <niels@netbox.org>]
10971 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10974 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10975 extension adding in x509 utility.
10978 *) Remove NOPROTO sections and error code comments.
10981 *) Partial rewrite of the DEF file generator to now parse the ANSI
10985 *) New Configure options --prefix=DIR and --openssldir=DIR.
10988 *) Complete rewrite of the error code script(s). It is all now handled
10989 by one script at the top level which handles error code gathering,
10990 header rewriting and C source file generation. It should be much better
10991 than the old method: it now uses a modified version of Ulf's parser to
10992 read the ANSI prototypes in all header files (thus the old K&R definitions
10993 aren't needed for error creation any more) and do a better job of
10994 translating function codes into names. The old 'ASN1 error code imbedded
10995 in a comment' is no longer necessary and it doesn't use .err files which
10996 have now been deleted. Also the error code call doesn't have to appear all
10997 on one line (which resulted in some large lines...).
11000 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11003 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11004 0 (which usually indicates a closed connection), but continue reading.
11007 *) Fix some race conditions.
11010 *) Add support for CRL distribution points extension. Add Certificate
11011 Policies and CRL distribution points documentation.
11014 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11017 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11018 8 of keying material. Merlin has also confirmed interop with this fix
11019 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11020 [Merlin Hughes <merlin@baltimore.ie>]
11022 *) Fix lots of warnings.
11023 [Richard Levitte <levitte@stacken.kth.se>]
11025 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11026 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11027 [Richard Levitte <levitte@stacken.kth.se>]
11029 *) Fix problems with sizeof(long) == 8.
11030 [Andy Polyakov <appro@fy.chalmers.se>]
11032 *) Change functions to ANSI C.
11035 *) Fix typos in error codes.
11036 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11038 *) Remove defunct assembler files from Configure.
11041 *) SPARC v8 assembler BIGNUM implementation.
11042 [Andy Polyakov <appro@fy.chalmers.se>]
11044 *) Support for Certificate Policies extension: both print and set.
11045 Various additions to support the r2i method this uses.
11048 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11049 return a const string when you are expecting an allocated buffer.
11052 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11053 types DirectoryString and DisplayText.
11056 *) Add code to allow r2i extensions to access the configuration database,
11057 add an LHASH database driver and add several ctx helper functions.
11060 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11061 fail when they extended the size of a BIGNUM.
11064 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11065 support typesafe stack.
11068 *) Fix typo in SSL_[gs]et_options().
11069 [Nils Frostberg <nils@medcom.se>]
11071 *) Delete various functions and files that belonged to the (now obsolete)
11072 old X509V3 handling code.
11075 *) New Configure option "rsaref".
11078 *) Don't auto-generate pem.h.
11081 *) Introduce type-safe ASN.1 SETs.
11084 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11085 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11087 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11088 that links with OpenSSL (well at least cause lots of warnings), but fear
11089 not: the conversion is trivial, and it eliminates loads of evil casts. A
11090 few STACKed things have been converted already. Feel free to convert more.
11091 In the fullness of time, I'll do away with the STACK type altogether.
11094 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11095 specified in <certfile> by updating the entry in the index.txt file.
11096 This way one no longer has to edit the index.txt file manually for
11097 revoking a certificate. The -revoke option does the gory details now.
11098 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11100 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11101 `-text' option at all and this way the `-noout -text' combination was
11102 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11103 [Ralf S. Engelschall]
11105 *) Make sure a corresponding plain text error message exists for the
11106 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11107 verify callback function determined that a certificate was revoked.
11108 [Ralf S. Engelschall]
11110 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11111 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11112 all available cipers including rc5, which was forgotten until now.
11113 In order to let the testing shell script know which algorithms
11114 are available, a new (up to now undocumented) command
11115 "openssl list-cipher-commands" is used.
11118 *) Bugfix: s_client occasionally would sleep in select() when
11119 it should have checked SSL_pending() first.
