5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Removed the aged BC-32 config and all its supporting scripts
10 *) Removed support for Ultrix as a platform.
13 *) Add support for HKDF.
16 *) Add support for blake2b and blake2s
19 *) Added support for "pipelining". Ciphers that have the
20 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
21 encryptions/decryptions simultaneously. There are currently no built-in
22 ciphers with this property but the expectation is that engines will be able
23 to offer it to significantly improve throughput. Support has been extended
24 into libssl so that multiple records for a single connection can be
25 processed in one go (for >=TLS 1.1).
28 *) Added the AFALG engine. This is an async capable engine which is able to
29 offload work to the Linux kernel. In this initial version it only supports
30 AES128-CBC. The kernel must be version 4.1.0 or greater.
33 *) OpenSSL now uses a new threading API. It is no longer necessary to
34 set locking callbacks to use OpenSSL in a multi-threaded environment. There
35 are two supported threading models: pthreads and windows threads. It is
36 also possible to configure OpenSSL at compile time for "no-threads". The
37 old threading API should no longer be used. The functions have been
38 replaced with "no-op" compatibility macros.
39 [Alessandro Ghedini, Matt Caswell]
41 *) Modify behavior of ALPN to invoke callback after SNI/servername
42 callback, such that updates to the SSL_CTX affect ALPN.
45 *) Add SSL_CIPHER queries for authentication and key-exchange.
48 *) Changes to the DEFAULT cipherlist:
49 - Prefer (EC)DHE handshakes over plain RSA.
50 - Prefer AEAD ciphers over legacy ciphers.
51 - Prefer ECDSA over RSA when both certificates are available.
52 - Prefer TLSv1.2 ciphers/PRF.
53 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
57 *) Change the ECC default curve list to be this, in order: x25519,
58 secp256r1, secp521r1, secp384r1.
61 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
62 disabled by default. They can be re-enabled using the
63 enable-weak-ssl-ciphers option to Configure.
66 *) If the server has ALPN configured, but supports no protocols that the
67 client advertises, send a fatal "no_application_protocol" alert.
68 This behaviour is SHALL in RFC 7301, though it isn't universally
69 implemented by other servers.
72 *) Add X25519 support.
73 Integrate support for X25519 into EC library. This includes support
74 for public and private key encoding using the format documented in
75 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
76 OID from that draft, encodes public keys using little endian
77 format in the ECPoint structure and private keys using
78 little endian form in the privateKey field of the ECPrivateKey
79 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
82 Note: the current version supports key generation, public and
83 private key encoding and ECDH key agreement using the EC API.
84 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
88 *) Deprecate SRP_VBASE_get_by_user.
89 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
90 In order to fix an unavoidable memory leak (CVE-2016-0798),
91 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
92 seed, even if the seed is configured.
94 Users should use SRP_VBASE_get1_by_user instead. Note that in
95 SRP_VBASE_get1_by_user, caller must free the returned value. Note
96 also that even though configuring the SRP seed attempts to hide
97 invalid usernames by continuing the handshake with fake
98 credentials, this behaviour is not constant time and no strong
99 guarantees are made that the handshake is indistinguishable from
100 that of a valid user.
103 *) Configuration change; it's now possible to build dynamic engines
104 without having to build shared libraries and vice versa. This
105 only applies to the engines in engines/, those in crypto/engine/
106 will always be built into libcrypto (i.e. "static").
108 Building dynamic engines is enabled by default; to disable, use
109 the configuration option "disable-dynamic-engine".
111 The only requirements for building dynamic engines are the
112 presence of the DSO module and building with position independent
113 code, so they will also automatically be disabled if configuring
114 with "disable-dso" or "disable-pic".
116 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
117 are also taken away from openssl/opensslconf.h, as they are
121 *) Configuration change; if there is a known flag to compile
122 position independent code, it will always be applied on the
123 libcrypto and libssl object files, and never on the application
124 object files. This means other libraries that use routines from
125 libcrypto / libssl can be made into shared libraries regardless
126 of how OpenSSL was configured.
128 If this isn't desirable, the configuration options "disable-pic"
129 or "no-pic" can be used to disable the use of PIC. This will
130 also disable building shared libraries and dynamic engines.
133 *) Removed JPAKE code. It was experimental and has no wide use.
136 *) The INSTALL_PREFIX Makefile variable has been renamed to
137 DESTDIR. That makes for less confusion on what this variable
138 is for. Also, the configuration option --install_prefix is
142 *) Heartbeat for TLS has been removed and is disabled by default
143 for DTLS; configure with enable-heartbeats. Code that uses the
144 old #define's might need to be updated.
145 [Emilia Käsper, Rich Salz]
147 *) Rename REF_CHECK to REF_DEBUG.
150 *) New "unified" build system
152 The "unified" build system is aimed to be a common system for all
153 platforms we support. With it comes new support for VMS.
155 This system builds supports building in a different directory tree
156 than the source tree. It produces one Makefile (for unix family
157 or lookalikes), or one descrip.mms (for VMS).
159 The source of information to make the Makefile / descrip.mms is
160 small files called 'build.info', holding the necessary
161 information for each directory with source to compile, and a
162 template in Configurations, like unix-Makefile.tmpl or
165 We rely heavily on the perl module Text::Template.
168 *) Added support for auto-initialisation and de-initialisation of the library.
169 OpenSSL no longer requires explicit init or deinit routines to be called,
170 except in certain circumstances. See the OPENSSL_init_crypto() and
171 OPENSSL_init_ssl() man pages for further information.
174 *) The arguments to the DTLSv1_listen function have changed. Specifically the
175 "peer" argument is now expected to be a BIO_ADDR object.
177 *) Rewrite of BIO networking library. The BIO library lacked consistent
178 support of IPv6, and adding it required some more extensive
179 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
180 which hold all types of addresses and chains of address information.
181 It also introduces a new API, with functions like BIO_socket,
182 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
183 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
184 have been adapted accordingly.
187 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
191 *) CRIME protection: disable compression by default, even if OpenSSL is
192 compiled with zlib enabled. Applications can still enable compression
193 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
194 using the SSL_CONF library to configure compression.
197 *) The signature of the session callback configured with
198 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
199 was explicitly marked as 'const unsigned char*' instead of
203 *) Always DPURIFY. Remove the use of uninitialized memory in the
204 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
207 *) Removed many obsolete configuration items, including
208 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
209 MD2_CHAR, MD2_INT, MD2_LONG
211 IDEA_SHORT, IDEA_LONG
212 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
213 [Rich Salz, with advice from Andy Polyakov]
215 *) Many BN internals have been moved to an internal header file.
216 [Rich Salz with help from Andy Polyakov]
218 *) Configuration and writing out the results from it has changed.
219 Files such as Makefile include/openssl/opensslconf.h and are now
220 produced through general templates, such as Makefile.in and
221 crypto/opensslconf.h.in and some help from the perl module
224 Also, the center of configuration information is no longer
225 Makefile. Instead, Configure produces a perl module in
226 configdata.pm which holds most of the config data (in the hash
227 table %config), the target data that comes from the target
228 configuration in one of the Configurations/*.conf files (in
232 *) To clarify their intended purposes, the Configure options
233 --prefix and --openssldir change their semantics, and become more
234 straightforward and less interdependent.
236 --prefix shall be used exclusively to give the location INSTALLTOP
237 where programs, scripts, libraries, include files and manuals are
238 going to be installed. The default is now /usr/local.
240 --openssldir shall be used exclusively to give the default
241 location OPENSSLDIR where certificates, private keys, CRLs are
242 managed. This is also where the default openssl.cnf gets
244 If the directory given with this option is a relative path, the
245 values of both the --prefix value and the --openssldir value will
246 be combined to become OPENSSLDIR.
247 The default for --openssldir is INSTALLTOP/ssl.
249 Anyone who uses --openssldir to specify where OpenSSL is to be
250 installed MUST change to use --prefix instead.
253 *) The GOST engine was out of date and therefore it has been removed. An up
254 to date GOST engine is now being maintained in an external repository.
255 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
256 support for GOST ciphersuites (these are only activated if a GOST engine
260 *) EGD is no longer supported by default; use enable-egd when
262 [Ben Kaduk and Rich Salz]
264 *) The distribution now has Makefile.in files, which are used to
265 create Makefile's when Configure is run. *Configure must be run
266 before trying to build now.*
269 *) The return value for SSL_CIPHER_description() for error conditions
273 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
275 Obtaining and performing DNSSEC validation of TLSA records is
276 the application's responsibility. The application provides
277 the TLSA records of its choice to OpenSSL, and these are then
278 used to authenticate the peer.
280 The TLSA records need not even come from DNS. They can, for
281 example, be used to implement local end-entity certificate or
282 trust-anchor "pinning", where the "pin" data takes the form
283 of TLSA records, which can augment or replace verification
284 based on the usual WebPKI public certification authorities.
287 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
288 continues to support deprecated interfaces in default builds.
289 However, applications are strongly advised to compile their
290 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
291 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
292 or the 1.1.0 releases.
294 In environments in which all applications have been ported to
295 not use any deprecated interfaces OpenSSL's Configure script
296 should be used with the --api=1.1.0 option to entirely remove
297 support for the deprecated features from the library and
298 unconditionally disable them in the installed headers.
299 Essentially the same effect can be achieved with the "no-deprecated"
300 argument to Configure, except that this will always restrict
301 the build to just the latest API, rather than a fixed API
304 As applications are ported to future revisions of the API,
305 they should update their compile-time OPENSSL_API_COMPAT define
306 accordingly, but in most cases should be able to continue to
307 compile with later releases.
309 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
310 0x10000000L and 0x00908000L, respectively. However those
311 versions did not support the OPENSSL_API_COMPAT feature, and
312 so applications are not typically tested for explicit support
313 of just the undeprecated features of either release.
316 *) Add support for setting the minimum and maximum supported protocol.
317 It can bet set via the SSL_set_min_proto_version() and
318 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
319 MaxProtcol. It's recommended to use the new APIs to disable
320 protocols instead of disabling individual protocols using
321 SSL_set_options() or SSL_CONF's Protocol. This change also
322 removes support for disabling TLS 1.2 in the OpenSSL TLS
323 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
326 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
329 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
330 and integrates ECDSA and ECDH functionality into EC. Implementations can
331 now redirect key generation and no longer need to convert to or from
334 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
335 include the ec.h header file instead.
338 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
339 ciphers who are no longer supported and drops support the ephemeral RSA key
340 exchange. The LOW ciphers currently doesn't have any ciphers in it.
343 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
344 opaque. For HMAC_CTX, the following constructors and destructors
347 HMAC_CTX *HMAC_CTX_new(void);
348 void HMAC_CTX_free(HMAC_CTX *ctx);
350 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
351 destroy such methods has been added. See EVP_MD_meth_new(3) and
352 EVP_CIPHER_meth_new(3) for documentation.
355 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
356 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
357 EVP_MD_CTX_reset() should be called instead to reinitialise
358 an already created structure.
359 2) For consistency with the majority of our object creators and
360 destructors, EVP_MD_CTX_(create|destroy) were renamed to
361 EVP_MD_CTX_(new|free). The old names are retained as macros
362 for deprecated builds.
365 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
366 cryptographic operations to be performed asynchronously as long as an
367 asynchronous capable engine is used. See the ASYNC_start_job() man page for
368 further details. Libssl has also had this capability integrated with the
369 introduction of the new mode SSL_MODE_ASYNC and associated error
370 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
371 pages. This work was developed in partnership with Intel Corp.
374 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
375 always enabled now. If you want to disable the support you should
376 exclude it using the list of supported ciphers. This also means that the
377 "-no_ecdhe" option has been removed from s_server.
380 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
381 SSL_{CTX_}set1_curves() which can set a list.
384 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
385 curve you want to support using SSL_{CTX_}set1_curves().
388 *) State machine rewrite. The state machine code has been significantly
389 refactored in order to remove much duplication of code and solve issues
390 with the old code (see ssl/statem/README for further details). This change
391 does have some associated API changes. Notably the SSL_state() function
392 has been removed and replaced by SSL_get_state which now returns an
393 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
394 altogether. The previous handshake states defined in ssl.h and ssl3.h have
398 *) All instances of the string "ssleay" in the public API were replaced
399 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
400 Some error codes related to internal RSA_eay API's were renamed.
403 *) The demo files in crypto/threads were moved to demo/threads.
406 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
408 [Matt Caswell, Rich Salz]
410 *) New ASN.1 embed macro.
412 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
413 structure is not allocated: it is part of the parent. That is instead of
421 This reduces memory fragmentation and make it impossible to accidentally
422 set a mandatory field to NULL.
424 This currently only works for some fields specifically a SEQUENCE, CHOICE,
425 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
426 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
430 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
433 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
434 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
435 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
436 DES and RC4 ciphersuites.
439 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
440 This changes the decoding behaviour for some invalid messages,
441 though the change is mostly in the more lenient direction, and
442 legacy behaviour is preserved as much as possible.
445 *) Fix no-stdio build.
446 [ David Woodhouse <David.Woodhouse@intel.com> and also
447 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
449 *) New testing framework
450 The testing framework has been largely rewritten and is now using
451 perl and the perl modules Test::Harness and an extended variant of
452 Test::More called OpenSSL::Test to do its work. All test scripts in
453 test/ have been rewritten into test recipes, and all direct calls to
454 executables in test/Makefile have become individual recipes using the
455 simplified testing OpenSSL::Test::Simple.
457 For documentation on our testing modules, do:
459 perldoc test/testlib/OpenSSL/Test/Simple.pm
460 perldoc test/testlib/OpenSSL/Test.pm
464 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
465 are used; the latter aborts on memory leaks (usually checked on exit).
466 Some undocumented "set malloc, etc., hooks" functions were removed
467 and others were changed. All are now documented.
470 *) In DSA_generate_parameters_ex, if the provided seed is too short,
472 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
474 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
475 from RFC4279, RFC4785, RFC5487, RFC5489.
477 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
478 original RSA_PSK patch.
481 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
482 era flag was never set throughout the codebase (only read). Also removed
483 SSL3_FLAGS_POP_BUFFER which was only used if
484 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
487 *) Changed the default name options in the "ca", "crl", "req" and "x509"
488 to be "oneline" instead of "compat".
491 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
492 not aware of clients that still exhibit this bug, and the workaround
493 hasn't been working properly for a while.
496 *) The return type of BIO_number_read() and BIO_number_written() as well as
497 the corresponding num_read and num_write members in the BIO structure has
498 changed from unsigned long to uint64_t. On platforms where an unsigned
499 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
503 *) Given the pervasive nature of TLS extensions it is inadvisable to run
504 OpenSSL without support for them. It also means that maintaining
505 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
506 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
509 *) Removed support for the two export grade static DH ciphersuites
510 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
511 were newly added (along with a number of other static DH ciphersuites) to
512 1.0.2. However the two export ones have *never* worked since they were
513 introduced. It seems strange in any case to be adding new export
514 ciphersuites, and given "logjam" it also does not seem correct to fix them.
517 *) Version negotiation has been rewritten. In particular SSLv23_method(),
518 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
519 and turned into macros which simply call the new preferred function names
520 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
521 should use the new names instead. Also as part of this change the ssl23.h
522 header file has been removed.
525 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
526 code and the associated standard is no longer considered fit-for-purpose.
529 *) RT2547 was closed. When generating a private key, try to make the
530 output file readable only by the owner. This behavior change might
531 be noticeable when interacting with other software.
533 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
537 *) Added HTTP GET support to the ocsp command.
540 *) Changed default digest for the dgst and enc commands from MD5 to
544 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
547 *) Added support for TLS extended master secret from
548 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
549 initial patch which was a great help during development.
552 *) All libssl internal structures have been removed from the public header
553 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
554 now redundant). Users should not attempt to access internal structures
555 directly. Instead they should use the provided API functions.
558 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
559 Access to deprecated functions can be re-enabled by running config with
560 "enable-deprecated". In addition applications wishing to use deprecated
561 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
562 will, by default, disable some transitive includes that previously existed
563 in the header files (e.g. ec.h will no longer, by default, include bn.h)
566 *) Added support for OCB mode. OpenSSL has been granted a patent license
567 compatible with the OpenSSL license for use of OCB. Details are available
568 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
569 for OCB can be removed by calling config with no-ocb.
572 *) SSLv2 support has been removed. It still supports receiving a SSLv2
573 compatible client hello.
576 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
577 done while fixing the error code for the key-too-small case.
578 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
580 *) CA.sh has been removmed; use CA.pl instead.
583 *) Removed old DES API.
586 *) Remove various unsupported platforms:
592 Sinix/ReliantUNIX RM400
597 16-bit platforms such as WIN16
600 *) Clean up OPENSSL_NO_xxx #define's
601 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
602 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
603 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
604 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
605 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
606 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
607 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
608 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
609 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
610 Remove MS_STATIC; it's a relic from platforms <32 bits.
613 *) Cleaned up dead code
614 Remove all but one '#ifdef undef' which is to be looked at.
617 *) Clean up calling of xxx_free routines.
618 Just like free(), fix most of the xxx_free routines to accept
619 NULL. Remove the non-null checks from callers. Save much code.
622 *) Add secure heap for storage of private keys (when possible).
623 Add BIO_s_secmem(), CBIGNUM, etc.
624 Contributed by Akamai Technologies under our Corporate CLA.
627 *) Experimental support for a new, fast, unbiased prime candidate generator,
628 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
629 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
631 *) New output format NSS in the sess_id command line tool. This allows
632 exporting the session id and the master key in NSS keylog format.
633 [Martin Kaiser <martin@kaiser.cx>]
635 *) Harmonize version and its documentation. -f flag is used to display
637 [mancha <mancha1@zoho.com>]
639 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
640 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
641 [mancha <mancha1@zoho.com>]
643 *) Fix some double frees. These are not thought to be exploitable.
644 [mancha <mancha1@zoho.com>]
646 *) A missing bounds check in the handling of the TLS heartbeat extension
647 can be used to reveal up to 64k of memory to a connected client or
650 Thanks for Neel Mehta of Google Security for discovering this bug and to
651 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
652 preparing the fix (CVE-2014-0160)
653 [Adam Langley, Bodo Moeller]
655 *) Fix for the attack described in the paper "Recovering OpenSSL
656 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
657 by Yuval Yarom and Naomi Benger. Details can be obtained from:
658 http://eprint.iacr.org/2014/140
660 Thanks to Yuval Yarom and Naomi Benger for discovering this
661 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
662 [Yuval Yarom and Naomi Benger]
664 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
665 this fixes a limitation in previous versions of OpenSSL.
668 *) Experimental encrypt-then-mac support.
670 Experimental support for encrypt then mac from
671 draft-gutmann-tls-encrypt-then-mac-02.txt
673 To enable it set the appropriate extension number (0x42 for the test
674 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
676 For non-compliant peers (i.e. just about everything) this should have no
679 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
683 *) Add EVP support for key wrapping algorithms, to avoid problems with
684 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
685 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
686 algorithms and include tests cases.
689 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
693 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
694 MGF1 digest and OAEP label.
697 *) Make openssl verify return errors.