11122 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11123 the raw DSA values prior to ASN.1 encoding.
11126 *) Tweaks to Configure
11127 [Niels Poppe <niels@netbox.org>]
11129 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11133 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11136 *) New config option to avoid instructions that are illegal on the 80386.
11137 The default code is faster, but requires at least a 486.
11140 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11141 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11142 same as SSL2_VERSION anyway.
11145 *) New "-showcerts" option for s_client.
11148 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11149 application. Various cleanups and fixes.
11152 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11153 modify error routines to work internally. Add error codes and PBE init
11154 to library startup routines.
11157 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11158 packing functions to asn1 and evp. Changed function names and error
11159 codes along the way.
11162 *) PKCS12 integration: and so it begins... First of several patches to
11163 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11164 objects to objects.h
11167 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11168 and display support for Thawte strong extranet extension.
11171 *) Add LinuxPPC support.
11172 [Jeff Dubrule <igor@pobox.org>]
11174 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11175 bn_div_words in alpha.s.
11176 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11178 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11179 OAEP isn't supported when OpenSSL is built with RSAref.
11180 [Ulf Moeller <ulf@fitug.de>]
11182 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11183 so they no longer are missing under -DNOPROTO.
11184 [Soren S. Jorvang <soren@t.dk>]
11187 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11189 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11190 doesn't work when the session is reused. Coming soon!
11193 *) Fix a security hole, that allows sessions to be reused in the wrong
11194 context thus bypassing client cert protection! All software that uses
11195 client certs and session caches in multiple contexts NEEDS PATCHING to
11196 allow session reuse! A fuller solution is in the works.
11197 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11199 *) Some more source tree cleanups (removed obsolete files
11200 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11201 permission on "config" script to be executable) and a fix for the INSTALL
11203 [Ulf Moeller <ulf@fitug.de>]
11205 *) Remove some legacy and erroneous uses of malloc, free instead of
11207 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11209 *) Make rsa_oaep_test return non-zero on error.
11210 [Ulf Moeller <ulf@fitug.de>]
11212 *) Add support for native Solaris shared libraries. Configure
11213 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11214 if someone would make that last step automatic.
11215 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11217 *) ctx_size was not built with the right compiler during "make links". Fixed.
11220 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11221 except NULL ciphers". This means the default cipher list will no longer
11222 enable NULL ciphers. They need to be specifically enabled e.g. with
11223 the string "DEFAULT:eNULL".
11226 *) Fix to RSA private encryption routines: if p < q then it would
11227 occasionally produce an invalid result. This will only happen with
11228 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11231 *) Be less restrictive and allow also `perl util/perlpath.pl
11232 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11233 because this way one can also use an interpreter named `perl5' (which is
11234 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11235 installed as `perl').
11236 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11238 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11239 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11241 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11242 advapi32.lib to Win32 build and change the pem test comparison
11243 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11244 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11245 and crypto/des/ede_cbcm_enc.c.
11248 *) DES quad checksum was broken on big-endian architectures. Fixed.
11251 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11252 Win32 test batch file so it (might) work again. The Win32 test batch file
11253 is horrible: I feel ill....
11256 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11257 in e_os.h. Audit of header files to check ANSI and non ANSI
11258 sections: 10 functions were absent from non ANSI section and not exported
11259 from Windows DLLs. Fixed up libeay.num for new functions.
11262 *) Make `openssl version' output lines consistent.
11263 [Ralf S. Engelschall]
11265 *) Fix Win32 symbol export lists for BIO functions: Added
11266 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11267 to ms/libeay{16,32}.def.
11268 [Ralf S. Engelschall]
11270 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11271 fine under Unix and passes some trivial tests I've now added. But the
11272 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11273 added to make sure no one expects that this stuff really works in the
11274 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11275 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11277 [Ralf S. Engelschall]
11279 *) Fix the generation of two part addresses in perl.
11280 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11282 *) Add config entry for Linux on MIPS.
11283 [John Tobey <jtobey@channel1.com>]
11285 *) Make links whenever Configure is run, unless we are on Windoze.
11288 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11289 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11293 *) Add a useful kludge to allow package maintainers to specify compiler and
11294 other platforms details on the command line without having to patch the
11295 Configure script everytime: One now can use ``perl Configure
11296 <id>:<details>'', i.e. platform ids are allowed to have details appended
11297 to them (separated by colons). This is treated as there would be a static
11298 pre-configured entry in Configure's %table under key <id> with value
11299 <details> and ``perl Configure <id>'' is called. So, when you want to
11300 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11301 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11302 now, which overrides the FreeBSD-elf entry on-the-fly.
11303 [Ralf S. Engelschall]
11305 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11308 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11309 on the `perl Configure ...' command line. This way one can compile
11310 OpenSSL libraries with Position Independent Code (PIC) which is needed
11311 for linking it into DSOs.
11312 [Ralf S. Engelschall]
11314 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11318 *) Cleaned up the LICENSE document: The official contact for any license
11319 questions now is the OpenSSL core team under openssl-core@openssl.org.
11320 And add a paragraph about the dual-license situation to make sure people
11321 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11322 to the OpenSSL toolkit.
11323 [Ralf S. Engelschall]
11325 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11326 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11327 Additionally cleaned up the `make links' target: Remove unnecessary
11328 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11329 to speed processing and no longer clutter the display with confusing
11330 stuff. Instead only the actually done links are displayed.
11331 [Ralf S. Engelschall]
11333 *) Permit null encryption ciphersuites, used for authentication only. It used
11334 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11335 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11339 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11340 signed attributes when verifying signatures (this would break them),
11341 the detached data encoding was wrong and public keys obtained using
11342 X509_get_pubkey() weren't freed.
11345 *) Add text documentation for the BUFFER functions. Also added a work around
11346 to a Win95 console bug. This was triggered by the password read stuff: the
11347 last character typed gets carried over to the next fread(). If you were
11348 generating a new cert request using 'req' for example then the last
11349 character of the passphrase would be CR which would then enter the first
11353 *) Added the new `Includes OpenSSL Cryptography Software' button as
11354 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11355 button and can be used by applications based on OpenSSL to show the
11356 relationship to the OpenSSL project.
11357 [Ralf S. Engelschall]
11359 *) Remove confusing variables in function signatures in files
11360 ssl/ssl_lib.c and ssl/ssl.h.
11361 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11363 *) Don't install bss_file.c under PREFIX/include/
11364 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11366 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11367 functions that return function pointers and has support for NT specific
11368 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11369 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11370 unsigned to signed types: this was killing the Win32 compile.
11373 *) Add new certificate file to stack functions,
11374 SSL_add_dir_cert_subjects_to_stack() and
11375 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11376 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11377 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11378 This means that Apache-SSL and similar packages don't have to mess around
11379 to add as many CAs as they want to the preferred list.
11382 *) Experiment with doxygen documentation. Currently only partially applied to
11384 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11385 openssl.doxy as the configuration file.
11388 *) Get rid of remaining C++-style comments which strict C compilers hate.
11389 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11391 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11392 compiled in by default: it has problems with large keys.
11395 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11396 DH private keys and/or callback functions which directly correspond to
11397 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11398 is needed for applications which have to configure certificates on a
11399 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11401 For the RSA certificate situation is makes no difference, but
11402 for the DSA certificate situation this fixes the "no shared cipher"
11403 problem where the OpenSSL cipher selection procedure failed because the
11404 temporary keys were not overtaken from the context and the API provided
11405 no way to reconfigure them.
11406 The new functions now let applications reconfigure the stuff and they
11407 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11408 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11409 non-public-API function ssl_cert_instantiate() is used as a helper
11410 function and also to reduce code redundancy inside ssl_rsa.c.