698 [Chris Palmer <palmer@google.com> and Ben Laurie]
700 *) New function ASN1_TIME_diff to calculate the difference between two
701 ASN1_TIME structures or one structure and the current time.
704 *) Update fips_test_suite to support multiple command line options. New
705 test to induce all self test errors in sequence and check expected
709 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
710 sign or verify all in one operation.
713 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
714 test programs and fips_test_suite. Includes functionality to parse
715 the minimal script output of fipsalgest.pl directly.
718 *) Add authorisation parameter to FIPS_module_mode_set().
721 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
724 *) Use separate DRBG fields for internal and external flags. New function
725 FIPS_drbg_health_check() to perform on demand health checking. Add
726 generation tests to fips_test_suite with reduced health check interval to
727 demonstrate periodic health checking. Add "nodh" option to
728 fips_test_suite to skip very slow DH test.
731 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
735 *) More extensive health check for DRBG checking many more failure modes.
736 New function FIPS_selftest_drbg_all() to handle every possible DRBG
737 combination: call this in fips_test_suite.
740 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
741 and POST to handle Dual EC cases.
744 *) Add support for canonical generation of DSA parameter 'g'. See
747 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
748 POST to handle HMAC cases.
751 *) Add functions FIPS_module_version() and FIPS_module_version_text()
752 to return numerical and string versions of the FIPS module number.
755 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
756 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
757 outside the validated module in the FIPS capable OpenSSL.
760 *) Minor change to DRBG entropy callback semantics. In some cases
761 there is no multiple of the block length between min_len and
762 max_len. Allow the callback to return more than max_len bytes
763 of entropy but discard any extra: it is the callback's responsibility
764 to ensure that the extra data discarded does not impact the
765 requested amount of entropy.
768 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
769 information in FIPS186-3, SP800-57 and SP800-131A.
772 *) CCM support via EVP. Interface is very similar to GCM case except we
773 must supply all data in one chunk (i.e. no update, final) and the
774 message length must be supplied if AAD is used. Add algorithm test
778 *) Initial version of POST overhaul. Add POST callback to allow the status
779 of POST to be monitored and/or failures induced. Modify fips_test_suite
780 to use callback. Always run all selftests even if one fails.
783 *) XTS support including algorithm test driver in the fips_gcmtest program.
784 Note: this does increase the maximum key length from 32 to 64 bytes but
785 there should be no binary compatibility issues as existing applications
786 will never use XTS mode.
789 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
790 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
791 performs algorithm blocking for unapproved PRNG types. Also do not
792 set PRNG type in FIPS_mode_set(): leave this to the application.
793 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
794 the standard OpenSSL PRNG: set additional data to a date time vector.
797 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
798 This shouldn't present any incompatibility problems because applications
799 shouldn't be using these directly and any that are will need to rethink
800 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
803 *) Extensive self tests and health checking required by SP800-90 DRBG.
804 Remove strength parameter from FIPS_drbg_instantiate and always
805 instantiate at maximum supported strength.
808 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
811 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
814 *) New function DH_compute_key_padded() to compute a DH key and pad with
815 leading zeroes if needed: this complies with SP800-56A et al.
818 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
819 anything, incomplete, subject to change and largely untested at present.
822 *) Modify fipscanisteronly build option to only build the necessary object
823 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
826 *) Add experimental option FIPSSYMS to give all symbols in
827 fipscanister.o and FIPS or fips prefix. This will avoid
828 conflicts with future versions of OpenSSL. Add perl script
829 util/fipsas.pl to preprocess assembly language source files
830 and rename any affected symbols.
833 *) Add selftest checks and algorithm block of non-fips algorithms in
834 FIPS mode. Remove DES2 from selftests.
837 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
838 return internal method without any ENGINE dependencies. Add new
839 tiny fips sign and verify functions.
842 *) New build option no-ec2m to disable characteristic 2 code.
845 *) New build option "fipscanisteronly". This only builds fipscanister.o
846 and (currently) associated fips utilities. Uses the file Makefile.fips
847 instead of Makefile.org as the prototype.
850 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
851 Update fips_gcmtest to use IV generator.
854 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
855 setting output buffer to NULL. The *Final function must be
856 called although it will not retrieve any additional data. The tag
857 can be set or retrieved with a ctrl. The IV length is by default 12
858 bytes (96 bits) but can be set to an alternative value. If the IV
859 length exceeds the maximum IV length (currently 16 bytes) it cannot be
863 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
864 underlying do_cipher function handles all cipher semantics itself
865 including padding and finalisation. This is useful if (for example)
866 an ENGINE cipher handles block padding itself. The behaviour of
867 do_cipher is subtly changed if this flag is set: the return value
868 is the number of characters written to the output buffer (zero is
869 no longer an error code) or a negative error code. Also if the
870 input buffer is NULL and length 0 finalisation should be performed.
873 *) If a candidate issuer certificate is already part of the constructed
874 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
877 *) Improve forward-security support: add functions
879 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
880 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
882 for use by SSL/TLS servers; the callback function will be called whenever a
883 new session is created, and gets to decide whether the session may be
884 cached to make it resumable (return 0) or not (return 1). (As by the
885 SSL/TLS protocol specifications, the session_id sent by the server will be
886 empty to indicate that the session is not resumable; also, the server will
887 not generate RFC 4507 (RFC 5077) session tickets.)
889 A simple reasonable callback implementation is to return is_forward_secure.
890 This parameter will be set to 1 or 0 depending on the ciphersuite selected
891 by the SSL/TLS server library, indicating whether it can provide forward
893 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
895 *) New -verify_name option in command line utilities to set verification
899 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
900 Add CMAC pkey methods.
903 *) Experimental renegotiation in s_server -www mode. If the client
904 browses /reneg connection is renegotiated. If /renegcert it is
905 renegotiated requesting a certificate.
908 *) Add an "external" session cache for debugging purposes to s_server. This
909 should help trace issues which normally are only apparent in deployed
910 multi-process servers.
913 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
914 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
915 BIO_set_cipher() and some obscure PEM functions were changed so they
916 can now return an error. The RAND changes required a change to the
917 RAND_METHOD structure.
920 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
921 a gcc attribute to warn if the result of a function is ignored. This
922 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
923 whose return value is often ignored.
926 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
927 These allow SCTs (signed certificate timestamps) to be requested and
928 validated when establishing a connection.
929 [Rob Percival <robpercival@google.com>]
931 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
933 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
934 Builds that are not configured with "enable-weak-ssl-ciphers" will not
935 provide any "EXPORT" or "LOW" strength ciphers.
938 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
939 is by default disabled at build-time. Builds that are not configured with
940 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
941 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
942 will need to explicitly call either of:
944 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
946 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
948 as appropriate. Even if either of those is used, or the application
949 explicitly uses the version-specific SSLv2_method() or its client and
950 server variants, SSLv2 ciphers vulnerable to exhaustive search key
951 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
952 ciphers, and SSLv2 56-bit DES are no longer available.
956 *) Fix a double-free in DSA code
958 A double free bug was discovered when OpenSSL parses malformed DSA private
959 keys and could lead to a DoS attack or memory corruption for applications
960 that receive DSA private keys from untrusted sources. This scenario is
963 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
968 *) Disable SRP fake user seed to address a server memory leak.
970 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
972 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
973 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
974 was changed to ignore the "fake user" SRP seed, even if the seed
977 Users should use SRP_VBASE_get1_by_user instead. Note that in
978 SRP_VBASE_get1_by_user, caller must free the returned value. Note
979 also that even though configuring the SRP seed attempts to hide
980 invalid usernames by continuing the handshake with fake
981 credentials, this behaviour is not constant time and no strong
982 guarantees are made that the handshake is indistinguishable from
983 that of a valid user.
987 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
989 In the BN_hex2bn function the number of hex digits is calculated using an
990 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
991 large values of |i| this can result in |bn_expand| not allocating any
992 memory because |i * 4| is negative. This can leave the internal BIGNUM data
993 field as NULL leading to a subsequent NULL ptr deref. For very large values
994 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
995 In this case memory is allocated to the internal BIGNUM data field, but it
996 is insufficiently sized leading to heap corruption. A similar issue exists
997 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
998 is ever called by user applications with very large untrusted hex/dec data.
999 This is anticipated to be a rare occurrence.
1001 All OpenSSL internal usage of these functions use data that is not expected
1002 to be untrusted, e.g. config file data or application command line
1003 arguments. If user developed applications generate config file data based
1004 on untrusted data then it is possible that this could also lead to security
1005 consequences. This is also anticipated to be rare.
1007 This issue was reported to OpenSSL by Guido Vranken.
1011 *) Fix memory issues in BIO_*printf functions
1013 The internal |fmtstr| function used in processing a "%s" format string in
1014 the BIO_*printf functions could overflow while calculating the length of a
1015 string and cause an OOB read when printing very long strings.
1017 Additionally the internal |doapr_outch| function can attempt to write to an
1018 OOB memory location (at an offset from the NULL pointer) in the event of a
1019 memory allocation failure. In 1.0.2 and below this could be caused where
1020 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1021 could be in processing a very long "%s" format string. Memory leaks can
1024 The first issue may mask the second issue dependent on compiler behaviour.
1025 These problems could enable attacks where large amounts of untrusted data
1026 is passed to the BIO_*printf functions. If applications use these functions
1027 in this way then they could be vulnerable. OpenSSL itself uses these
1028 functions when printing out human-readable dumps of ASN.1 data. Therefore
1029 applications that print this data could be vulnerable if the data is from
1030 untrusted sources. OpenSSL command line applications could also be
1031 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1032 as command line arguments.
1034 Libssl is not considered directly vulnerable. Additionally certificates etc
1035 received via remote connections via libssl are also unlikely to be able to
1036 trigger these issues because of message size limits enforced within libssl.
1038 This issue was reported to OpenSSL Guido Vranken.
1042 *) Side channel attack on modular exponentiation
1044 A side-channel attack was found which makes use of cache-bank conflicts on
1045 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1046 of RSA keys. The ability to exploit this issue is limited as it relies on
1047 an attacker who has control of code in a thread running on the same
1048 hyper-threaded core as the victim thread which is performing decryptions.
1050 This issue was reported to OpenSSL by Yuval Yarom, The University of
1051 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1052 Nadia Heninger, University of Pennsylvania with more information at
1053 http://cachebleed.info.
1057 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1058 if no keysize is specified with default_bits. This fixes an
1059 omission in an earlier change that changed all RSA/DSA key generation
1060 apps to use 2048 bits by default.
1063 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1064 *) DH small subgroups
1066 Historically OpenSSL only ever generated DH parameters based on "safe"
1067 primes. More recently (in version 1.0.2) support was provided for
1068 generating X9.42 style parameter files such as those required for RFC 5114
1069 support. The primes used in such files may not be "safe". Where an
1070 application is using DH configured with parameters based on primes that are
1071 not "safe" then an attacker could use this fact to find a peer's private
1072 DH exponent. This attack requires that the attacker complete multiple
1073 handshakes in which the peer uses the same private DH exponent. For example
1074 this could be used to discover a TLS server's private DH exponent if it's
1075 reusing the private DH exponent or it's using a static DH ciphersuite.
1077 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1078 TLS. It is not on by default. If the option is not set then the server
1079 reuses the same private DH exponent for the life of the server process and
1080 would be vulnerable to this attack. It is believed that many popular
1081 applications do set this option and would therefore not be at risk.
1083 The fix for this issue adds an additional check where a "q" parameter is
1084 available (as is the case in X9.42 based parameters). This detects the
1085 only known attack, and is the only possible defense for static DH
1086 ciphersuites. This could have some performance impact.
1088 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1089 default and cannot be disabled. This could have some performance impact.
1091 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1095 *) SSLv2 doesn't block disabled ciphers
1097 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1098 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1099 been disabled, provided that the SSLv2 protocol was not also disabled via
1102 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1103 and Sebastian Schinzel.
1107 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1109 *) BN_mod_exp may produce incorrect results on x86_64
1111 There is a carry propagating bug in the x86_64 Montgomery squaring
1112 procedure. No EC algorithms are affected. Analysis suggests that attacks
1113 against RSA and DSA as a result of this defect would be very difficult to
1114 perform and are not believed likely. Attacks against DH are considered just
1115 feasible (although very difficult) because most of the work necessary to
1116 deduce information about a private key may be performed offline. The amount
1117 of resources required for such an attack would be very significant and
1118 likely only accessible to a limited number of attackers. An attacker would
1119 additionally need online access to an unpatched system using the target
1120 private key in a scenario with persistent DH parameters and a private
1121 key that is shared between multiple clients. For example this can occur by
1122 default in OpenSSL DHE based SSL/TLS ciphersuites.
1124 This issue was reported to OpenSSL by Hanno Böck.
1128 *) Certificate verify crash with missing PSS parameter
1130 The signature verification routines will crash with a NULL pointer
1131 dereference if presented with an ASN.1 signature using the RSA PSS
1132 algorithm and absent mask generation function parameter. Since these
1133 routines are used to verify certificate signature algorithms this can be
1134 used to crash any certificate verification operation and exploited in a
1135 DoS attack. Any application which performs certificate verification is
1136 vulnerable including OpenSSL clients and servers which enable client
1139 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1143 *) X509_ATTRIBUTE memory leak
1145 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1146 memory. This structure is used by the PKCS#7 and CMS routines so any
1147 application which reads PKCS#7 or CMS data from untrusted sources is
1148 affected. SSL/TLS is not affected.
1150 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1155 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1156 This changes the decoding behaviour for some invalid messages,
1157 though the change is mostly in the more lenient direction, and
1158 legacy behaviour is preserved as much as possible.
1161 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1163 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1165 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1167 *) Alternate chains certificate forgery
1169 During certificate verfification, OpenSSL will attempt to find an
1170 alternative certificate chain if the first attempt to build such a chain
1171 fails. An error in the implementation of this logic can mean that an
1172 attacker could cause certain checks on untrusted certificates to be
1173 bypassed, such as the CA flag, enabling them to use a valid leaf
1174 certificate to act as a CA and "issue" an invalid certificate.
1176 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1180 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1182 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1183 incompatibility in the handling of HMAC. The previous ABI has now been
1187 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1189 *) Malformed ECParameters causes infinite loop
1191 When processing an ECParameters structure OpenSSL enters an infinite loop
1192 if the curve specified is over a specially malformed binary polynomial
1195 This can be used to perform denial of service against any
1196 system which processes public keys, certificate requests or
1197 certificates. This includes TLS clients and TLS servers with
1198 client authentication enabled.
1200 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1204 *) Exploitable out-of-bounds read in X509_cmp_time
1206 X509_cmp_time does not properly check the length of the ASN1_TIME
1207 string and can read a few bytes out of bounds. In addition,
1208 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1211 An attacker can use this to craft malformed certificates and CRLs of
1212 various sizes and potentially cause a segmentation fault, resulting in
1213 a DoS on applications that verify certificates or CRLs. TLS clients
1214 that verify CRLs are affected. TLS clients and servers with client
1215 authentication enabled may be affected if they use custom verification
1218 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1219 independently by Hanno Böck.
1223 *) PKCS7 crash with missing EnvelopedContent
1225 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1226 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1227 with missing content and trigger a NULL pointer dereference on parsing.
1229 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1230 structures from untrusted sources are affected. OpenSSL clients and
1231 servers are not affected.
1233 This issue was reported to OpenSSL by Michal Zalewski (Google).
1237 *) CMS verify infinite loop with unknown hash function
1239 When verifying a signedData message the CMS code can enter an infinite loop
1240 if presented with an unknown hash function OID. This can be used to perform
1241 denial of service against any system which verifies signedData messages using
1243 This issue was reported to OpenSSL by Johannes Bauer.
1247 *) Race condition handling NewSessionTicket
1249 If a NewSessionTicket is received by a multi-threaded client when attempting to
1250 reuse a previous ticket then a race condition can occur potentially leading to
1251 a double free of the ticket data.
1255 *) Only support 256-bit or stronger elliptic curves with the
1256 'ecdh_auto' setting (server) or by default (client). Of supported
1257 curves, prefer P-256 (both).
1260 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1262 *) ClientHello sigalgs DoS fix
1264 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1265 invalid signature algorithms extension a NULL pointer dereference will
1266 occur. This can be exploited in a DoS attack against the server.
1268 This issue was was reported to OpenSSL by David Ramos of Stanford
1271 [Stephen Henson and Matt Caswell]
1273 *) Multiblock corrupted pointer fix
1275 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1276 feature only applies on 64 bit x86 architecture platforms that support AES
1277 NI instructions. A defect in the implementation of "multiblock" can cause
1278 OpenSSL's internal write buffer to become incorrectly set to NULL when
1279 using non-blocking IO. Typically, when the user application is using a
1280 socket BIO for writing, this will only result in a failed connection.
1281 However if some other BIO is used then it is likely that a segmentation
1282 fault will be triggered, thus enabling a potential DoS attack.
1284 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1288 *) Segmentation fault in DTLSv1_listen fix
1290 The DTLSv1_listen function is intended to be stateless and processes the
1291 initial ClientHello from many peers. It is common for user code to loop
1292 over the call to DTLSv1_listen until a valid ClientHello is received with
1293 an associated cookie. A defect in the implementation of DTLSv1_listen means
1294 that state is preserved in the SSL object from one invocation to the next
1295 that can lead to a segmentation fault. Errors processing the initial
1296 ClientHello can trigger this scenario. An example of such an error could be
1297 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1300 This issue was reported to OpenSSL by Per Allansson.
1304 *) Segmentation fault in ASN1_TYPE_cmp fix
1306 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1307 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1308 certificate signature algorithm consistency this can be used to crash any
1309 certificate verification operation and exploited in a DoS attack. Any
1310 application which performs certificate verification is vulnerable including
1311 OpenSSL clients and servers which enable client authentication.
1315 *) Segmentation fault for invalid PSS parameters fix
1317 The signature verification routines will crash with a NULL pointer
1318 dereference if presented with an ASN.1 signature using the RSA PSS
1319 algorithm and invalid parameters. Since these routines are used to verify
1320 certificate signature algorithms this can be used to crash any
1321 certificate verification operation and exploited in a DoS attack. Any
1322 application which performs certificate verification is vulnerable including
1323 OpenSSL clients and servers which enable client authentication.
1325 This issue was was reported to OpenSSL by Brian Carpenter.
1329 *) ASN.1 structure reuse memory corruption fix
1331 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1332 memory corruption via an invalid write. Such reuse is and has been
1333 strongly discouraged and is believed to be rare.
1335 Applications that parse structures containing CHOICE or ANY DEFINED BY
1336 components may be affected. Certificate parsing (d2i_X509 and related
1337 functions) are however not affected. OpenSSL clients and servers are
1342 *) PKCS7 NULL pointer dereferences fix
1344 The PKCS#7 parsing code does not handle missing outer ContentInfo
1345 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1346 missing content and trigger a NULL pointer dereference on parsing.
1348 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1349 otherwise parse PKCS#7 structures from untrusted sources are
1350 affected. OpenSSL clients and servers are not affected.
1352 This issue was reported to OpenSSL by Michal Zalewski (Google).
1356 *) DoS via reachable assert in SSLv2 servers fix
1358 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1359 servers that both support SSLv2 and enable export cipher suites by sending
1360 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1362 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1363 (OpenSSL development team).