11411 [Ralf S. Engelschall]
11413 *) Move s_server -dcert and -dkey options out of the undocumented feature
11414 area because they are useful for the DSA situation and should be
11415 recognized by the users.
11416 [Ralf S. Engelschall]
11418 *) Fix the cipher decision scheme for export ciphers: the export bits are
11419 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11420 SSL_EXP_MASK. So, the original variable has to be used instead of the
11421 already masked variable.
11422 [Richard Levitte <levitte@stacken.kth.se>]
11424 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11425 [Richard Levitte <levitte@stacken.kth.se>]
11427 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11428 from `int' to `unsigned int' because it's a length and initialized by
11429 EVP_DigestFinal() which expects an `unsigned int *'.
11430 [Richard Levitte <levitte@stacken.kth.se>]
11432 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11433 script. Instead use the usual Shell->Perl transition trick.
11434 [Ralf S. Engelschall]
11436 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11437 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11438 -noout -modulus' as it's already the case for `openssl rsa -noout
11439 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11440 currently the public key is printed (a decision which was already done by
11441 `openssl dsa -modulus' in the past) which serves a similar purpose.
11442 Additionally the NO_RSA no longer completely removes the whole -modulus
11443 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11445 [Ralf S. Engelschall]
11447 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11448 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11449 [Arne Ansper <arne@ats.cyber.ee>]
11451 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11452 to be added. Now both 'req' and 'ca' can use new objects defined in the
11456 *) Add cool BIO that does syslog (or event log on NT).
11457 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11459 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11460 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11461 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11462 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11465 *) Add preliminary config info for new extension code.
11468 *) Make RSA_NO_PADDING really use no padding.
11469 [Ulf Moeller <ulf@fitug.de>]
11471 *) Generate errors when private/public key check is done.
11474 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11475 for some CRL extensions and new objects added.
11478 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11479 key usage extension and fuller support for authority key id.
11482 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11483 padding method for RSA, which is recommended for new applications in PKCS
11484 #1 v2.0 (RFC 2437, October 1998).
11485 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11486 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11487 against Bleichbacher's attack on RSA.
11488 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11491 *) Updates to the new SSL compression code
11492 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11494 *) Fix so that the version number in the master secret, when passed
11495 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11496 (because the server will not accept higher), that the version number
11497 is 0x03,0x01, not 0x03,0x00
11498 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11500 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11501 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11502 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11505 *) Support for RAW extensions where an arbitrary extension can be
11506 created by including its DER encoding. See apps/openssl.cnf for
11510 *) Make sure latest Perl versions don't interpret some generated C array
11511 code as Perl array code in the crypto/err/err_genc.pl script.
11512 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11514 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11515 not many people have the assembler. Various Win32 compilation fixes and
11516 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11517 build instructions.
11520 *) Modify configure script 'Configure' to automatically create crypto/date.h
11521 file under Win32 and also build pem.h from pem.org. New script
11522 util/mkfiles.pl to create the MINFO file on environments that can't do a
11523 'make files': perl util/mkfiles.pl >MINFO should work.
11526 *) Major rework of DES function declarations, in the pursuit of correctness
11527 and purity. As a result, many evil casts evaporated, and some weirdness,
11528 too. You may find this causes warnings in your code. Zapping your evil
11529 casts will probably fix them. Mostly.
11532 *) Fix for a typo in asn1.h. Bug fix to object creation script
11533 obj_dat.pl. It considered a zero in an object definition to mean
11534 "end of object": none of the objects in objects.h have any zeros
11535 so it wasn't spotted.
11536 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11538 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11539 Masking (CBCM). In the absence of test vectors, the best I have been able
11540 to do is check that the decrypt undoes the encrypt, so far. Send me test
11541 vectors if you have them.
11544 *) Correct calculation of key length for export ciphers (too much space was
11545 allocated for null ciphers). This has not been tested!