1367 *) Empty CKE with client auth and DHE fix
1369 If client auth is used then a server can seg fault in the event of a DHE
1370 ciphersuite being selected and a zero length ClientKeyExchange message
1371 being sent by the client. This could be exploited in a DoS attack.
1375 *) Handshake with unseeded PRNG fix
1377 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1378 with an unseeded PRNG. The conditions are:
1379 - The client is on a platform where the PRNG has not been seeded
1380 automatically, and the user has not seeded manually
1381 - A protocol specific client method version has been used (i.e. not
1382 SSL_client_methodv23)
1383 - A ciphersuite is used that does not require additional random data from
1384 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1386 If the handshake succeeds then the client random that has been used will
1387 have been generated from a PRNG with insufficient entropy and therefore the
1388 output may be predictable.
1390 For example using the following command with an unseeded openssl will
1391 succeed on an unpatched platform:
1393 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1397 *) Use After Free following d2i_ECPrivatekey error fix
1399 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1400 could cause a use after free condition. This, in turn, could cause a double
1401 free in several private key parsing functions (such as d2i_PrivateKey
1402 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1403 for applications that receive EC private keys from untrusted
1404 sources. This scenario is considered rare.
1406 This issue was discovered by the BoringSSL project and fixed in their
1411 *) X509_to_X509_REQ NULL pointer deref fix
1413 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1414 the certificate key is invalid. This function is rarely used in practice.
1416 This issue was discovered by Brian Carpenter.
1420 *) Removed the export ciphers from the DEFAULT ciphers
1423 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1425 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1426 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1427 So far those who have to target multiple plaforms would compromise
1428 and argue that binary targeting say ARMv5 would still execute on
1429 ARMv8. "Universal" build resolves this compromise by providing
1430 near-optimal performance even on newer platforms.
1433 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1434 (other platforms pending).
1435 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1437 *) Add support for the SignedCertificateTimestampList certificate and
1438 OCSP response extensions from RFC6962.
1441 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1442 for corner cases. (Certain input points at infinity could lead to
1443 bogus results, with non-infinity inputs mapped to infinity too.)
1446 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1447 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1448 common cases are optimized and there still is room for further
1449 improvements. Vector Permutation AES for Altivec is also added.
1452 *) Add support for little-endian ppc64 Linux target.
1453 [Marcelo Cerri (IBM)]
1455 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1456 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1457 are optimized and there still is room for further improvements.
1458 Both 32- and 64-bit modes are supported.
1459 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1461 *) Improved ARMv7 NEON support.
1464 *) Support for SPARC Architecture 2011 crypto extensions, first
1465 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1466 SHA256/512, MD5, GHASH and modular exponentiation.
1467 [Andy Polyakov, David Miller]
1469 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1471 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1473 *) Support for new and upcoming Intel processors, including AVX2,
1474 BMI and SHA ISA extensions. This includes additional "stitched"
1475 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1478 This work was sponsored by Intel Corp.
1481 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1482 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1483 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1486 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1487 this fixes a limiation in previous versions of OpenSSL.
1490 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1491 MGF1 digest and OAEP label.
1494 *) Add EVP support for key wrapping algorithms, to avoid problems with
1495 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1496 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1497 algorithms and include tests cases.
1500 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1502 [Douglas E. Engert, Steve Henson]
1504 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1505 difference in days and seconds between two tm or ASN1_TIME structures.
1508 *) Add -rev test option to s_server to just reverse order of characters
1509 received by client and send back to server. Also prints an abbreviated
1510 summary of the connection parameters.
1513 *) New option -brief for s_client and s_server to print out a brief summary
1514 of connection parameters.
1517 *) Add callbacks for arbitrary TLS extensions.
1518 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1520 *) New option -crl_download in several openssl utilities to download CRLs
1521 from CRLDP extension in certificates.
1524 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1527 *) New function X509_CRL_diff to generate a delta CRL from the difference
1528 of two full CRLs. Add support to "crl" utility.
1531 *) New functions to set lookup_crls function and to retrieve
1532 X509_STORE from X509_STORE_CTX.
1535 *) Print out deprecated issuer and subject unique ID fields in
1539 *) Extend OCSP I/O functions so they can be used for simple general purpose
1540 HTTP as well as OCSP. New wrapper function which can be used to download
1541 CRLs using the OCSP API.
1544 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1547 *) SSL_CONF* functions. These provide a common framework for application
1548 configuration using configuration files or command lines.
1551 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1552 message callback and prints the results. Needs compile time option
1553 "enable-ssl-trace". New options to s_client and s_server to enable
1557 *) New ctrl and macro to retrieve supported points extensions.
1558 Print out extension in s_server and s_client.
1561 *) New functions to retrieve certificate signature and signature
1565 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1569 *) New Suite B modes for TLS code. These use and enforce the requirements
1570 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1571 only use Suite B curves. The Suite B modes can be set by using the
1572 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1575 *) New chain verification flags for Suite B levels of security. Check
1576 algorithms are acceptable when flags are set in X509_verify_cert.
1579 *) Make tls1_check_chain return a set of flags indicating checks passed
1580 by a certificate chain. Add additional tests to handle client
1581 certificates: checks for matching certificate type and issuer name
1585 *) If an attempt is made to use a signature algorithm not in the peer
1586 preference list abort the handshake. If client has no suitable
1587 signature algorithms in response to a certificate request do not
1588 use the certificate.
1591 *) If server EC tmp key is not in client preference list abort handshake.
1594 *) Add support for certificate stores in CERT structure. This makes it
1595 possible to have different stores per SSL structure or one store in
1596 the parent SSL_CTX. Include distint stores for certificate chain
1597 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1598 to build and store a certificate chain in CERT structure: returing
1599 an error if the chain cannot be built: this will allow applications
1600 to test if a chain is correctly configured.
1602 Note: if the CERT based stores are not set then the parent SSL_CTX
1603 store is used to retain compatibility with existing behaviour.
1607 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1608 mask based on the current session, check mask when sending client
1609 hello and checking the requested ciphersuite.
1612 *) New ctrls to retrieve and set certificate types in a certificate
1613 request message. Print out received values in s_client. If certificate
1614 types is not set with custom values set sensible values based on
1615 supported signature algorithms.
1618 *) Support for distinct client and server supported signature algorithms.
1621 *) Add certificate callback. If set this is called whenever a certificate
1622 is required by client or server. An application can decide which
1623 certificate chain to present based on arbitrary criteria: for example
1624 supported signature algorithms. Add very simple example to s_server.
1625 This fixes many of the problems and restrictions of the existing client
1626 certificate callback: for example you can now clear an existing
1627 certificate and specify the whole chain.
1630 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1631 the certificate can be used for (if anything). Set valid_flags field
1632 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1633 to have similar checks in it.
1635 Add new "cert_flags" field to CERT structure and include a "strict mode".
1636 This enforces some TLS certificate requirements (such as only permitting
1637 certificate signature algorithms contained in the supported algorithms
1638 extension) which some implementations ignore: this option should be used
1639 with caution as it could cause interoperability issues.
1642 *) Update and tidy signature algorithm extension processing. Work out
1643 shared signature algorithms based on preferences and peer algorithms
1644 and print them out in s_client and s_server. Abort handshake if no
1645 shared signature algorithms.
1648 *) Add new functions to allow customised supported signature algorithms
1649 for SSL and SSL_CTX structures. Add options to s_client and s_server
1653 *) New function SSL_certs_clear() to delete all references to certificates
1654 from an SSL structure. Before this once a certificate had been added
1655 it couldn't be removed.
1658 *) Integrate hostname, email address and IP address checking with certificate
1659 verification. New verify options supporting checking in opensl utility.
1662 *) Fixes and wildcard matching support to hostname and email checking
1663 functions. Add manual page.
1664 [Florian Weimer (Red Hat Product Security Team)]
1666 *) New functions to check a hostname email or IP address against a
1667 certificate. Add options x509 utility to print results of checks against
1671 *) Fix OCSP checking.
1672 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1674 *) Initial experimental support for explicitly trusted non-root CAs.
1675 OpenSSL still tries to build a complete chain to a root but if an
1676 intermediate CA has a trust setting included that is used. The first
1677 setting is used: whether to trust (e.g., -addtrust option to the x509
1681 *) Add -trusted_first option which attempts to find certificates in the
1682 trusted store even if an untrusted chain is also supplied.
1685 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1686 platform support for Linux and Android.
1689 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1692 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1693 When in FIPS mode the approved implementations are used as normal,
1694 when not in FIPS mode the internal unapproved versions are used instead.
1695 This means that the FIPS capable OpenSSL isn't forced to use the
1696 (often lower performance) FIPS implementations outside FIPS mode.
1699 *) Transparently support X9.42 DH parameters when calling
1700 PEM_read_bio_DHparameters. This means existing applications can handle
1701 the new parameter format automatically.
1704 *) Initial experimental support for X9.42 DH parameter format: mainly
1705 to support use of 'q' parameter for RFC5114 parameters.
1708 *) Add DH parameters from RFC5114 including test data to dhtest.
1711 *) Support for automatic EC temporary key parameter selection. If enabled
1712 the most preferred EC parameters are automatically used instead of
1713 hardcoded fixed parameters. Now a server just has to call:
1714 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1715 support ECDH and use the most appropriate parameters.
1718 *) Enhance and tidy EC curve and point format TLS extension code. Use
1719 static structures instead of allocation if default values are used.
1720 New ctrls to set curves we wish to support and to retrieve shared curves.
1721 Print out shared curves in s_server. New options to s_server and s_client
1722 to set list of supported curves.
1725 *) New ctrls to retrieve supported signature algorithms and
1726 supported curve values as an array of NIDs. Extend openssl utility
1727 to print out received values.
1730 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1731 between NIDs and the more common NIST names such as "P-256". Enhance
1732 ecparam utility and ECC method to recognise the NIST names for curves.
1735 *) Enhance SSL/TLS certificate chain handling to support different
1736 chains for each certificate instead of one chain in the parent SSL_CTX.
1739 *) Support for fixed DH ciphersuite client authentication: where both
1740 server and client use DH certificates with common parameters.
1743 *) Support for fixed DH ciphersuites: those requiring DH server
1747 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1749 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1750 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1751 X509_CINF_get_signature were reverted post internal team review.
1753 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1755 *) Build fixes for the Windows and OpenVMS platforms
1756 [Matt Caswell and Richard Levitte]
1758 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1760 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1761 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1762 dereference. This could lead to a Denial Of Service attack. Thanks to
1763 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1767 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1768 dtls1_buffer_record function under certain conditions. In particular this
1769 could occur if an attacker sent repeated DTLS records with the same
1770 sequence number but for the next epoch. The memory leak could be exploited
1771 by an attacker in a Denial of Service attack through memory exhaustion.
1772 Thanks to Chris Mueller for reporting this issue.
1776 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1777 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1778 method would be set to NULL which could later result in a NULL pointer
1779 dereference. Thanks to Frank Schmirler for reporting this issue.
1783 *) Abort handshake if server key exchange message is omitted for ephemeral
1786 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1787 reporting this issue.
1791 *) Remove non-export ephemeral RSA code on client and server. This code
1792 violated the TLS standard by allowing the use of temporary RSA keys in
1793 non-export ciphersuites and could be used by a server to effectively
1794 downgrade the RSA key length used to a value smaller than the server
1795 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1796 INRIA or reporting this issue.
1800 *) Fixed issue where DH client certificates are accepted without verification.
1801 An OpenSSL server will accept a DH certificate for client authentication
1802 without the certificate verify message. This effectively allows a client to
1803 authenticate without the use of a private key. This only affects servers
1804 which trust a client certificate authority which issues certificates
1805 containing DH keys: these are extremely rare and hardly ever encountered.
1806 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1811 *) Ensure that the session ID context of an SSL is updated when its
1812 SSL_CTX is updated via SSL_set_SSL_CTX.
1814 The session ID context is typically set from the parent SSL_CTX,
1815 and can vary with the CTX.
1818 *) Fix various certificate fingerprint issues.
1820 By using non-DER or invalid encodings outside the signed portion of a
1821 certificate the fingerprint can be changed without breaking the signature.
1822 Although no details of the signed portion of the certificate can be changed
1823 this can cause problems with some applications: e.g. those using the
1824 certificate fingerprint for blacklists.
1826 1. Reject signatures with non zero unused bits.
1828 If the BIT STRING containing the signature has non zero unused bits reject
1829 the signature. All current signature algorithms require zero unused bits.
1831 2. Check certificate algorithm consistency.
1833 Check the AlgorithmIdentifier inside TBS matches the one in the
1834 certificate signature. NB: this will result in signature failure
1835 errors for some broken certificates.
1837 Thanks to Konrad Kraszewski from Google for reporting this issue.
1839 3. Check DSA/ECDSA signatures use DER.
1841 Reencode DSA/ECDSA signatures and compare with the original received
1842 signature. Return an error if there is a mismatch.
1844 This will reject various cases including garbage after signature
1845 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1846 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1847 (negative or with leading zeroes).
1849 Further analysis was conducted and fixes were developed by Stephen Henson
1850 of the OpenSSL core team.
1855 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1856 results on some platforms, including x86_64. This bug occurs at random
1857 with a very low probability, and is not known to be exploitable in any
1858 way, though its exact impact is difficult to determine. Thanks to Pieter
1859 Wuille (Blockstream) who reported this issue and also suggested an initial
1860 fix. Further analysis was conducted by the OpenSSL development team and
1861 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1862 the OpenSSL core team.
1866 *) Do not resume sessions on the server if the negotiated protocol
1867 version does not match the session's version. Resuming with a different
1868 version, while not strictly forbidden by the RFC, is of questionable
1869 sanity and breaks all known clients.
1870 [David Benjamin, Emilia Käsper]
1872 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1873 early CCS messages during renegotiation. (Note that because
1874 renegotiation is encrypted, this early CCS was not exploitable.)
1877 *) Tighten client-side session ticket handling during renegotiation:
1878 ensure that the client only accepts a session ticket if the server sends
1879 the extension anew in the ServerHello. Previously, a TLS client would
1880 reuse the old extension state and thus accept a session ticket if one was
1881 announced in the initial ServerHello.
1883 Similarly, ensure that the client requires a session ticket if one
1884 was advertised in the ServerHello. Previously, a TLS client would
1885 ignore a missing NewSessionTicket message.
1888 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1890 *) SRTP Memory Leak.
1892 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1893 sends a carefully crafted handshake message, to cause OpenSSL to fail
1894 to free up to 64k of memory causing a memory leak. This could be
1895 exploited in a Denial Of Service attack. This issue affects OpenSSL
1896 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1897 whether SRTP is used or configured. Implementations of OpenSSL that
1898 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1900 The fix was developed by the OpenSSL team.
1904 *) Session Ticket Memory Leak.
1906 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1907 integrity of that ticket is first verified. In the event of a session
1908 ticket integrity check failing, OpenSSL will fail to free memory
1909 causing a memory leak. By sending a large number of invalid session
1910 tickets an attacker could exploit this issue in a Denial Of Service
1915 *) Build option no-ssl3 is incomplete.
1917 When OpenSSL is configured with "no-ssl3" as a build option, servers
1918 could accept and complete a SSL 3.0 handshake, and clients could be
1919 configured to send them.
1921 [Akamai and the OpenSSL team]
1923 *) Add support for TLS_FALLBACK_SCSV.
1924 Client applications doing fallback retries should call
1925 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1927 [Adam Langley, Bodo Moeller]
1929 *) Add additional DigestInfo checks.
1931 Reencode DigestInto in DER and check against the original when
1932 verifying RSA signature: this will reject any improperly encoded
1933 DigestInfo structures.
1935 Note: this is a precautionary measure and no attacks are currently known.
1939 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1941 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1942 SRP code can be overrun an internal buffer. Add sanity check that
1943 g, A, B < N to SRP code.
1945 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1946 Group for discovering this issue.
1950 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1951 TLS 1.0 instead of higher protocol versions when the ClientHello message
1952 is badly fragmented. This allows a man-in-the-middle attacker to force a
1953 downgrade to TLS 1.0 even if both the server and the client support a
1954 higher protocol version, by modifying the client's TLS records.
1956 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1957 researching this issue.
1961 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1962 to a denial of service attack. A malicious server can crash the client
1963 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1964 ciphersuite and sending carefully crafted handshake messages.
1966 Thanks to Felix Gröbert (Google) for discovering and researching this
1971 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1972 to leak memory. This can be exploited through a Denial of Service attack.
1973 Thanks to Adam Langley for discovering and researching this issue.
1977 *) An attacker can force openssl to consume large amounts of memory whilst
1978 processing DTLS handshake messages. This can be exploited through a
1979 Denial of Service attack.
1980 Thanks to Adam Langley for discovering and researching this issue.
1984 *) An attacker can force an error condition which causes openssl to crash
1985 whilst processing DTLS packets due to memory being freed twice. This
1986 can be exploited through a Denial of Service attack.
1987 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1992 *) If a multithreaded client connects to a malicious server using a resumed
1993 session and the server sends an ec point format extension it could write
1994 up to 255 bytes to freed memory.
1996 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2001 *) A malicious server can crash an OpenSSL client with a null pointer
2002 dereference (read) by specifying an SRP ciphersuite even though it was not
2003 properly negotiated with the client. This can be exploited through a
2004 Denial of Service attack.
2006 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2007 discovering and researching this issue.
2011 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2012 X509_name_oneline, X509_name_print_ex et al. to leak some information
2013 from the stack. Applications may be affected if they echo pretty printing
2014 output to the attacker.
2016 Thanks to Ivan Fratric (Google) for discovering this issue.
2018 [Emilia Käsper, and Steve Henson]
2020 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2021 for corner cases. (Certain input points at infinity could lead to
2022 bogus results, with non-infinity inputs mapped to infinity too.)
2025 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2027 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2028 handshake can force the use of weak keying material in OpenSSL
2029 SSL/TLS clients and servers.
2031 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2032 researching this issue. (CVE-2014-0224)
2033 [KIKUCHI Masashi, Steve Henson]
2035 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2036 OpenSSL DTLS client the code can be made to recurse eventually crashing
2039 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2041 [Imre Rad, Steve Henson]
2043 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2044 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2045 client or server. This is potentially exploitable to run arbitrary
2046 code on a vulnerable client or server.
2048 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2049 [Jüri Aedla, Steve Henson]
2051 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2052 are subject to a denial of service attack.
2054 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2055 this issue. (CVE-2014-3470)
2056 [Felix Gröbert, Ivan Fratric, Steve Henson]
2058 *) Harmonize version and its documentation. -f flag is used to display
2060 [mancha <mancha1@zoho.com>]
2062 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2063 in i2d_ECPrivateKey.
2064 [mancha <mancha1@zoho.com>]
2066 *) Fix some double frees. These are not thought to be exploitable.
2067 [mancha <mancha1@zoho.com>]
2069 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2071 *) A missing bounds check in the handling of the TLS heartbeat extension
2072 can be used to reveal up to 64k of memory to a connected client or
2075 Thanks for Neel Mehta of Google Security for discovering this bug and to
2076 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2077 preparing the fix (CVE-2014-0160)
2078 [Adam Langley, Bodo Moeller]
2080 *) Fix for the attack described in the paper "Recovering OpenSSL
2081 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2082 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2083 http://eprint.iacr.org/2014/140
2085 Thanks to Yuval Yarom and Naomi Benger for discovering this
2086 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2087 [Yuval Yarom and Naomi Benger]
2089 *) TLS pad extension: draft-agl-tls-padding-03
2091 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2092 TLS client Hello record length value would otherwise be > 255 and
2093 less that 512 pad with a dummy extension containing zeroes so it
2094 is at least 512 bytes long.
2096 [Adam Langley, Steve Henson]
2098 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2100 *) Fix for TLS record tampering bug. A carefully crafted invalid
2101 handshake could crash OpenSSL with a NULL pointer exception.
2102 Thanks to Anton Johansson for reporting this issues.
2105 *) Keep original DTLS digest and encryption contexts in retransmission
2106 structures so we can use the previous session parameters if they need
2107 to be resent. (CVE-2013-6450)
2110 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2111 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2112 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2113 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2114 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2115 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2116 [Rob Stradling, Adam Langley]
2118 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2120 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2121 supporting platforms or when small records were transferred.
2122 [Andy Polyakov, Steve Henson]
2124 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2126 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2128 This addresses the flaw in CBC record processing discovered by
2129 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2130 at: http://www.isg.rhul.ac.uk/tls/
2132 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2133 Security Group at Royal Holloway, University of London
2134 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2135 Emilia Käsper for the initial patch.
2137 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2139 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2140 ciphersuites which can be exploited in a denial of service attack.
2141 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2142 and detecting this bug and to Wolfgang Ettlinger
2143 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2147 *) Return an error when checking OCSP signatures when key is NULL.
2148 This fixes a DoS attack. (CVE-2013-0166)
2151 *) Make openssl verify return errors.
2152 [Chris Palmer <palmer@google.com> and Ben Laurie]
2154 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2155 the right response is stapled. Also change SSL_get_certificate()
2156 so it returns the certificate actually sent.
2157 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2158 [Rob Stradling <rob.stradling@comodo.com>]
2160 *) Fix possible deadlock when decoding public keys.
2163 *) Don't use TLS 1.0 record version number in initial client hello
2167 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2169 *) Sanity check record length before skipping explicit IV in TLS
2170 1.2, 1.1 and DTLS to fix DoS attack.
2172 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2173 fuzzing as a service testing platform.
2177 *) Initialise tkeylen properly when encrypting CMS messages.
2178 Thanks to Solar Designer of Openwall for reporting this issue.
2181 *) In FIPS mode don't try to use composite ciphers as they are not
2185 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2187 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2188 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2189 mean any application compiled against OpenSSL 1.0.0 headers setting
2190 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2191 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2192 0x10000000L Any application which was previously compiled against
2193 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2194 will need to be recompiled as a result. Letting be results in
2195 inability to disable specifically TLS 1.1 and in client context,
2196 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2199 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2200 disable just protocol X, but all protocols above X *if* there are
2201 protocols *below* X still enabled. In more practical terms it means
2202 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2203 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2204 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2208 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2210 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2211 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2212 in CRYPTO_realloc_clean.
2214 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2215 issue and to Adam Langley <agl@chromium.org> for fixing it.
2217 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2219 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2222 *) Workarounds for some broken servers that "hang" if a client hello
2223 record length exceeds 255 bytes.
2225 1. Do not use record version number > TLS 1.0 in initial client
2226 hello: some (but not all) hanging servers will now work.
2227 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2228 the number of ciphers sent in the client hello. This should be
2229 set to an even number, such as 50, for example by passing:
2230 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2231 Most broken servers should now work.
2232 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2233 TLS 1.2 client support entirely.
2236 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2239 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2241 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2242 STRING form instead of a DigestInfo.
2245 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2246 and the RSA_sign/RSA_verify functions. This was made more apparent when
2247 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2248 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2249 the correct format in RSA_verify so both forms transparently work.
2252 *) Some servers which support TLS 1.0 can choke if we initially indicate
2253 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2254 encrypted premaster secret. As a workaround use the maximum permitted
2255 client version in client hello, this should keep such servers happy
2256 and still work with previous versions of OpenSSL.
2259 *) Add support for TLS/DTLS heartbeats.
2260 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2262 *) Add support for SCTP.
2263 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2265 *) Improved PRNG seeding for VOS.
2266 [Paul Green <Paul.Green@stratus.com>]
2268 *) Extensive assembler packs updates, most notably:
2270 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2271 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2272 - x86_64: bit-sliced AES implementation;
2273 - ARM: NEON support, contemporary platforms optimizations;
2274 - s390x: z196 support;
2275 - *: GHASH and GF(2^m) multiplication implementations;
2279 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2280 (removal of unnecessary code)
2281 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2283 *) Add TLS key material exporter from RFC 5705.
2286 *) Add DTLS-SRTP negotiation from RFC 5764.
2289 *) Add Next Protocol Negotiation,
2290 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2291 disabled with a no-npn flag to config or Configure. Code donated
2293 [Adam Langley <agl@google.com> and Ben Laurie]
2295 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2296 NIST-P256, NIST-P521, with constant-time single point multiplication on
2297 typical inputs. Compiler support for the nonstandard type __uint128_t is
2298 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2299 Code made available under Apache License version 2.0.
2301 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2302 line to include this in your build of OpenSSL, and run "make depend" (or
2303 "make update"). This enables the following EC_METHODs:
2305 EC_GFp_nistp224_method()
2306 EC_GFp_nistp256_method()
2307 EC_GFp_nistp521_method()
2309 EC_GROUP_new_by_curve_name() will automatically use these (while
2310 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2312 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2314 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2315 all platforms. Move ssize_t definition from e_os.h to the public
2316 header file e_os2.h as it now appears in public header file cms.h
2319 *) New -sigopt option to the ca, req and x509 utilities. Additional
2320 signature parameters can be passed using this option and in
2324 *) Add RSA PSS signing function. This will generate and set the
2325 appropriate AlgorithmIdentifiers for PSS based on those in the
2326 corresponding EVP_MD_CTX structure. No application support yet.
2329 *) Support for companion algorithm specific ASN1 signing routines.
2330 New function ASN1_item_sign_ctx() signs a pre-initialised
2331 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2332 the appropriate parameters.
2335 *) Add new algorithm specific ASN1 verification initialisation function
2336 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2337 handling will be the same no matter what EVP_PKEY_METHOD is used.
2338 Add a PSS handler to support verification of PSS signatures: checked
2339 against a number of sample certificates.
2342 *) Add signature printing for PSS. Add PSS OIDs.
2343 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2345 *) Add algorithm specific signature printing. An individual ASN1 method
2346 can now print out signatures instead of the standard hex dump.
2348 More complex signatures (e.g. PSS) can print out more meaningful
2349 information. Include DSA version that prints out the signature
2353 *) Password based recipient info support for CMS library: implementing
2357 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2358 neatly separates the code into cipher and PBE sections and is required
2359 for some algorithms that split PBES2 into separate pieces (such as
2360 password based CMS).
2363 *) Session-handling fixes:
2364 - Fix handling of connections that are resuming with a session ID,
2365 but also support Session Tickets.
2366 - Fix a bug that suppressed issuing of a new ticket if the client
2367 presented a ticket with an expired session.
2368 - Try to set the ticket lifetime hint to something reasonable.
2369 - Make tickets shorter by excluding irrelevant information.
2370 - On the client side, don't ignore renewed tickets.
2371 [Adam Langley, Bodo Moeller (Google)]
2373 *) Fix PSK session representation.
2376 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2378 This work was sponsored by Intel.
2381 *) Add GCM support to TLS library. Some custom code is needed to split
2382 the IV between the fixed (from PRF) and explicit (from TLS record)
2383 portions. This adds all GCM ciphersuites supported by RFC5288 and
2384 RFC5289. Generalise some AES* cipherstrings to include GCM and
2385 add a special AESGCM string for GCM only.
2388 *) Expand range of ctrls for AES GCM. Permit setting invocation
2389 field on decrypt and retrieval of invocation field only on encrypt.
2392 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2393 As required by RFC5289 these ciphersuites cannot be used if for
2394 versions of TLS earlier than 1.2.
2397 *) For FIPS capable OpenSSL interpret a NULL default public key method
2398 as unset and return the appropriate default but do *not* set the default.
2399 This means we can return the appropriate method in applications that
2400 switch between FIPS and non-FIPS modes.
2403 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2404 ENGINE is used then we cannot handle that in the FIPS module so we
2405 keep original code iff non-FIPS operations are allowed.
2408 *) Add -attime option to openssl utilities.
2409 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2411 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2414 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2415 FIPS EC methods unconditionally for now.
2418 *) New build option no-ec2m to disable characteristic 2 code.
2421 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2422 all cases can be covered as some introduce binary incompatibilities.
2425 *) Redirect RSA operations to FIPS module including keygen,
2426 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2429 *) Add similar low level API blocking to ciphers.
2432 *) Low level digest APIs are not approved in FIPS mode: any attempt
2433 to use these will cause a fatal error. Applications that *really* want
2434 to use them can use the private_* version instead.
2437 *) Redirect cipher operations to FIPS module for FIPS builds.
2440 *) Redirect digest operations to FIPS module for FIPS builds.
2443 *) Update build system to add "fips" flag which will link in fipscanister.o
2444 for static and shared library builds embedding a signature if needed.
2447 *) Output TLS supported curves in preference order instead of numerical
2448 order. This is currently hardcoded for the highest order curves first.
2449 This should be configurable so applications can judge speed vs strength.
2452 *) Add TLS v1.2 server support for client authentication.
2455 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2459 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2460 FIPS modules versions.
2463 *) Add TLS v1.2 client side support for client authentication. Keep cache
2464 of handshake records longer as we don't know the hash algorithm to use
2465 until after the certificate request message is received.
2468 *) Initial TLS v1.2 client support. Add a default signature algorithms
2469 extension including all the algorithms we support. Parse new signature
2470 format in client key exchange. Relax some ECC signing restrictions for
2471 TLS v1.2 as indicated in RFC5246.
2474 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2475 to new signature format when needed using client digest preference.
2476 All server ciphersuites should now work correctly in TLS v1.2. No client
2477 support yet and no support for client certificates.
2480 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2481 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2482 ciphersuites. At present only RSA key exchange ciphersuites work with
2483 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2484 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2485 and version checking.
2488 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2489 with this defined it will not be affected by any changes to ssl internal
2490 structures. Add several utility functions to allow openssl application
2491 to work with OPENSSL_NO_SSL_INTERN defined.
2495 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2497 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2500 *) Permit abbreviated handshakes when renegotiating using the function
2501 SSL_renegotiate_abbreviated().
2502 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2504 *) Add call to ENGINE_register_all_complete() to
2505 ENGINE_load_builtin_engines(), so some implementations get used
2506 automatically instead of needing explicit application support.
2509 *) Add support for TLS key exporter as described in RFC5705.
2510 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2512 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2513 a few changes are required:
2515 Add SSL_OP_NO_TLSv1_1 flag.
2516 Add TLSv1_1 methods.
2517 Update version checking logic to handle version 1.1.
2518 Add explicit IV handling (ported from DTLS code).
2519 Add command line options to s_client/s_server.
2522 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2524 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2525 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2526 content decryption and always return the same error. Note: this attack
2527 needs on average 2^20 messages so it only affects automated senders. The
2528 old behaviour can be reenabled in the CMS code by setting the
2529 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2530 an MMA defence is not necessary.
2531 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2532 this issue. (CVE-2012-0884)
2535 *) Fix CVE-2011-4619: make sure we really are receiving a
2536 client hello before rejecting multiple SGC restarts. Thanks to
2537 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2540 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2542 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2543 Thanks to Antonio Martin, Enterprise Secure Access Research and
2544 Development, Cisco Systems, Inc. for discovering this bug and
2545 preparing a fix. (CVE-2012-0050)
2548 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2550 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2551 of the Vaudenay padding oracle attack on CBC mode encryption
2552 which enables an efficient plaintext recovery attack against
2553 the OpenSSL implementation of DTLS. Their attack exploits timing
2554 differences arising during decryption processing. A research
2555 paper describing this attack can be found at:
2556 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2557 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2558 Security Group at Royal Holloway, University of London
2559 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2560 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2561 for preparing the fix. (CVE-2011-4108)
2562 [Robin Seggelmann, Michael Tuexen]
2564 *) Clear bytes used for block padding of SSL 3.0 records.
2566 [Adam Langley (Google)]
2568 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2569 Kadianakis <desnacked@gmail.com> for discovering this issue and
2570 Adam Langley for preparing the fix. (CVE-2011-4619)
2571 [Adam Langley (Google)]
2573 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2574 [Andrey Kulikov <amdeich@gmail.com>]
2576 *) Prevent malformed RFC3779 data triggering an assertion failure.
2577 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2578 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2579 [Rob Austein <sra@hactrn.net>]
2581 *) Improved PRNG seeding for VOS.
2582 [Paul Green <Paul.Green@stratus.com>]
2584 *) Fix ssl_ciph.c set-up race.
2585 [Adam Langley (Google)]
2587 *) Fix spurious failures in ecdsatest.c.
2588 [Emilia Käsper (Google)]
2590 *) Fix the BIO_f_buffer() implementation (which was mixing different
2591 interpretations of the '..._len' fields).
2592 [Adam Langley (Google)]
2594 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2595 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2596 threads won't reuse the same blinding coefficients.
2598 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2599 lock to call BN_BLINDING_invert_ex, and avoids one use of
2600 BN_BLINDING_update for each BN_BLINDING structure (previously,
2601 the last update always remained unused).
2602 [Emilia Käsper (Google)]
2604 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2605 [Bob Buckholz (Google)]
2607 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2609 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2610 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2611 [Kaspar Brand <ossl@velox.ch>]
2613 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2614 for multi-threaded use of ECDH. (CVE-2011-3210)
2615 [Adam Langley (Google)]
2617 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2620 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2621 signature public key algorithm by using OID xref utilities instead.
2622 Before this you could only use some ECC ciphersuites with SHA1 only.
2625 *) Add protection against ECDSA timing attacks as mentioned in the paper
2626 by Billy Bob Brumley and Nicola Tuveri, see:
2628 http://eprint.iacr.org/2011/232.pdf
2630 [Billy Bob Brumley and Nicola Tuveri]
2632 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2634 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2635 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2637 *) Fix bug in string printing code: if *any* escaping is enabled we must
2638 escape the escape character (backslash) or the resulting string is
2642 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2644 *) Disable code workaround for ancient and obsolete Netscape browsers
2645 and servers: an attacker can use it in a ciphersuite downgrade attack.
2646 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2649 *) Fixed J-PAKE implementation error, originally discovered by
2650 Sebastien Martini, further info and confirmation from Stefan
2651 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2654 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2656 *) Fix extension code to avoid race conditions which can result in a buffer
2657 overrun vulnerability: resumed sessions must not be modified as they can
2658 be shared by multiple threads. CVE-2010-3864
2661 *) Fix WIN32 build system to correctly link an ENGINE directory into
2665 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2667 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2669 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2671 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2673 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2674 context. The operation can be customised via the ctrl mechanism in
2675 case ENGINEs want to include additional functionality.
2678 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2681 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2682 output hashes compatible with older versions of OpenSSL.
2683 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2685 *) Fix compression algorithm handling: if resuming a session use the
2686 compression algorithm of the resumed session instead of determining
2687 it from client hello again. Don't allow server to change algorithm.
2690 *) Add load_crls() function to apps tidying load_certs() too. Add option
2691 to verify utility to allow additional CRLs to be included.
2694 *) Update OCSP request code to permit adding custom headers to the request:
2695 some responders need this.
2698 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2700 [Julia Lawall <julia@diku.dk>]
2702 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2703 needlessly dereferenced structures, used obsolete functions and
2704 didn't handle all updated verify codes correctly.
2707 *) Disable MD2 in the default configuration.
2710 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2711 indicate the initial BIO being pushed or popped. This makes it possible
2712 to determine whether the BIO is the one explicitly called or as a result
2713 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2714 it handles reference counts correctly and doesn't zero out the I/O bio
2715 when it is not being explicitly popped. WARNING: applications which
2716 included workarounds for the old buggy behaviour will need to be modified
2717 or they could free up already freed BIOs.
2720 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2721 renaming to all platforms (within the 0.9.8 branch, this was
2722 done conditionally on Netware platforms to avoid a name clash).
2723 [Guenter <lists@gknw.net>]
2725 *) Add ECDHE and PSK support to DTLS.
2726 [Michael Tuexen <tuexen@fh-muenster.de>]
2728 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2732 *) Add "missing" function EVP_MD_flags() (without this the only way to
2733 retrieve a digest flags is by accessing the structure directly. Update
2734 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2735 or cipher is registered as in the "from" argument. Print out all
2736 registered digests in the dgst usage message instead of manually
2737 attempting to work them out.
2740 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2741 this allows the use of compression and extensions. Change default cipher
2742 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2743 by default unless an application cipher string requests it.
2746 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2747 key ids to find matching certificates and keys but some PKCS#12 files
2748 don't follow the (somewhat unwritten) rules and this strategy fails.
2749 Now just gather all certificates together and the first private key
2750 then look for the first certificate that matches the key.
2753 *) Support use of registered digest and cipher names for dgst and cipher
2754 commands instead of having to add each one as a special case. So now
2761 openssl dgst -sha256 foo
2763 and this works for ENGINE based algorithms too.
2767 *) Update Gost ENGINE to support parameter files.
2768 [Victor B. Wagner <vitus@cryptocom.ru>]
2770 *) Support GeneralizedTime in ca utility.
2771 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2773 *) Enhance the hash format used for certificate directory links. The new
2774 form uses the canonical encoding (meaning equivalent names will work
2775 even if they aren't identical) and uses SHA1 instead of MD5. This form
2776 is incompatible with the older format and as a result c_rehash should
2777 be used to rebuild symbolic links.
2780 *) Make PKCS#8 the default write format for private keys, replacing the
2781 traditional format. This form is standardised, more secure and doesn't
2782 include an implicit MD5 dependency.
2785 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2786 committed to OpenSSL should pass this lot as a minimum.
2789 *) Add session ticket override functionality for use by EAP-FAST.
2790 [Jouni Malinen <j@w1.fi>]
2792 *) Modify HMAC functions to return a value. Since these can be implemented
2793 in an ENGINE errors can occur.
2796 *) Type-checked OBJ_bsearch_ex.
2799 *) Type-checked OBJ_bsearch. Also some constification necessitated
2800 by type-checking. Still to come: TXT_DB, bsearch(?),
2801 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2805 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2806 seconds to a tm structure directly, instead of going through OS
2807 specific date routines. This avoids any issues with OS routines such
2808 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2809 and X509_time_adj_ex() to cover the extended range. The existing
2810 X509_time_adj() is still usable and will no longer have any date issues.
2813 *) Delta CRL support. New use deltas option which will attempt to locate
2814 and search any appropriate delta CRLs available.
2816 This work was sponsored by Google.
2819 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2820 code and add additional score elements. Validate alternate CRL paths
2821 as part of the CRL checking and indicate a new error "CRL path validation
2822 error" in this case. Applications wanting additional details can use
2823 the verify callback and check the new "parent" field. If this is not
2824 NULL CRL path validation is taking place. Existing applications wont
2825 see this because it requires extended CRL support which is off by
2828 This work was sponsored by Google.
2831 *) Support for freshest CRL extension.