11548 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11549 message is now correct (it understands "crypto" and "ssl" on its
11550 command line). There is also now an "update" option. This will update
11551 the util/ssleay.num and util/libeay.num files with any new functions.
11553 perl util/mkdef.pl crypto ssl update
11554 it will update them.
11557 *) Overhauled the Perl interface (perl/*):
11558 - ported BN stuff to OpenSSL's different BN library
11559 - made the perl/ source tree CVS-aware
11560 - renamed the package from SSLeay to OpenSSL (the files still contain
11561 their history because I've copied them in the repository)
11562 - removed obsolete files (the test scripts will be replaced
11563 by better Test::Harness variants in the future)
11564 [Ralf S. Engelschall]
11566 *) First cut for a very conservative source tree cleanup:
11567 1. merge various obsolete readme texts into doc/ssleay.txt
11568 where we collect the old documents and readme texts.
11569 2. remove the first part of files where I'm already sure that we no
11570 longer need them because of three reasons: either they are just temporary
11571 files which were left by Eric or they are preserved original files where
11572 I've verified that the diff is also available in the CVS via "cvs diff
11573 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11574 the crypto/md/ stuff).
11575 [Ralf S. Engelschall]
11577 *) More extension code. Incomplete support for subject and issuer alt
11578 name, issuer and authority key id. Change the i2v function parameters
11579 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11580 what that's for :-) Fix to ASN1 macro which messed up
11581 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11584 *) Preliminary support for ENUMERATED type. This is largely copied from the
11588 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11589 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11591 *) Make sure `make rehash' target really finds the `openssl' program.
11592 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11594 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11595 like to hear about it if this slows down other processors.
11598 *) Add CygWin32 platform information to Configure script.
11599 [Alan Batie <batie@aahz.jf.intel.com>]
11601 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11602 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11604 *) New program nseq to manipulate netscape certificate sequences
11607 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11611 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11612 but the BN code had some problems that would cause failures when
11613 doing certificate verification and some other functions.
11614 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11616 *) Add ASN1 and PEM code to support netscape certificate sequences.
11619 *) Add ASN1 and PEM code to support netscape certificate sequences.
11622 *) Add several PKIX and private extended key usage OIDs.
11625 *) Modify the 'ca' program to handle the new extension code. Modify
11626 openssl.cnf for new extension format, add comments.
11629 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11630 and add a sample to openssl.cnf so req -x509 now adds appropriate
11634 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11635 error code, add initial support to X509_print() and x509 application.
11638 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11639 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11640 stuff is currently isolated and isn't even compiled yet.
11643 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11644 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11645 Removed the versions check from X509 routines when loading extensions:
11646 this allows certain broken certificates that don't set the version
11647 properly to be processed.
11650 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11651 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11652 can still be regenerated with "make depend".
11655 *) Spelling mistake in C version of CAST-128.
11656 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11658 *) Changes to the error generation code. The perl script err-code.pl
11659 now reads in the old error codes and retains the old numbers, only
11660 adding new ones if necessary. It also only changes the .err files if new
11661 codes are added. The makefiles have been modified to only insert errors
11662 when needed (to avoid needlessly modifying header files). This is done
11663 by only inserting errors if the .err file is newer than the auto generated
11664 C file. To rebuild all the error codes from scratch (the old behaviour)
11665 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11666 or delete all the .err files.
11669 *) CAST-128 was incorrectly implemented for short keys. The C version has
11670 been fixed, but is untested. The assembler versions are also fixed, but
11671 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11672 to regenerate it if needed.
11673 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11674 Hagino <itojun@kame.net>]
11676 *) File was opened incorrectly in randfile.c.
11677 [Ulf Möller <ulf@fitug.de>]
11679 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11680 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11681 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11682 al: it's just almost always a UTCTime. Note this patch adds new error
11683 codes so do a "make errors" if there are problems.