2833 This work was sponsored by Google.
2836 *) Initial indirect CRL support. Currently only supported in the CRLs
2837 passed directly and not via lookup. Process certificate issuer
2838 CRL entry extension and lookup CRL entries by bother issuer name
2839 and serial number. Check and process CRL issuer entry in IDP extension.
2841 This work was sponsored by Google.
2844 *) Add support for distinct certificate and CRL paths. The CRL issuer
2845 certificate is validated separately in this case. Only enabled if
2846 an extended CRL support flag is set: this flag will enable additional
2847 CRL functionality in future.
2849 This work was sponsored by Google.
2852 *) Add support for policy mappings extension.
2854 This work was sponsored by Google.
2857 *) Fixes to pathlength constraint, self issued certificate handling,
2858 policy processing to align with RFC3280 and PKITS tests.
2860 This work was sponsored by Google.
2863 *) Support for name constraints certificate extension. DN, email, DNS
2864 and URI types are currently supported.
2866 This work was sponsored by Google.
2869 *) To cater for systems that provide a pointer-based thread ID rather
2870 than numeric, deprecate the current numeric thread ID mechanism and
2871 replace it with a structure and associated callback type. This
2872 mechanism allows a numeric "hash" to be extracted from a thread ID in
2873 either case, and on platforms where pointers are larger than 'long',
2874 mixing is done to help ensure the numeric 'hash' is usable even if it
2875 can't be guaranteed unique. The default mechanism is to use "&errno"
2876 as a pointer-based thread ID to distinguish between threads.
2878 Applications that want to provide their own thread IDs should now use
2879 CRYPTO_THREADID_set_callback() to register a callback that will call
2880 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2882 Note that ERR_remove_state() is now deprecated, because it is tied
2883 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2884 to free the current thread's error state should be replaced by
2885 ERR_remove_thread_state(NULL).
2887 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2888 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2889 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2890 application was previously providing a numeric thread callback that
2891 was inappropriate for distinguishing threads, then uniqueness might
2892 have been obtained with &errno that happened immediately in the
2893 intermediate development versions of OpenSSL; this is no longer the
2894 case, the numeric thread callback will now override the automatic use
2896 [Geoff Thorpe, with help from Bodo Moeller]
2898 *) Initial support for different CRL issuing certificates. This covers a
2899 simple case where the self issued certificates in the chain exist and
2900 the real CRL issuer is higher in the existing chain.
2902 This work was sponsored by Google.
2905 *) Removed effectively defunct crypto/store from the build.
2908 *) Revamp of STACK to provide stronger type-checking. Still to come:
2909 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2910 ASN1_STRING, CONF_VALUE.
2913 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2914 RAM on SSL connections. This option can save about 34k per idle SSL.
2917 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2918 STACK, TXT_DB, bsearch, qsort.
2921 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2922 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2923 support for data, signedData, compressedData, digestedData and
2924 encryptedData, envelopedData types included. Scripts to check against
2925 RFC4134 examples draft and interop and consistency checks of many
2926 content types and variants.
2929 *) Add options to enc utility to support use of zlib compression BIO.
2932 *) Extend mk1mf to support importing of options and assembly language
2933 files from Configure script, currently only included in VC-WIN32.
2934 The assembly language rules can now optionally generate the source
2935 files from the associated perl scripts.
2938 *) Implement remaining functionality needed to support GOST ciphersuites.
2939 Interop testing has been performed using CryptoPro implementations.
2940 [Victor B. Wagner <vitus@cryptocom.ru>]
2942 *) s390x assembler pack.
2945 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2949 *) Implement Opaque PRF Input TLS extension as specified in
2950 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2951 official specification yet and no extension type assignment by
2952 IANA exists, this extension (for now) will have to be explicitly
2953 enabled when building OpenSSL by providing the extension number
2954 to use. For example, specify an option
2956 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2958 to the "config" or "Configure" script to enable the extension,
2959 assuming extension number 0x9527 (which is a completely arbitrary
2960 and unofficial assignment based on the MD5 hash of the Internet
2961 Draft). Note that by doing so, you potentially lose
2962 interoperability with other TLS implementations since these might
2963 be using the same extension number for other purposes.
2965 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2966 opaque PRF input value to use in the handshake. This will create
2967 an interal copy of the length-'len' string at 'src', and will
2968 return non-zero for success.
2970 To get more control and flexibility, provide a callback function
2973 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2974 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2978 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2981 Callback function 'cb' will be called in handshakes, and is
2982 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2983 Argument 'arg' is for application purposes (the value as given to
2984 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2985 be provided to the callback function). The callback function
2986 has to return non-zero to report success: usually 1 to use opaque
2987 PRF input just if possible, or 2 to enforce use of the opaque PRF
2988 input. In the latter case, the library will abort the handshake
2989 if opaque PRF input is not successfully negotiated.
2991 Arguments 'peerinput' and 'len' given to the callback function
2992 will always be NULL and 0 in the case of a client. A server will
2993 see the client's opaque PRF input through these variables if
2994 available (NULL and 0 otherwise). Note that if the server
2995 provides an opaque PRF input, the length must be the same as the
2996 length of the client's opaque PRF input.
2998 Note that the callback function will only be called when creating
2999 a new session (session resumption can resume whatever was
3000 previously negotiated), and will not be called in SSL 2.0
3001 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3002 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3003 for applications that need to enforce opaque PRF input.
3007 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3010 [Victor B. Wagner <vitus@cryptocom.ru>]
3012 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3013 RFC4507bis. The encrypted ticket format is an encrypted encoded
3014 SSL_SESSION structure, that way new session features are automatically
3017 If a client application caches session in an SSL_SESSION structure
3018 support is transparent because tickets are now stored in the encoded
3021 The SSL_CTX structure automatically generates keys for ticket
3022 protection in servers so again support should be possible
3023 with no application modification.
3025 If a client or server wishes to disable RFC4507 support then the option
3026 SSL_OP_NO_TICKET can be set.
3028 Add a TLS extension debugging callback to allow the contents of any client
3029 or server extensions to be examined.
3031 This work was sponsored by Google.
3034 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3035 OpenSSL should now compile cleanly on gcc 4.2
3036 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3038 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3039 support including streaming MAC support: this is required for GOST
3040 ciphersuite support.
3041 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3043 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3044 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3045 to output in BER and PEM format.
3048 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3049 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3050 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3051 ENGINE support for HMAC keys which are unextractable. New -mac and
3052 -macopt options to dgst utility.
3055 *) New option -sigopt to dgst utility. Update dgst to use
3056 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3057 alternative signing parameters such as X9.31 or PSS in the dgst
3061 *) Change ssl_cipher_apply_rule(), the internal function that does
3062 the work each time a ciphersuite string requests enabling
3063 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3064 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3065 the order of disabled ciphersuites such that those ciphersuites
3066 that most recently went from enabled to disabled not only stay
3067 in order with respect to each other, but also have higher priority
3068 than other disabled ciphersuites the next time ciphersuites are
3071 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3072 the same ciphersuites as with "HIGH" alone, but in a specific
3073 order where the PSK ciphersuites come first (since they are the
3074 most recently disabled ciphersuites when "HIGH" is parsed).
3076 Also, change ssl_create_cipher_list() (using this new
3077 funcionality) such that between otherwise identical
3078 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3082 *) Change ssl_create_cipher_list() so that it automatically
3083 arranges the ciphersuites in reasonable order before starting
3084 to process the rule string. Thus, the definition for "DEFAULT"
3085 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3086 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3087 This makes it much easier to arrive at a reasonable default order
3088 in applications for which anonymous ciphers are OK (meaning
3089 that you can't actually use DEFAULT).
3090 [Bodo Moeller; suggested by Victor Duchovni]
3092 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3093 processing) into multiple integers instead of setting
3094 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3095 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3096 (These masks as well as the individual bit definitions are hidden
3097 away into the non-exported interface ssl/ssl_locl.h, so this
3098 change to the definition of the SSL_CIPHER structure shouldn't
3099 affect applications.) This give us more bits for each of these
3100 categories, so there is no longer a need to coagulate AES128 and
3101 AES256 into a single algorithm bit, and to coagulate Camellia128
3102 and Camellia256 into a single algorithm bit, which has led to all
3105 Thus, among other things, the kludge introduced in 0.9.7m and
3106 0.9.8e for masking out AES256 independently of AES128 or masking
3107 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3109 With the change, we also introduce new ciphersuite aliases that
3110 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3114 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3115 Use the leftmost N bytes of the signature input if the input is
3116 larger than the prime q (with N being the size in bytes of q).
3119 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3120 it yet and it is largely untested.
3123 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3126 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3127 some compilers (gcc 4.2 and later) reject their use. Safestack is
3128 reimplemented. Update ASN1 to avoid use of legacy functions.
3131 *) Win32/64 targets are linked with Winsock2.
3134 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3135 to external functions. This can be used to increase CRL handling
3136 efficiency especially when CRLs are very large by (for example) storing
3137 the CRL revoked certificates in a database.
3140 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3141 new CRLs added to a directory can be used. New command line option
3142 -verify_return_error to s_client and s_server. This causes real errors
3143 to be returned by the verify callback instead of carrying on no matter
3144 what. This reflects the way a "real world" verify callback would behave.
3147 *) GOST engine, supporting several GOST algorithms and public key formats.
3148 Kindly donated by Cryptocom.
3151 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3152 partitioned by DP are handled but no indirect CRL or reason partitioning
3153 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3154 selected via a scoring technique which handles IDP and AKID in CRLs.
3157 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3158 will ultimately be used for all verify operations: this will remove the
3159 X509_STORE dependency on certificate verification and allow alternative
3160 lookup methods. X509_STORE based implementations of these two callbacks.
3163 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3164 Modify get_crl() to find a valid (unexpired) CRL if possible.
3167 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3168 this would be called X509_CRL_cmp() but that name is already used by
3169 a function that just compares CRL issuer names. Cache several CRL
3170 extensions in X509_CRL structure and cache CRLDP in X509.
3173 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3174 this maps equivalent X509_NAME structures into a consistent structure.
3175 Name comparison can then be performed rapidly using memcmp().
3178 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3182 *) Allow digests to supply their own micalg string for S/MIME type using
3183 the ctrl EVP_MD_CTRL_MICALG.
3186 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3187 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3188 ctrl. It can then customise the structure before and/or after signing
3192 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3193 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3194 to free up any added signature OIDs.
3197 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3198 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3199 digest and cipher tables. New options added to openssl utility:
3200 list-message-digest-algorithms and list-cipher-algorithms.
3203 *) Change the array representation of binary polynomials: the list
3204 of degrees of non-zero coefficients is now terminated with -1.
3205 Previously it was terminated with 0, which was also part of the
3206 value; thus, the array representation was not applicable to
3207 polynomials where t^0 has coefficient zero. This change makes
3208 the array representation useful in a more general context.
3211 *) Various modifications and fixes to SSL/TLS cipher string
3212 handling. For ECC, the code now distinguishes between fixed ECDH
3213 with RSA certificates on the one hand and with ECDSA certificates
3214 on the other hand, since these are separate ciphersuites. The
3215 unused code for Fortezza ciphersuites has been removed.
3217 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3218 (not "ECDHE"). For consistency with the code for DH
3219 certificates, use of ECDH certificates is now considered ECDH
3220 authentication, not RSA or ECDSA authentication (the latter is
3221 merely the CA's signing algorithm and not actively used in the
3224 The temporary ciphersuite alias "ECCdraft" is no longer
3225 available, and ECC ciphersuites are no longer excluded from "ALL"
3226 and "DEFAULT". The following aliases now exist for RFC 4492
3227 ciphersuites, most of these by analogy with the DH case:
3229 kECDHr - ECDH cert, signed with RSA
3230 kECDHe - ECDH cert, signed with ECDSA
3231 kECDH - ECDH cert (signed with either RSA or ECDSA)
3232 kEECDH - ephemeral ECDH
3233 ECDH - ECDH cert or ephemeral ECDH
3239 AECDH - anonymous ECDH
3240 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3244 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3245 Use correct micalg parameters depending on digest(s) in signed message.
3248 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3249 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3252 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3253 an engine to register a method. Add ENGINE lookups for methods and
3254 functional reference processing.
3257 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3258 EVP_{Sign,Verify}* which allow an application to customise the signature
3262 *) New -resign option to smime utility. This adds one or more signers
3263 to an existing PKCS#7 signedData structure. Also -md option to use an
3264 alternative message digest algorithm for signing.
3267 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3268 create PKCS7 structures containing multiple signers. Update smime
3269 application to support multiple signers.
3272 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3276 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3277 Reorganize PBE internals to lookup from a static table using NIDs,
3278 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3279 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3280 PRF which will be automatically used with PBES2.
3283 *) Replace the algorithm specific calls to generate keys in "req" with the
3287 *) Update PKCS#7 enveloped data routines to use new API. This is now
3288 supported by any public key method supporting the encrypt operation. A
3289 ctrl is added to allow the public key algorithm to examine or modify
3290 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3294 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3295 a default digest type to use. In most cases this will be SHA1 but some
3296 algorithms (such as GOST) need to specify an alternative digest. The
3297 return value indicates how strong the preference is 1 means optional and
3298 2 is mandatory (that is it is the only supported type). Modify
3299 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3300 use the default md. Update openssl utilities to use the default digest
3301 type for signing if it is not explicitly indicated.
3304 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3305 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3306 signing method from the key type. This effectively removes the link
3307 between digests and public key types.
3310 *) Add an OID cross reference table and utility functions. Its purpose is to
3311 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3312 rsaEncryption. This will allow some of the algorithm specific hackery
3313 needed to use the correct OID to be removed.
3316 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3317 structures for PKCS7_sign(). They are now set up by the relevant public
3321 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3324 *) Add support for key derivation (agreement) in the API, DH method and
3328 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3329 public and private key formats. As a side effect these add additional
3330 command line functionality not previously available: DSA signatures can be
3331 generated and verified using pkeyutl and DH key support and generation in
3336 [Oliver Tappe <zooey@hirschkaefer.de>]
3338 *) New make target "install_html_docs" installs HTML renditions of the
3340 [Oliver Tappe <zooey@hirschkaefer.de>]
3342 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3343 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3344 support key and parameter generation and add initial key generation
3345 functionality for RSA.
3348 *) Add functions for main EVP_PKEY_method operations. The undocumented
3349 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3350 EVP_PKEY_{encrypt,decrypt}_old.
3353 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3354 key API, doesn't do much yet.
3357 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3358 public key algorithms. New option to openssl utility:
3359 "list-public-key-algorithms" to print out info.
3362 *) Implement the Supported Elliptic Curves Extension for
3363 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3366 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3367 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3370 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3371 utilities such as rsa, dsa, dsaparam etc except they process any key
3375 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3376 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3377 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3381 *) Initial support for pluggable public key ASN1.
3382 De-spaghettify the public key ASN1 handling. Move public and private
3383 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3384 algorithm specific handling to a single module within the relevant
3385 algorithm directory. Add functions to allow (near) opaque processing
3386 of public and private key structures.
3389 *) Implement the Supported Point Formats Extension for
3390 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3393 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3394 for the psk identity [hint] and the psk callback functions to the
3395 SSL_SESSION, SSL and SSL_CTX structure.
3398 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3402 SSL_CTX_use_psk_identity_hint
3403 SSL_get_psk_identity_hint
3404 SSL_get_psk_identity
3405 SSL_use_psk_identity_hint
3407 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3409 *) Add RFC 3161 compliant time stamp request creation, response generation
3410 and response verification functionality.
3411 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3413 *) Add initial support for TLS extensions, specifically for the server_name
3414 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3415 have new members for a host name. The SSL data structure has an
3416 additional member SSL_CTX *initial_ctx so that new sessions can be
3417 stored in that context to allow for session resumption, even after the
3418 SSL has been switched to a new SSL_CTX in reaction to a client's
3419 server_name extension.
3421 New functions (subject to change):
3423 SSL_get_servername()
3424 SSL_get_servername_type()
3427 New CTRL codes and macros (subject to change):
3429 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3430 - SSL_CTX_set_tlsext_servername_callback()
3431 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3432 - SSL_CTX_set_tlsext_servername_arg()
3433 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3435 openssl s_client has a new '-servername ...' option.
3437 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3438 '-key2 ...', '-servername_fatal' (subject to change). This allows
3439 testing the HostName extension for a specific single host name ('-cert'
3440 and '-key' remain fallbacks for handshakes without HostName
3441 negotiation). If the unrecognized_name alert has to be sent, this by
3442 default is a warning; it becomes fatal with the '-servername_fatal'
3445 [Peter Sylvester, Remy Allais, Christophe Renou]
3447 *) Whirlpool hash implementation is added.
3450 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3451 bn(64,32). Because of instruction set limitations it doesn't have
3452 any negative impact on performance. This was done mostly in order
3453 to make it possible to share assembler modules, such as bn_mul_mont
3454 implementations, between 32- and 64-bit builds without hassle.
3457 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3458 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3462 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3463 dedicated Montgomery multiplication procedure, is introduced.
3464 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3465 "64-bit" performance on certain 32-bit targets.
3468 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3469 in SSL structures. New SSL ctrl to set maximum send fragment size.
3470 Save memory by seeting the I/O buffer sizes dynamically instead of
3471 using the maximum available value.
3474 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3475 in addition to the text details.
3478 *) Very, very preliminary EXPERIMENTAL support for printing of general
3479 ASN1 structures. This currently produces rather ugly output and doesn't
3480 handle several customised structures at all.
3483 *) Integrated support for PVK file format and some related formats such
3484 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3485 these in the 'rsa' and 'dsa' utilities.
3488 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3491 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3492 place for the (very old) "NETSCAPE" format certificates which are now
3493 handled using new ASN1 code equivalents.
3496 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3497 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3498 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3501 *) Modify CRL distribution points extension code to print out previously
3502 unsupported fields. Enhance extension setting code to allow setting of
3506 *) Add print and set support for Issuing Distribution Point CRL extension.
3509 *) Change 'Configure' script to enable Camellia by default.
3512 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3514 *) When rejecting SSL/TLS records due to an incorrect version number, never
3515 update s->server with a new major version number. As of
3516 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3517 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3518 the previous behavior could result in a read attempt at NULL when
3519 receiving specific incorrect SSL/TLS records once record payload
3520 protection is active. (CVE-2010-0740)
3521 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3523 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3524 could be crashed if the relevant tables were not present (e.g. chrooted).
3525 [Tomas Hoger <thoger@redhat.com>]
3527 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3529 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3530 [Martin Olsson, Neel Mehta]
3532 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3533 accommodate for stack sorting, always a write lock!).
3536 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3537 excessive delays in the RAND_poll(): over a minute. As a workaround
3538 include a time check in the inner Heap32Next loop too.
3541 *) The code that handled flushing of data in SSL/TLS originally used the
3542 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3543 the problem outlined in PR#1949. The fix suggested there however can
3544 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3545 of Apache). So instead simplify the code to flush unconditionally.
3546 This should be fine since flushing with no data to flush is a no op.
3549 *) Handle TLS versions 2.0 and later properly and correctly use the
3550 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3551 off ancient servers have a habit of sticking around for a while...
3554 *) Modify compression code so it frees up structures without using the
3555 ex_data callbacks. This works around a problem where some applications
3556 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3557 restarting) then use compression (e.g. SSL with compression) later.
3558 This results in significant per-connection memory leaks and
3559 has caused some security issues including CVE-2008-1678 and
3563 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3564 change when encrypting or decrypting.
3567 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3568 connect and renegotiate with servers which do not support RI.
3569 Until RI is more widely deployed this option is enabled by default.
3572 *) Add "missing" ssl ctrls to clear options and mode.
3575 *) If client attempts to renegotiate and doesn't support RI respond with
3576 a no_renegotiation alert as required by RFC5746. Some renegotiating
3577 TLS clients will continue a connection gracefully when they receive
3578 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3579 waiting for a server hello which it will never receive. Now we treat a
3580 received no_renegotiation alert as a fatal error. This is because
3581 applications requesting a renegotiation might well expect it to succeed
3582 and would have no code in place to handle the server denying it so the
3583 only safe thing to do is to terminate the connection.
3586 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3587 peer supports secure renegotiation and 0 otherwise. Print out peer
3588 renegotiation support in s_client/s_server.
3591 *) Replace the highly broken and deprecated SPKAC certification method with
3592 the updated NID creation version. This should correctly handle UTF8.
3595 *) Implement RFC5746. Re-enable renegotiation but require the extension
3596 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3597 turns out to be a bad idea. It has been replaced by
3598 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3599 SSL_CTX_set_options(). This is really not recommended unless you
3600 know what you are doing.
3601 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3603 *) Fixes to stateless session resumption handling. Use initial_ctx when
3604 issuing and attempting to decrypt tickets in case it has changed during
3605 servername handling. Use a non-zero length session ID when attempting
3606 stateless session resumption: this makes it possible to determine if
3607 a resumption has occurred immediately after receiving server hello
3608 (several places in OpenSSL subtly assume this) instead of later in
3612 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3613 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3614 fixes for a few places where the return code is not checked
3616 [Julia Lawall <julia@diku.dk>]
3618 *) Add --strict-warnings option to Configure script to include devteam
3619 warnings in other configurations.
3622 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3623 makes it possible to install openssl libraries in locations which
3624 have names other than "lib", for example "/usr/lib64" which some
3626 [Steve Henson, based on patch from Jeremy Utley]
3628 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3629 X690 8.9.12 and can produce some misleading textual output of OIDs.
3630 [Steve Henson, reported by Dan Kaminsky]
3632 *) Delete MD2 from algorithm tables. This follows the recommendation in
3633 several standards that it is not used in new applications due to
3634 several cryptographic weaknesses. For binary compatibility reasons
3635 the MD2 API is still compiled in by default.
3638 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3642 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3643 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3645 [Guenter <lists@gknw.net>]
3647 *) Fix the server certificate chain building code to use X509_verify_cert(),
3648 it used to have an ad-hoc builder which was unable to cope with anything
3649 other than a simple chain.
3650 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3652 *) Don't check self signed certificate signatures in X509_verify_cert()
3653 by default (a flag can override this): it just wastes time without
3654 adding any security. As a useful side effect self signed root CAs
3655 with non-FIPS digests are now usable in FIPS mode.
3658 *) In dtls1_process_out_of_seq_message() the check if the current message
3659 is already buffered was missing. For every new message was memory
3660 allocated, allowing an attacker to perform an denial of service attack
3661 with sending out of seq handshake messages until there is no memory
3662 left. Additionally every future messege was buffered, even if the
3663 sequence number made no sense and would be part of another handshake.
3664 So only messages with sequence numbers less than 10 in advance will be
3665 buffered. (CVE-2009-1378)
3666 [Robin Seggelmann, discovered by Daniel Mentz]
3668 *) Records are buffered if they arrive with a future epoch to be
3669 processed after finishing the corresponding handshake. There is
3670 currently no limitation to this buffer allowing an attacker to perform
3671 a DOS attack with sending records with future epochs until there is no
3672 memory left. This patch adds the pqueue_size() function to determine
3673 the size of a buffer and limits the record buffer to 100 entries.
3675 [Robin Seggelmann, discovered by Daniel Mentz]
3677 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3678 parent structure is freed. (CVE-2009-1379)
3681 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3682 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3685 [Ilya O. <vrghost@gmail.com>]
3687 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3689 *) Disable renegotiation completely - this fixes a severe security
3690 problem (CVE-2009-3555) at the cost of breaking all
3691 renegotiation. Renegotiation can be re-enabled by setting
3692 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3693 run-time. This is really not recommended unless you know what
3697 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3699 *) Don't set val to NULL when freeing up structures, it is freed up by
3700 underlying code. If sizeof(void *) > sizeof(long) this can result in
3701 zeroing past the valid field. (CVE-2009-0789)
3702 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3704 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3705 checked correctly. This would allow some invalid signed attributes to
3706 appear to verify correctly. (CVE-2009-0591)
3707 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3709 *) Reject UniversalString and BMPString types with invalid lengths. This
3710 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3711 a legal length. (CVE-2009-0590)
3714 *) Set S/MIME signing as the default purpose rather than setting it
3715 unconditionally. This allows applications to override it at the store
3719 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3720 to handle some structures.
3723 *) Improve efficiency of mem_gets: don't search whole buffer each time
3725 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3727 *) New -hex option for openssl rand.
3730 *) Print out UTF8String and NumericString when parsing ASN1.
3733 *) Support NumericString type for name components.
3736 *) Allow CC in the environment to override the automatically chosen
3737 compiler. Note that nothing is done to ensure flags work with the
3741 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3743 *) Properly check EVP_VerifyFinal() and similar return values
3745 [Ben Laurie, Bodo Moeller, Google Security Team]
3747 *) Enable TLS extensions by default.
3750 *) Allow the CHIL engine to be loaded, whether the application is
3751 multithreaded or not. (This does not release the developer from the
3752 obligation to set up the dynamic locking callbacks.)
3753 [Sander Temme <sander@temme.net>]
3755 *) Use correct exit code if there is an error in dgst command.
3756 [Steve Henson; problem pointed out by Roland Dirlewanger]
3758 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3759 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3762 *) Add experimental JPAKE support, including demo authentication in
3763 s_client and s_server.
3766 *) Set the comparison function in v3_addr_canonize().
3767 [Rob Austein <sra@hactrn.net>]
3769 *) Add support for XMPP STARTTLS in s_client.
3770 [Philip Paeps <philip@freebsd.org>]
3772 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3773 to ensure that even with this option, only ciphersuites in the
3774 server's preference list will be accepted. (Note that the option
3775 applies only when resuming a session, so the earlier behavior was
3776 just about the algorithm choice for symmetric cryptography.)
3779 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3781 *) Fix NULL pointer dereference if a DTLS server received
3782 ChangeCipherSpec as first record (CVE-2009-1386).
3785 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3786 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3789 *) The fix in 0.9.8c that supposedly got rid of unsafe
3790 double-checked locking was incomplete for RSA blinding,
3791 addressing just one layer of what turns out to have been
3792 doubly unsafe triple-checked locking.
3794 So now fix this for real by retiring the MONT_HELPER macro
3795 in crypto/rsa/rsa_eay.c.
3797 [Bodo Moeller; problem pointed out by Marius Schilder]
3799 *) Various precautionary measures:
3801 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3803 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3804 (NB: This would require knowledge of the secret session ticket key
3805 to exploit, in which case you'd be SOL either way.)
3807 - Change bn_nist.c so that it will properly handle input BIGNUMs
3808 outside the expected range.
3810 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3813 [Neel Mehta, Bodo Moeller]
3815 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3816 the load fails. Useful for distros.
3817 [Ben Laurie and the FreeBSD team]
3819 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3822 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3825 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3827 This work was sponsored by Logica.
3830 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3831 keystores. Support for SSL/TLS client authentication too.
3832 Not compiled unless enable-capieng specified to Configure.
3834 This work was sponsored by Logica.
3837 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3838 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3839 attribute creation routines such as certifcate requests and PKCS#12
3843 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3845 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3846 handshake which could lead to a cilent crash as found using the
3847 Codenomicon TLS test suite (CVE-2008-1672)
3848 [Steve Henson, Mark Cox]
3850 *) Fix double free in TLS server name extensions which could lead to
3851 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3854 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3856 Clear the error queue to ensure that error entries left from
3857 older function calls do not interfere with the correct operation.
3858 [Lutz Jaenicke, Erik de Castro Lopo]
3860 *) Remove root CA certificates of commercial CAs:
3862 The OpenSSL project does not recommend any specific CA and does not
3863 have any policy with respect to including or excluding any CA.
3864 Therefore it does not make any sense to ship an arbitrary selection
3865 of root CA certificates with the OpenSSL software.
3868 *) RSA OAEP patches to fix two separate invalid memory reads.
3869 The first one involves inputs when 'lzero' is greater than
3870 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3871 before the beginning of from). The second one involves inputs where
3872 the 'db' section contains nothing but zeroes (there is a one-byte
3873 invalid read after the end of 'db').
3874 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3876 *) Partial backport from 0.9.9-dev:
3878 Introduce bn_mul_mont (dedicated Montgomery multiplication
3879 procedure) as a candidate for BIGNUM assembler implementation.
3880 While 0.9.9-dev uses assembler for various architectures, only
3881 x86_64 is available by default here in the 0.9.8 branch, and
3882 32-bit x86 is available through a compile-time setting.
3884 To try the 32-bit x86 assembler implementation, use Configure
3885 option "enable-montasm" (which exists only for this backport).
3887 As "enable-montasm" for 32-bit x86 disclaims code stability
3888 anyway, in this constellation we activate additional code
3889 backported from 0.9.9-dev for further performance improvements,
3890 namely BN_from_montgomery_word. (To enable this otherwise,
3891 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3893 [Andy Polyakov (backport partially by Bodo Moeller)]
3895 *) Add TLS session ticket callback. This allows an application to set
3896 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3897 values. This is useful for key rollover for example where several key
3898 sets may exist with different names.
3901 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3902 This was broken until now in 0.9.8 releases, such that the only way
3903 a registered ENGINE could be used (assuming it initialises
3904 successfully on the host) was to explicitly set it as the default
3905 for the relevant algorithms. This is in contradiction with 0.9.7
3906 behaviour and the documentation. With this fix, when an ENGINE is
3907 registered into a given algorithm's table of implementations, the
3908 'uptodate' flag is reset so that auto-discovery will be used next
3909 time a new context for that algorithm attempts to select an
3911 [Ian Lister (tweaked by Geoff Thorpe)]
3913 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3914 implemention in the following ways:
3916 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3919 Lack of BER streaming support means one pass streaming processing is
3920 only supported if data is detached: setting the streaming flag is
3921 ignored for embedded content.
3923 CMS support is disabled by default and must be explicitly enabled
3924 with the enable-cms configuration option.
3927 *) Update the GMP engine glue to do direct copies between BIGNUM and
3928 mpz_t when openssl and GMP use the same limb size. Otherwise the
3929 existing "conversion via a text string export" trick is still used.
3930 [Paul Sheer <paulsheer@gmail.com>]
3932 *) Zlib compression BIO. This is a filter BIO which compressed and
3933 uncompresses any data passed through it.
3936 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3937 RFC3394 compatible AES key wrapping.
3940 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3941 sets string data without copying. X509_ALGOR_set0() and
3942 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3943 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3944 from an X509_ATTRIBUTE structure optionally checking it occurs only
3945 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3949 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3950 to get the expected BN_FLG_CONSTTIME behavior.
3951 [Bodo Moeller (Google)]
3955 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3956 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3957 - added some more tests to do_tests.pl
3958 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3959 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3960 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3961 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3962 - various changes to netware.pl to enable gcc-cross builds on Win32
3964 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3965 - various changes to fix missing prototype warnings
3966 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3967 - added AES, WHIRLPOOL and CPUID assembler code to build files
3968 - added missing AES assembler make rules to mk1mf.pl
3969 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3970 [Guenter Knauf <eflash@gmx.net>]
3972 *) Implement certificate status request TLS extension defined in RFC3546.
3973 A client can set the appropriate parameters and receive the encoded
3974 OCSP response via a callback. A server can query the supplied parameters
3975 and set the encoded OCSP response in the callback. Add simplified examples
3976 to s_client and s_server.
3979 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3981 *) Fix various bugs:
3982 + Binary incompatibility of ssl_ctx_st structure
3983 + DTLS interoperation with non-compliant servers
3984 + Don't call get_session_cb() without proposed session
3985 + Fix ia64 assembler code
3986 [Andy Polyakov, Steve Henson]
3988 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3990 *) DTLS Handshake overhaul. There were longstanding issues with
3991 OpenSSL DTLS implementation, which were making it impossible for
3992 RFC 4347 compliant client to communicate with OpenSSL server.
3993 Unfortunately just fixing these incompatibilities would "cut off"
3994 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3995 server keeps tolerating non RFC compliant syntax. The opposite is
3996 not true, 0.9.8f client can not communicate with earlier server.
3997 This update even addresses CVE-2007-4995.
4000 *) Changes to avoid need for function casts in OpenSSL: some compilers
4001 (gcc 4.2 and later) reject their use.
4002 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4005 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4006 RFC4507bis. The encrypted ticket format is an encrypted encoded
4007 SSL_SESSION structure, that way new session features are automatically
4010 If a client application caches session in an SSL_SESSION structure
4011 support is transparent because tickets are now stored in the encoded
4014 The SSL_CTX structure automatically generates keys for ticket
4015 protection in servers so again support should be possible
4016 with no application modification.
4018 If a client or server wishes to disable RFC4507 support then the option
4019 SSL_OP_NO_TICKET can be set.
4021 Add a TLS extension debugging callback to allow the contents of any client
4022 or server extensions to be examined.
4024 This work was sponsored by Google.
4027 *) Add initial support for TLS extensions, specifically for the server_name
4028 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4029 have new members for a host name. The SSL data structure has an
4030 additional member SSL_CTX *initial_ctx so that new sessions can be
4031 stored in that context to allow for session resumption, even after the
4032 SSL has been switched to a new SSL_CTX in reaction to a client's
4033 server_name extension.
4035 New functions (subject to change):
4037 SSL_get_servername()
4038 SSL_get_servername_type()
4041 New CTRL codes and macros (subject to change):
4043 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4044 - SSL_CTX_set_tlsext_servername_callback()
4045 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4046 - SSL_CTX_set_tlsext_servername_arg()
4047 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4049 openssl s_client has a new '-servername ...' option.
4051 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4052 '-key2 ...', '-servername_fatal' (subject to change). This allows
4053 testing the HostName extension for a specific single host name ('-cert'
4054 and '-key' remain fallbacks for handshakes without HostName
4055 negotiation). If the unrecognized_name alert has to be sent, this by
4056 default is a warning; it becomes fatal with the '-servername_fatal'
4059 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4061 *) Add AES and SSE2 assembly language support to VC++ build.
4064 *) Mitigate attack on final subtraction in Montgomery reduction.
4067 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4068 (which previously caused an internal error).
4071 *) Squeeze another 10% out of IGE mode when in != out.
4074 *) AES IGE mode speedup.
4075 [Dean Gaudet (Google)]
4077 *) Add the Korean symmetric 128-bit cipher SEED (see
4078 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4079 add SEED ciphersuites from RFC 4162:
4081 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4082 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4083 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4084 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4086 To minimize changes between patchlevels in the OpenSSL 0.9.8
4087 series, SEED remains excluded from compilation unless OpenSSL
4088 is configured with 'enable-seed'.
4089 [KISA, Bodo Moeller]
4091 *) Mitigate branch prediction attacks, which can be practical if a
4092 single processor is shared, allowing a spy process to extract
4093 information. For detailed background information, see
4094 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4095 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4096 and Necessary Software Countermeasures"). The core of the change
4097 are new versions BN_div_no_branch() and
4098 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4099 respectively, which are slower, but avoid the security-relevant
4100 conditional branches. These are automatically called by BN_div()
4101 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4102 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4103 remove a conditional branch.
4105 BN_FLG_CONSTTIME is the new name for the previous
4106 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4107 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4108 in the exponent causes BN_mod_exp_mont() to use the alternative
4109 implementation in BN_mod_exp_mont_consttime().) The old name
4110 remains as a deprecated alias.
4112 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4113 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4114 constant-time implementations for more than just exponentiation.
4115 Here too the old name is kept as a deprecated alias.
4117 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4118 the BN_BLINDING structure gets an independent copy of the
4119 modulus. This means that the previous "BIGNUM *m" argument to
4120 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4121 essentially becomes "const BIGNUM *m", although we can't actually
4122 change this in the header file before 0.9.9. It allows
4123 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4124 enable BN_FLG_CONSTTIME.
4126 [Matthew D Wood (Intel Corp)]
4128 *) In the SSL/TLS server implementation, be strict about session ID
4129 context matching (which matters if an application uses a single
4130 external cache for different purposes). Previously,
4131 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4132 set. This did ensure strict client verification, but meant that,
4133 with applications using a single external cache for quite
4134 different requirements, clients could circumvent ciphersuite
4135 restrictions for a given session ID context by starting a session
4136 in a different context.
4139 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4140 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4141 authentication-only ciphersuites.
4144 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4145 not complete and could lead to a possible single byte overflow
4146 (CVE-2007-5135) [Ben Laurie]
4148 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4150 *) Since AES128 and AES256 (and similarly Camellia128 and
4151 Camellia256) share a single mask bit in the logic of
4152 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4153 kludge to work properly if AES128 is available and AES256 isn't
4154 (or if Camellia128 is available and Camellia256 isn't).
4157 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4158 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4159 When a point or a seed is encoded in a BIT STRING, we need to
4160 prevent the removal of trailing zero bits to get the proper DER
4161 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4162 of a NamedBitList, for which trailing 0 bits need to be removed.)
4165 *) Have SSL/TLS server implementation tolerate "mismatched" record
4166 protocol version while receiving ClientHello even if the
4167 ClientHello is fragmented. (The server can't insist on the
4168 particular protocol version it has chosen before the ServerHello
4169 message has informed the client about his choice.)
4172 *) Add RFC 3779 support.
4173 [Rob Austein for ARIN, Ben Laurie]
4175 *) Load error codes if they are not already present instead of using a
4176 static variable. This allows them to be cleanly unloaded and reloaded.
4177 Improve header file function name parsing.
4180 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4181 or CAPABILITY handshake as required by RFCs.
4184 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4186 *) Introduce limits to prevent malicious keys being able to
4187 cause a denial of service. (CVE-2006-2940)
4188 [Steve Henson, Bodo Moeller]
4190 *) Fix ASN.1 parsing of certain invalid structures that can result
4191 in a denial of service. (CVE-2006-2937) [Steve Henson]
4193 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4194 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4196 *) Fix SSL client code which could crash if connecting to a
4197 malicious SSLv2 server. (CVE-2006-4343)
4198 [Tavis Ormandy and Will Drewry, Google Security Team]
4200 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4201 match only those. Before that, "AES256-SHA" would be interpreted
4202 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4203 the same strength classification in 0.9.7h) as we currently only
4204 have a single AES bit in the ciphersuite description bitmap.