11686 *) Correct Linux 1 recognition in config.
11687 [Ulf Möller <ulf@fitug.de>]
11689 *) Remove pointless MD5 hash when using DSA keys in ca.
11690 [Anonymous <nobody@replay.com>]
11692 *) Generate an error if given an empty string as a cert directory. Also
11693 generate an error if handed NULL (previously returned 0 to indicate an
11694 error, but didn't set one).
11695 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11697 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11700 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11701 parameters. This was causing a warning which killed off the Win32 compile.
11704 *) Remove C++ style comments from crypto/bn/bn_local.h.
11705 [Neil Costigan <neil.costigan@celocom.com>]
11707 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11708 based on a text string, looking up short and long names and finally
11709 "dot" format. The "dot" format stuff didn't work. Added new function
11710 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11711 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11712 OID is not part of the table.
11715 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11716 X509_LOOKUP_by_alias().
11719 *) Sort openssl functions by name.
11722 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11723 encryption from sample DSA keys (in case anyone is interested the password
11727 *) Make _all_ *_free functions accept a NULL pointer.
11728 [Frans Heymans <fheymans@isaserver.be>]
11730 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11732 [Anonymous <nobody@replay.com>]
11734 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11735 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11737 *) Don't blow it for numeric -newkey arguments to apps/req.
11738 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11740 *) Temp key "for export" tests were wrong in s3_srvr.c.
11741 [Anonymous <nobody@replay.com>]
11743 *) Add prototype for temp key callback functions
11744 SSL_CTX_set_tmp_{rsa,dh}_callback().
11747 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11748 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11751 *) X509_name_add_entry() freed the wrong thing after an error.
11752 [Arne Ansper <arne@ats.cyber.ee>]
11754 *) rsa_eay.c would attempt to free a NULL context.
11755 [Arne Ansper <arne@ats.cyber.ee>]
11757 *) BIO_s_socket() had a broken should_retry() on Windoze.
11758 [Arne Ansper <arne@ats.cyber.ee>]
11760 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11761 [Arne Ansper <arne@ats.cyber.ee>]
11763 *) Make sure the already existing X509_STORE->depth variable is initialized
11764 in X509_STORE_new(), but document the fact that this variable is still
11765 unused in the certificate verification process.
11766 [Ralf S. Engelschall]
11768 *) Fix the various library and apps files to free up pkeys obtained from
11769 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11772 *) Fix reference counting in X509_PUBKEY_get(). This makes
11773 demos/maurice/example2.c work, amongst others, probably.
11774 [Steve Henson and Ben Laurie]
11776 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11777 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11778 are no longer created. This way we have a single and consistent command
11779 line interface `openssl <command>', similar to `cvs <command>'.
11780 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11782 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11783 BIT STRING wrapper always have zero unused bits.
11786 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11789 *) Make the top-level INSTALL documentation easier to understand.
11792 *) Makefiles updated to exit if an error occurs in a sub-directory
11793 make (including if user presses ^C) [Paul Sutton]
11795 *) Make Montgomery context stuff explicit in RSA data structure.
11798 *) Fix build order of pem and err to allow for generated pem.h.
11801 *) Fix renumbering bug in X509_NAME_delete_entry().
11804 *) Enhanced the err-ins.pl script so it makes the error library number
11805 global and can add a library name. This is needed for external ASN1 and
11806 other error libraries.
11809 *) Fixed sk_insert which never worked properly.
11812 *) Fix ASN1 macros so they can handle indefinite length constructed
11813 EXPLICIT tags. Some non standard certificates use these: they can now
11817 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11818 into a single doc/ssleay.txt bundle. This way the information is still
11819 preserved but no longer messes up this directory. Now it's new room for
11820 the new set of documentation files.