4205 That change, however, also applied to ciphersuite strings such as
4206 "RC4-MD5" that intentionally matched multiple ciphersuites --
4207 namely, SSL 2.0 ciphersuites in addition to the more common ones
4208 from SSL 3.0/TLS 1.0.
4210 So we change the selection algorithm again: Naming an explicit
4211 ciphersuite selects this one ciphersuite, and any other similar
4212 ciphersuite (same bitmap) from *other* protocol versions.
4213 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4214 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4216 Since SSL 2.0 does not have any ciphersuites for which the
4217 128/256 bit distinction would be relevant, this works for now.
4218 The proper fix will be to use different bits for AES128 and
4219 AES256, which would have avoided the problems from the beginning;
4220 however, bits are scarce, so we can only do this in a new release
4221 (not just a patchlevel) when we can change the SSL_CIPHER
4222 definition to split the single 'unsigned long mask' bitmap into
4223 multiple values to extend the available space.
4227 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4229 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4230 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4232 *) Add AES IGE and biIGE modes.
4235 *) Change the Unix randomness entropy gathering to use poll() when
4236 possible instead of select(), since the latter has some
4237 undesirable limitations.
4238 [Darryl Miles via Richard Levitte and Bodo Moeller]
4240 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4241 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4242 cannot be implicitly activated as part of, e.g., the "AES" alias.
4243 However, please upgrade to OpenSSL 0.9.9[-dev] for
4244 non-experimental use of the ECC ciphersuites to get TLS extension
4245 support, which is required for curve and point format negotiation
4246 to avoid potential handshake problems.
4249 *) Disable rogue ciphersuites:
4251 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4252 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4253 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4255 The latter two were purportedly from
4256 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4259 Also deactivate the remaining ciphersuites from
4260 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4261 unofficial, and the ID has long expired.
4264 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4265 dual-core machines) and other potential thread-safety issues.
4268 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4269 versions), which is now available for royalty-free use
4270 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4271 Also, add Camellia TLS ciphersuites from RFC 4132.
4273 To minimize changes between patchlevels in the OpenSSL 0.9.8
4274 series, Camellia remains excluded from compilation unless OpenSSL
4275 is configured with 'enable-camellia'.
4278 *) Disable the padding bug check when compression is in use. The padding
4279 bug check assumes the first packet is of even length, this is not
4280 necessarily true if compresssion is enabled and can result in false
4281 positives causing handshake failure. The actual bug test is ancient
4282 code so it is hoped that implementations will either have fixed it by
4283 now or any which still have the bug do not support compression.
4286 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4288 *) When applying a cipher rule check to see if string match is an explicit
4289 cipher suite and only match that one cipher suite if it is.
4292 *) Link in manifests for VC++ if needed.
4293 [Austin Ziegler <halostatue@gmail.com>]
4295 *) Update support for ECC-based TLS ciphersuites according to
4296 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4297 TLS extensions, which are supported starting with the 0.9.9
4298 branch, not in the OpenSSL 0.9.8 branch).
4301 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4302 opaque EVP_CIPHER_CTX handling.
4305 *) Fixes and enhancements to zlib compression code. We now only use
4306 "zlib1.dll" and use the default __cdecl calling convention on Win32
4307 to conform with the standards mentioned here:
4308 http://www.zlib.net/DLL_FAQ.txt
4309 Static zlib linking now works on Windows and the new --with-zlib-include
4310 --with-zlib-lib options to Configure can be used to supply the location
4311 of the headers and library. Gracefully handle case where zlib library
4315 *) Several fixes and enhancements to the OID generation code. The old code
4316 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4317 handle numbers larger than ULONG_MAX, truncated printing and had a
4318 non standard OBJ_obj2txt() behaviour.
4321 *) Add support for building of engines under engine/ as shared libraries
4322 under VC++ build system.
4325 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4326 Hopefully, we will not see any false combination of paths any more.
4329 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4331 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4332 (part of SSL_OP_ALL). This option used to disable the
4333 countermeasure against man-in-the-middle protocol-version
4334 rollback in the SSL 2.0 server implementation, which is a bad
4335 idea. (CVE-2005-2969)
4337 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4338 for Information Security, National Institute of Advanced Industrial
4339 Science and Technology [AIST], Japan)]
4341 *) Add two function to clear and return the verify parameter flags.
4344 *) Keep cipherlists sorted in the source instead of sorting them at
4345 runtime, thus removing the need for a lock.
4348 *) Avoid some small subgroup attacks in Diffie-Hellman.
4349 [Nick Mathewson and Ben Laurie]
4351 *) Add functions for well-known primes.
4354 *) Extended Windows CE support.
4355 [Satoshi Nakamura and Andy Polyakov]
4357 *) Initialize SSL_METHOD structures at compile time instead of during
4358 runtime, thus removing the need for a lock.
4361 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4362 attempting to decrypt each encrypted key in turn. Add support to
4366 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4368 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4371 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4374 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4375 key into the same file any more.
4378 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4381 *) Add -utf8 command line and config file option to 'ca'.
4382 [Stefan <stf@udoma.org]
4384 *) Removed the macro des_crypt(), as it seems to conflict with some
4385 libraries. Use DES_crypt().
4388 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4389 involves renaming the source and generated shared-libs for
4390 both. The engines will accept the corrected or legacy ids
4391 ('ncipher' and '4758_cca' respectively) when binding. NB,
4392 this only applies when building 'shared'.
4393 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4395 *) Add attribute functions to EVP_PKEY structure. Modify
4396 PKCS12_create() to recognize a CSP name attribute and
4397 use it. Make -CSP option work again in pkcs12 utility.
4400 *) Add new functionality to the bn blinding code:
4401 - automatic re-creation of the BN_BLINDING parameters after
4402 a fixed number of uses (currently 32)
4403 - add new function for parameter creation
4404 - introduce flags to control the update behaviour of the
4405 BN_BLINDING parameters
4406 - hide BN_BLINDING structure
4407 Add a second BN_BLINDING slot to the RSA structure to improve
4408 performance when a single RSA object is shared among several
4412 *) Add support for DTLS.
4413 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4415 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4416 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4419 *) Remove buggy and incomplete DH cert support from
4420 ssl/ssl_rsa.c and ssl/s3_both.c
4423 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4424 the apps/openssl applications.
4427 *) Compile clean with "-Wall -Wmissing-prototypes
4428 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4429 DEBUG_SAFESTACK must also be set.
4432 *) Change ./Configure so that certain algorithms can be disabled by default.
4433 The new counterpiece to "no-xxx" is "enable-xxx".
4435 The patented RC5 and MDC2 algorithms will now be disabled unless
4436 "enable-rc5" and "enable-mdc2", respectively, are specified.
4438 (IDEA remains enabled despite being patented. This is because IDEA
4439 is frequently required for interoperability, and there is no license
4440 fee for non-commercial use. As before, "no-idea" can be used to
4441 avoid this algorithm.)
4445 *) Add processing of proxy certificates (see RFC 3820). This work was
4446 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4447 EGEE (Enabling Grids for E-science in Europe).
4450 *) RC4 performance overhaul on modern architectures/implementations, such
4451 as Intel P4, IA-64 and AMD64.
4454 *) New utility extract-section.pl. This can be used specify an alternative
4455 section number in a pod file instead of having to treat each file as
4456 a separate case in Makefile. This can be done by adding two lines to the
4459 =for comment openssl_section:XXX
4461 The blank line is mandatory.
4465 *) New arguments -certform, -keyform and -pass for s_client and s_server
4466 to allow alternative format key and certificate files and passphrase
4470 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4471 update associated structures and add various utility functions.
4473 Add new policy related verify parameters, include policy checking in
4474 standard verify code. Enhance 'smime' application with extra parameters
4475 to support policy checking and print out.
4478 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4479 Nehemiah processors. These extensions support AES encryption in hardware
4480 as well as RNG (though RNG support is currently disabled).
4481 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4483 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4486 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4487 [Andy Polyakov and a number of other people]
4489 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4490 implementation contributed by IBM.
4491 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4493 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4494 exponent rather than 'unsigned long'. There is a corresponding change to
4495 the new 'rsa_keygen' element of the RSA_METHOD structure.
4496 [Jelte Jansen, Geoff Thorpe]
4498 *) Functionality for creating the initial serial number file is now
4499 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4501 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4502 number file to 1, which is bound to cause problems. To avoid
4503 the problems while respecting compatibility between different 0.9.7
4504 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4505 CA.pl for serial number initialization. With the new release 0.9.8,
4506 we can fix the problem directly in the 'ca' utility.)
4509 *) Reduced header interdepencies by declaring more opaque objects in
4510 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4511 give fewer recursive includes, which could break lazy source code - so
4512 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4513 developers should define this symbol when building and using openssl to
4514 ensure they track the recommended behaviour, interfaces, [etc], but
4515 backwards-compatible behaviour prevails when this isn't defined.
4518 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4521 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4522 This will generate a random key of the appropriate length based on the
4523 cipher context. The EVP_CIPHER can provide its own random key generation
4524 routine to support keys of a specific form. This is used in the des and
4525 3des routines to generate a key of the correct parity. Update S/MIME
4526 code to use new functions and hence generate correct parity DES keys.
4527 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4528 valid (weak or incorrect parity).
4531 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4532 as looking them up. This is useful when the verified structure may contain
4533 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4534 present unless the new PKCS7_NO_CRL flag is asserted.
4537 *) Extend ASN1 oid configuration module. It now additionally accepts the
4540 shortName = some long name, 1.2.3.4
4543 *) Reimplemented the BN_CTX implementation. There is now no more static
4544 limitation on the number of variables it can handle nor the depth of the
4545 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4546 information can now expand as required, and rather than having a single
4547 static array of bignums, BN_CTX now uses a linked-list of such arrays
4548 allowing it to expand on demand whilst maintaining the usefulness of
4549 BN_CTX's "bundling".
4552 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4553 to allow all RSA operations to function using a single BN_CTX.
4556 *) Preliminary support for certificate policy evaluation and checking. This
4557 is initially intended to pass the tests outlined in "Conformance Testing
4558 of Relying Party Client Certificate Path Processing Logic" v1.07.
4561 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4562 remained unused and not that useful. A variety of other little bignum
4563 tweaks and fixes have also been made continuing on from the audit (see
4567 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4568 associated ASN1, EVP and SSL functions and old ASN1 macros.
4571 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4572 and this should never fail. So the return value from the use of
4573 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4574 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4577 *) BN_CTX_get() should return zero-valued bignums, providing the same
4578 initialised value as BN_new().
4579 [Geoff Thorpe, suggested by Ulf Möller]
4581 *) Support for inhibitAnyPolicy certificate extension.
4584 *) An audit of the BIGNUM code is underway, for which debugging code is
4585 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4586 is considered valid when processing BIGNUMs, and causes execution to
4587 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4588 further steps are taken to deliberately pollute unused data in BIGNUM
4589 structures to try and expose faulty code further on. For now, openssl will
4590 (in its default mode of operation) continue to tolerate the inconsistent
4591 forms that it has tolerated in the past, but authors and packagers should
4592 consider trying openssl and their own applications when compiled with
4593 these debugging symbols defined. It will help highlight potential bugs in
4594 their own code, and will improve the test coverage for OpenSSL itself. At
4595 some point, these tighter rules will become openssl's default to improve
4596 maintainability, though the assert()s and other overheads will remain only
4597 in debugging configurations. See bn.h for more details.
4598 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4600 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4601 that can only be obtained through BN_CTX_new() (which implicitly
4602 initialises it). The presence of this function only made it possible
4603 to overwrite an existing structure (and cause memory leaks).
4606 *) Because of the callback-based approach for implementing LHASH as a
4607 template type, lh_insert() adds opaque objects to hash-tables and
4608 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4609 to clean up those corresponding objects before destroying the hash table
4610 (and losing the object pointers). So some over-zealous constifications in
4611 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4612 objects as "const" and the lh_doall[_arg] callback wrappers are not
4613 prototyped to have "const" restrictions on the object pointers they are
4614 given (and so aren't required to cast them away any more).
4617 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4618 (speed) prefers to use its own implementation. The two implementations
4619 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4620 its object type properly exposed (MS_TM) instead of casting to/from "char
4621 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4622 aren't necessarily the greatest nomenclatures - but this is what was used
4623 internally to the implementation so I've used that for now.
4626 *) Ensure that deprecated functions do not get compiled when
4627 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4628 the self-tests were still using deprecated key-generation functions so
4629 these have been updated also.
4632 *) Reorganise PKCS#7 code to separate the digest location functionality
4633 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4634 New function PKCS7_set_digest() to set the digest type for PKCS#7
4635 digestedData type. Add additional code to correctly generate the
4636 digestedData type and add support for this type in PKCS7 initialization
4640 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4641 structure of type "other".
4644 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4645 sure the loop does correctly stop and breaking ("division by zero")
4646 modulus operations are not performed. The (pre-generated) prime
4647 table crypto/bn/bn_prime.h was already correct, but it could not be
4648 re-generated on some platforms because of the "division by zero"
4649 situation in the script.
4650 [Ralf S. Engelschall]
4652 *) Update support for ECC-based TLS ciphersuites according to
4653 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4654 SHA-1 now is only used for "small" curves (where the
4655 representation of a field element takes up to 24 bytes); for
4656 larger curves, the field element resulting from ECDH is directly
4657 used as premaster secret.
4658 [Douglas Stebila (Sun Microsystems Laboratories)]
4660 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4661 curve secp160r1 to the tests.
4662 [Douglas Stebila (Sun Microsystems Laboratories)]
4664 *) Add the possibility to load symbols globally with DSO.
4665 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4667 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4668 control of the error stack.
4671 *) Add support for STORE in ENGINE.
4674 *) Add the STORE type. The intention is to provide a common interface
4675 to certificate and key stores, be they simple file-based stores, or
4676 HSM-type store, or LDAP stores, or...
4677 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4680 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4681 pass a list of arguments to any function as well as provide a way
4682 for a function to pass data back to the caller.
4685 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4686 works like BUF_strdup() but can be used to duplicate a portion of
4687 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4691 *) Add the function sk_find_ex() which works like sk_find(), but will
4692 return an index to an element even if an exact match couldn't be
4693 found. The index is guaranteed to point at the element where the
4694 searched-for key would be inserted to preserve sorting order.
4697 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4698 takes an extra flags argument for optional functionality. Currently,
4699 the following flags are defined:
4701 OBJ_BSEARCH_VALUE_ON_NOMATCH
4702 This one gets OBJ_bsearch_ex() to return a pointer to the first
4703 element where the comparing function returns a negative or zero
4706 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4707 This one gets OBJ_bsearch_ex() to return a pointer to the first
4708 element where the comparing function returns zero. This is useful
4709 if there are more than one element where the comparing function
4713 *) Make it possible to create self-signed certificates with 'openssl ca'
4714 in such a way that the self-signed certificate becomes part of the
4715 CA database and uses the same mechanisms for serial number generation
4716 as all other certificate signing. The new flag '-selfsign' enables
4717 this functionality. Adapt CA.sh and CA.pl.in.
4720 *) Add functionality to check the public key of a certificate request
4721 against a given private. This is useful to check that a certificate
4722 request can be signed by that key (self-signing).
4725 *) Make it possible to have multiple active certificates with the same
4726 subject in the CA index file. This is done only if the keyword
4727 'unique_subject' is set to 'no' in the main CA section (default
4728 if 'CA_default') of the configuration file. The value is saved
4729 with the database itself in a separate index attribute file,
4730 named like the index file with '.attr' appended to the name.
4733 *) Generate muti valued AVAs using '+' notation in config files for
4737 *) Support for nameConstraints certificate extension.
4740 *) Support for policyConstraints certificate extension.
4743 *) Support for policyMappings certificate extension.
4746 *) Make sure the default DSA_METHOD implementation only uses its
4747 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4748 and change its own handlers to be NULL so as to remove unnecessary
4749 indirection. This lets alternative implementations fallback to the
4750 default implementation more easily.
4753 *) Support for directoryName in GeneralName related extensions
4757 *) Make it possible to link applications using Makefile.shared.
4758 Make that possible even when linking against static libraries!
4761 *) Support for single pass processing for S/MIME signing. This now
4762 means that S/MIME signing can be done from a pipe, in addition
4763 cleartext signing (multipart/signed type) is effectively streaming
4764 and the signed data does not need to be all held in memory.
4766 This is done with a new flag PKCS7_STREAM. When this flag is set
4767 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4768 is done after the data is output (and digests calculated) in
4769 SMIME_write_PKCS7().
4772 *) Add full support for -rpath/-R, both in shared libraries and
4773 applications, at least on the platforms where it's known how
4777 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4778 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4779 will now compute a table of multiples of the generator that
4780 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4781 faster (notably in the case of a single point multiplication,
4782 scalar * generator).
4783 [Nils Larsch, Bodo Moeller]
4785 *) IPv6 support for certificate extensions. The various extensions
4786 which use the IP:a.b.c.d can now take IPv6 addresses using the
4787 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4791 *) Added an ENGINE that implements RSA by performing private key
4792 exponentiations with the GMP library. The conversions to and from
4793 GMP's mpz_t format aren't optimised nor are any montgomery forms
4794 cached, and on x86 it appears OpenSSL's own performance has caught up.
4795 However there are likely to be other architectures where GMP could
4796 provide a boost. This ENGINE is not built in by default, but it can be
4797 specified at Configure time and should be accompanied by the necessary
4798 linker additions, eg;
4799 ./config -DOPENSSL_USE_GMP -lgmp
4802 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4803 testing availability of engines with "-t" - the old behaviour is
4804 produced by increasing the feature's verbosity with "-tt".
4807 *) ECDSA routines: under certain error conditions uninitialized BN objects
4808 could be freed. Solution: make sure initialization is performed early
4809 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4813 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4814 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4815 software implementations. For DSA and DH, parameter generation can
4816 also be overriden by providing the appropriate method callbacks.
4819 *) Change the "progress" mechanism used in key-generation and
4820 primality testing to functions that take a new BN_GENCB pointer in
4821 place of callback/argument pairs. The new API functions have "_ex"
4822 postfixes and the older functions are reimplemented as wrappers for
4823 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4824 declarations of the old functions to help (graceful) attempts to
4825 migrate to the new functions. Also, the new key-generation API
4826 functions operate on a caller-supplied key-structure and return
4827 success/failure rather than returning a key or NULL - this is to
4828 help make "keygen" another member function of RSA_METHOD etc.
4830 Example for using the new callback interface:
4832 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4836 BN_GENCB_set(&my_cb, my_callback, my_arg);
4838 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4839 /* For the meaning of a, b in calls to my_callback(), see the
4840 * documentation of the function that calls the callback.
4841 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4842 * my_callback should return 1 if it wants BN_is_prime_ex()
4843 * to continue, or 0 to stop.
4848 *) Change the ZLIB compression method to be stateful, and make it
4849 available to TLS with the number defined in
4850 draft-ietf-tls-compression-04.txt.