11821 [Ralf S. Engelschall]
11823 *) SETs were incorrectly DER encoded. This was a major pain, because they
11824 shared code with SEQUENCEs, which aren't coded the same. This means that
11825 almost everything to do with SETs or SEQUENCEs has either changed name or
11826 number of arguments.
11827 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11829 *) Fix test data to work with the above.
11832 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11833 was already fixed by Eric for 0.9.1 it seems.
11834 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11836 *) Autodetect FreeBSD3.
11839 *) Fix various bugs in Configure. This affects the following platforms:
11843 unixware-2.0-pentium
11847 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11848 before they are needed.
11851 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11855 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11857 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11858 changed SSLeay to OpenSSL in version strings.
11859 [Ralf S. Engelschall]
11861 *) Some fixups to the top-level documents.
11864 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11865 because the symlink to include/ was missing.
11866 [Ralf S. Engelschall]
11868 *) Incorporated the popular no-RSA/DSA-only patches
11869 which allow to compile a RSA-free SSLeay.
11870 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11872 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11873 when "ssleay" is still not found.
11874 [Ralf S. Engelschall]
11876 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11877 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11879 *) Updated the README file.
11880 [Ralf S. Engelschall]
11882 *) Added various .cvsignore files in the CVS repository subdirs
11883 to make a "cvs update" really silent.
11884 [Ralf S. Engelschall]
11886 *) Recompiled the error-definition header files and added
11887 missing symbols to the Win32 linker tables.
11888 [Ralf S. Engelschall]
11890 *) Cleaned up the top-level documents;
11891 o new files: CHANGES and LICENSE
11892 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11893 o merged COPYRIGHT into LICENSE
11894 o removed obsolete TODO file
11895 o renamed MICROSOFT to INSTALL.W32
11896 [Ralf S. Engelschall]
11898 *) Removed dummy files from the 0.9.1b source tree:
11899 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11900 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11901 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11902 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11903 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11904 [Ralf S. Engelschall]
11906 *) Added various platform portability fixes.
11909 *) The Genesis of the OpenSSL rpject:
11910 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11911 Young and Tim J. Hudson created while they were working for C2Net until
11913 [The OpenSSL Project]
11916 Changes between 0.9.0b and 0.9.1b [not released]
11918 *) Updated a few CA certificates under certs/
11921 *) Changed some BIGNUM api stuff.
11924 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11925 DGUX x86, Linux Alpha, etc.
11928 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11929 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11933 *) Add -strparse option to asn1pars program which parses nested
11935 [Dr Stephen Henson <shenson@bigfoot.com>]
11937 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11940 *) DSA fix for "ca" program.
11943 *) Added "-genkey" option to "dsaparam" program.
11946 *) Added RIPE MD160 (rmd160) message digest.
11949 *) Added -a (all) option to "ssleay version" command.
11952 *) Added PLATFORM define which is the id given to Configure.
11955 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11958 *) Extended the ASN.1 parser routines.
11961 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11964 *) Added a BN_CTX to the BN library.
11967 *) Fixed the weak key values in DES library
11970 *) Changed API in EVP library for cipher aliases.
11973 *) Added support for RC2/64bit cipher.
11976 *) Converted the lhash library to the crypto/mem.c functions.
11979 *) Added more recognized ASN.1 object ids.
11982 *) Added more RSA padding checks for SSL/TLS.
11985 *) Added BIO proxy/filter functionality.
11988 *) Added extra_certs to SSL_CTX which can be used
11989 send extra CA certificates to the client in the CA cert chain sending
11990 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11993 *) Now Fortezza is denied in the authentication phase because
11994 this is key exchange mechanism is not supported by SSLeay at all.
11997 *) Additional PKCS1 checks.
12000 *) Support the string "TLSv1" for all TLS v1 ciphers.
12003 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12004 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12007 *) Fixed a few memory leaks.
12010 *) Fixed various code and comment typos.
12013 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12014 bytes sent in the client random.
12015 [Edward Bishop <ebishop@spyglass.com>]