4853 *) Add the ASN.1 structures and functions for CertificatePair, which
4854 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4856 CertificatePair ::= SEQUENCE {
4857 forward [0] Certificate OPTIONAL,
4858 reverse [1] Certificate OPTIONAL,
4859 -- at least one of the pair shall be present -- }
4861 Also implement the PEM functions to read and write certificate
4862 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4864 This needed to be defined, mostly for the sake of the LDAP
4865 attribute crossCertificatePair, but may prove useful elsewhere as
4869 *) Make it possible to inhibit symlinking of shared libraries in
4870 Makefile.shared, for Cygwin's sake.
4873 *) Extend the BIGNUM API by creating a function
4874 void BN_set_negative(BIGNUM *a, int neg);
4875 and a macro that behave like
4876 int BN_is_negative(const BIGNUM *a);
4878 to avoid the need to access 'a->neg' directly in applications.
4881 *) Implement fast modular reduction for pseudo-Mersenne primes
4882 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4883 EC_GROUP_new_curve_GFp() will now automatically use this
4885 [Nils Larsch <nla@trustcenter.de>]
4887 *) Add new lock type (CRYPTO_LOCK_BN).
4890 *) Change the ENGINE framework to automatically load engines
4891 dynamically from specific directories unless they could be
4892 found to already be built in or loaded. Move all the
4893 current engines except for the cryptodev one to a new
4895 The engines in engines/ are built as shared libraries if
4896 the "shared" options was given to ./Configure or ./config.
4897 Otherwise, they are inserted in libcrypto.a.
4898 /usr/local/ssl/engines is the default directory for dynamic
4899 engines, but that can be overriden at configure time through
4900 the usual use of --prefix and/or --openssldir, and at run
4901 time with the environment variable OPENSSL_ENGINES.
4902 [Geoff Thorpe and Richard Levitte]
4904 *) Add Makefile.shared, a helper makefile to build shared
4905 libraries. Adapt Makefile.org.
4908 *) Add version info to Win32 DLLs.
4909 [Peter 'Luna' Runestig" <peter@runestig.com>]
4911 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4912 can be added using this API to created arbitrary PKCS#12
4913 files while avoiding the low level API.
4915 New options to PKCS12_create(), key or cert can be NULL and
4916 will then be omitted from the output file. The encryption
4917 algorithm NIDs can be set to -1 for no encryption, the mac
4918 iteration count can be set to 0 to omit the mac.
4920 Enhance pkcs12 utility by making the -nokeys and -nocerts
4921 options work when creating a PKCS#12 file. New option -nomac
4922 to omit the mac, NONE can be set for an encryption algorithm.
4923 New code is modified to use the enhanced PKCS12_create()
4924 instead of the low level API.
4927 *) Extend ASN1 encoder to support indefinite length constructed
4928 encoding. This can output sequences tags and octet strings in
4929 this form. Modify pk7_asn1.c to support indefinite length
4930 encoding. This is experimental and needs additional code to
4931 be useful, such as an ASN1 bio and some enhanced streaming
4934 Extend template encode functionality so that tagging is passed
4935 down to the template encoder.
4938 *) Let 'openssl req' fail if an argument to '-newkey' is not
4939 recognized instead of using RSA as a default.
4942 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4943 As these are not official, they are not included in "ALL";
4944 the "ECCdraft" ciphersuite group alias can be used to select them.
4945 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4947 *) Add ECDH engine support.
4948 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4950 *) Add ECDH in new directory crypto/ecdh/.
4951 [Douglas Stebila (Sun Microsystems Laboratories)]
4953 *) Let BN_rand_range() abort with an error after 100 iterations
4954 without success (which indicates a broken PRNG).
4957 *) Change BN_mod_sqrt() so that it verifies that the input value
4958 is really the square of the return value. (Previously,
4959 BN_mod_sqrt would show GIGO behaviour.)
4962 *) Add named elliptic curves over binary fields from X9.62, SECG,
4963 and WAP/WTLS; add OIDs that were still missing.
4965 [Sheueling Chang Shantz and Douglas Stebila
4966 (Sun Microsystems Laboratories)]
4968 *) Extend the EC library for elliptic curves over binary fields
4969 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4972 EC_GF2m_simple_method
4976 EC_GROUP_new_curve_GF2m
4977 EC_GROUP_set_curve_GF2m
4978 EC_GROUP_get_curve_GF2m
4979 EC_POINT_set_affine_coordinates_GF2m
4980 EC_POINT_get_affine_coordinates_GF2m
4981 EC_POINT_set_compressed_coordinates_GF2m
4983 Point compression for binary fields is disabled by default for
4984 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4987 As binary polynomials are represented as BIGNUMs, various members
4988 of the EC_GROUP and EC_POINT data structures can be shared
4989 between the implementations for prime fields and binary fields;
4990 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4991 are essentially identical to their ..._GFp counterparts.
4992 (For simplicity, the '..._GFp' prefix has been dropped from
4993 various internal method names.)
4995 An internal 'field_div' method (similar to 'field_mul' and
4996 'field_sqr') has been added; this is used only for binary fields.
4998 [Sheueling Chang Shantz and Douglas Stebila
4999 (Sun Microsystems Laboratories)]
5001 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5002 through methods ('mul', 'precompute_mult').
5004 The generic implementations (now internally called 'ec_wNAF_mul'
5005 and 'ec_wNAF_precomputed_mult') remain the default if these
5006 methods are undefined.
5008 [Sheueling Chang Shantz and Douglas Stebila
5009 (Sun Microsystems Laboratories)]
5011 *) New function EC_GROUP_get_degree, which is defined through
5012 EC_METHOD. For curves over prime fields, this returns the bit
5013 length of the modulus.
5015 [Sheueling Chang Shantz and Douglas Stebila
5016 (Sun Microsystems Laboratories)]
5018 *) New functions EC_GROUP_dup, EC_POINT_dup.
5019 (These simply call ..._new and ..._copy).
5021 [Sheueling Chang Shantz and Douglas Stebila
5022 (Sun Microsystems Laboratories)]
5024 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5025 Polynomials are represented as BIGNUMs (where the sign bit is not
5026 used) in the following functions [macros]:
5029 BN_GF2m_sub [= BN_GF2m_add]
5030 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5031 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5032 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5034 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5035 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5036 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5037 BN_GF2m_cmp [= BN_ucmp]
5039 (Note that only the 'mod' functions are actually for fields GF(2^m).
5040 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5042 For some functions, an the irreducible polynomial defining a
5043 field can be given as an 'unsigned int[]' with strictly
5044 decreasing elements giving the indices of those bits that are set;
5045 i.e., p[] represents the polynomial
5046 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5048 p[0] > p[1] > ... > p[k] = 0.
5049 This applies to the following functions:
5054 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5055 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5057 BN_GF2m_mod_sqrt_arr
5058 BN_GF2m_mod_solve_quad_arr
5062 Conversion can be performed by the following functions:
5067 bntest.c has additional tests for binary polynomial arithmetic.
5069 Two implementations for BN_GF2m_mod_div() are available.
5070 The default algorithm simply uses BN_GF2m_mod_inv() and
5071 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5072 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5073 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5075 [Sheueling Chang Shantz and Douglas Stebila
5076 (Sun Microsystems Laboratories)]
5078 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5079 functionality is disabled at compile-time.
5080 [Douglas Stebila <douglas.stebila@sun.com>]
5082 *) Change default behaviour of 'openssl asn1parse' so that more
5083 information is visible when viewing, e.g., a certificate:
5085 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5086 mode the content of non-printable OCTET STRINGs is output in a
5087 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5088 avoid the appearance of a printable string.
5089 [Nils Larsch <nla@trustcenter.de>]
5091 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5093 EC_GROUP_set_asn1_flag()
5094 EC_GROUP_get_asn1_flag()
5095 EC_GROUP_set_point_conversion_form()
5096 EC_GROUP_get_point_conversion_form()
5097 These control ASN1 encoding details:
5098 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5099 has been set to OPENSSL_EC_NAMED_CURVE.
5100 - Points are encoded in uncompressed form by default; options for
5101 asn1_for are as for point2oct, namely
5102 POINT_CONVERSION_COMPRESSED
5103 POINT_CONVERSION_UNCOMPRESSED
5104 POINT_CONVERSION_HYBRID
5106 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5109 EC_GROUP_get0_seed()
5110 EC_GROUP_get_seed_len()
5111 This is used only for ASN1 purposes (so far).
5112 [Nils Larsch <nla@trustcenter.de>]
5114 *) Add 'field_type' member to EC_METHOD, which holds the NID
5115 of the appropriate field type OID. The new function
5116 EC_METHOD_get_field_type() returns this value.
5117 [Nils Larsch <nla@trustcenter.de>]
5122 EC_POINT_point2hex()
5123 EC_POINT_hex2point()
5124 providing useful interfaces to EC_POINT_point2oct() and
5125 EC_POINT_oct2point().
5126 [Nils Larsch <nla@trustcenter.de>]
5128 *) Change internals of the EC library so that the functions
5129 EC_GROUP_set_generator()
5130 EC_GROUP_get_generator()
5131 EC_GROUP_get_order()
5132 EC_GROUP_get_cofactor()
5133 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5134 to methods, which would lead to unnecessary code duplication when
5135 adding different types of curves.
5136 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5138 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5139 arithmetic, and such that modified wNAFs are generated
5140 (which avoid length expansion in many cases).
5143 *) Add a function EC_GROUP_check_discriminant() (defined via
5144 EC_METHOD) that verifies that the curve discriminant is non-zero.
5146 Add a function EC_GROUP_check() that makes some sanity tests
5147 on a EC_GROUP, its generator and order. This includes
5148 EC_GROUP_check_discriminant().
5149 [Nils Larsch <nla@trustcenter.de>]
5151 *) Add ECDSA in new directory crypto/ecdsa/.
5153 Add applications 'openssl ecparam' and 'openssl ecdsa'
5154 (these are based on 'openssl dsaparam' and 'openssl dsa').
5156 ECDSA support is also included in various other files across the
5157 library. Most notably,
5158 - 'openssl req' now has a '-newkey ecdsa:file' option;
5159 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5160 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5161 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5162 them suitable for ECDSA where domain parameters must be
5163 extracted before the specific public key;
5164 - ECDSA engine support has been added.
5165 [Nils Larsch <nla@trustcenter.de>]
5167 *) Include some named elliptic curves, and add OIDs from X9.62,
5168 SECG, and WAP/WTLS. Each curve can be obtained from the new
5170 EC_GROUP_new_by_curve_name(),
5171 and the list of available named curves can be obtained with
5172 EC_get_builtin_curves().
5173 Also add a 'curve_name' member to EC_GROUP objects, which can be
5175 EC_GROUP_set_curve_name()
5176 EC_GROUP_get_curve_name()
5177 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5179 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5180 was actually never needed) and in BN_mul(). The removal in BN_mul()
5181 required a small change in bn_mul_part_recursive() and the addition
5182 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5183 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5184 bn_sub_words() and bn_add_words() except they take arrays with
5188 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5190 *) Cleanse PEM buffers before freeing them since they may contain
5192 [Benjamin Bennett <ben@psc.edu>]
5194 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5195 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5196 authentication-only ciphersuites.
5199 *) Since AES128 and AES256 share a single mask bit in the logic of
5200 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5201 kludge to work properly if AES128 is available and AES256 isn't.
5204 *) Expand security boundary to match 1.1.1 module.
5207 *) Remove redundant features: hash file source, editing of test vectors
5208 modify fipsld to use external fips_premain.c signature.
5211 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5212 run algorithm test programs.
5215 *) Make algorithm test programs more tolerant of whitespace.
5218 *) Have SSL/TLS server implementation tolerate "mismatched" record
5219 protocol version while receiving ClientHello even if the
5220 ClientHello is fragmented. (The server can't insist on the
5221 particular protocol version it has chosen before the ServerHello
5222 message has informed the client about his choice.)
5225 *) Load error codes if they are not already present instead of using a
5226 static variable. This allows them to be cleanly unloaded and reloaded.
5229 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5231 *) Introduce limits to prevent malicious keys being able to
5232 cause a denial of service. (CVE-2006-2940)
5233 [Steve Henson, Bodo Moeller]
5235 *) Fix ASN.1 parsing of certain invalid structures that can result
5236 in a denial of service. (CVE-2006-2937) [Steve Henson]
5238 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5239 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5241 *) Fix SSL client code which could crash if connecting to a
5242 malicious SSLv2 server. (CVE-2006-4343)
5243 [Tavis Ormandy and Will Drewry, Google Security Team]
5245 *) Change ciphersuite string processing so that an explicit
5246 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5247 will no longer include "AES128-SHA"), and any other similar
5248 ciphersuite (same bitmap) from *other* protocol versions (so that
5249 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5250 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5251 changes from 0.9.8b and 0.9.8d.
5254 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5256 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5257 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5259 *) Change the Unix randomness entropy gathering to use poll() when
5260 possible instead of select(), since the latter has some
5261 undesirable limitations.
5262 [Darryl Miles via Richard Levitte and Bodo Moeller]
5264 *) Disable rogue ciphersuites:
5266 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5267 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5268 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5270 The latter two were purportedly from
5271 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5274 Also deactive the remaining ciphersuites from
5275 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5276 unofficial, and the ID has long expired.
5279 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5280 dual-core machines) and other potential thread-safety issues.
5283 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5285 *) Adapt fipsld and the build system to link against the validated FIPS
5286 module in FIPS mode.
5289 *) Fixes for VC++ 2005 build under Windows.
5292 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5293 from a Windows bash shell such as MSYS. It is autodetected from the
5294 "config" script when run from a VC++ environment. Modify standard VC++
5295 build to use fipscanister.o from the GNU make build.
5298 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5300 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5301 The value now differs depending on if you build for FIPS or not.
5302 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5303 safely run with a non-FIPSed libcrypto, as it may crash because of
5304 the difference induced by this change.
5307 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5309 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5310 (part of SSL_OP_ALL). This option used to disable the
5311 countermeasure against man-in-the-middle protocol-version
5312 rollback in the SSL 2.0 server implementation, which is a bad
5313 idea. (CVE-2005-2969)
5315 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5316 for Information Security, National Institute of Advanced Industrial
5317 Science and Technology [AIST], Japan)]
5319 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5320 mainly for FIPS compliance and not fully integrated at this stage.
5323 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5324 the exponentiation using a fixed-length exponent. (Otherwise,
5325 the information leaked through timing could expose the secret key
5326 after many signatures; cf. Bleichenbacher's attack on DSA with
5330 *) Make a new fixed-window mod_exp implementation the default for
5331 RSA, DSA, and DH private-key operations so that the sequence of
5332 squares and multiplies and the memory access pattern are
5333 independent of the particular secret key. This will mitigate
5334 cache-timing and potential related attacks.
5336 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5337 and this is automatically used by BN_mod_exp_mont() if the new flag
5338 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5339 will use this BN flag for private exponents unless the flag
5340 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5341 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5343 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5345 *) Change the client implementation for SSLv23_method() and
5346 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5347 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5348 (Previously, the SSL 2.0 backwards compatible Client Hello
5349 message format would be used even with SSL_OP_NO_SSLv2.)
5352 *) Add support for smime-type MIME parameter in S/MIME messages which some
5356 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5357 a threadsafe manner. Modify rsa code to use new function and add calls
5358 to dsa and dh code (which had race conditions before).
5361 *) Include the fixed error library code in the C error file definitions
5362 instead of fixing them up at runtime. This keeps the error code
5363 structures constant.
5366 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5368 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5371 *) Fixes for newer kerberos headers. NB: the casts are needed because
5372 the 'length' field is signed on one version and unsigned on another
5373 with no (?) obvious way to tell the difference, without these VC++
5374 complains. Also the "definition" of FAR (blank) is no longer included
5375 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5376 some needed definitions.
5379 *) Undo Cygwin change.
5382 *) Added support for proxy certificates according to RFC 3820.
5383 Because they may be a security thread to unaware applications,
5384 they must be explicitly allowed in run-time. See
5385 docs/HOWTO/proxy_certificates.txt for further information.
5388 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5390 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5391 server and client random values. Previously
5392 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5393 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5395 This change has negligible security impact because:
5397 1. Server and client random values still have 24 bytes of pseudo random
5400 2. Server and client random values are sent in the clear in the initial
5403 3. The master secret is derived using the premaster secret (48 bytes in
5404 size for static RSA ciphersuites) as well as client server and random
5407 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5410 [Stephen Henson, reported by UK NISCC]
5412 *) Use Windows randomness collection on Cygwin.
5415 *) Fix hang in EGD/PRNGD query when communication socket is closed
5416 prematurely by EGD/PRNGD.
5417 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5419 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5422 *) Back-port of selected performance improvements from development
5423 branch, as well as improved support for PowerPC platforms.
5426 *) Add lots of checks for memory allocation failure, error codes to indicate
5427 failure and freeing up memory if a failure occurs.
5428 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5430 *) Add new -passin argument to dgst.
5433 *) Perform some character comparisons of different types in X509_NAME_cmp:
5434 this is needed for some certificates that reencode DNs into UTF8Strings
5435 (in violation of RFC3280) and can't or wont issue name rollover
5439 *) Make an explicit check during certificate validation to see that
5440 the CA setting in each certificate on the chain is correct. As a
5441 side effect always do the following basic checks on extensions,
5442 not just when there's an associated purpose to the check:
5444 - if there is an unhandled critical extension (unless the user
5445 has chosen to ignore this fault)
5446 - if the path length has been exceeded (if one is set at all)
5447 - that certain extensions fit the associated purpose (if one has
5451 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5453 *) Avoid a race condition when CRLs are checked in a multi threaded
5454 environment. This would happen due to the reordering of the revoked
5455 entries during signature checking and serial number lookup. Now the
5456 encoding is cached and the serial number sort performed under a lock.
5457 Add new STACK function sk_is_sorted().
5460 *) Add Delta CRL to the extension code.
5463 *) Various fixes to s3_pkt.c so alerts are sent properly.
5464 [David Holmes <d.holmes@f5.com>]
5466 *) Reduce the chances of duplicate issuer name and serial numbers (in
5467 violation of RFC3280) using the OpenSSL certificate creation utilities.
5468 This is done by creating a random 64 bit value for the initial serial
5469 number when a serial number file is created or when a self signed
5470 certificate is created using 'openssl req -x509'. The initial serial
5471 number file is created using 'openssl x509 -next_serial' in CA.pl
5472 rather than being initialized to 1.
5475 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5477 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5478 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5479 [Joe Orton, Steve Henson]
5481 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5483 [Joe Orton, Steve Henson]
5485 *) Make it possible to have multiple active certificates with the same
5486 subject in the CA index file. This is done only if the keyword
5487 'unique_subject' is set to 'no' in the main CA section (default
5488 if 'CA_default') of the configuration file. The value is saved
5489 with the database itself in a separate index attribute file,
5490 named like the index file with '.attr' appended to the name.
5493 *) X509 verify fixes. Disable broken certificate workarounds when
5494 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5495 keyUsage extension present. Don't accept CRLs with unhandled critical
5496 extensions: since verify currently doesn't process CRL extensions this
5497 rejects a CRL with *any* critical extensions. Add new verify error codes
5501 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5502 A clarification of RFC2560 will require the use of OCTET STRINGs and
5503 some implementations cannot handle the current raw format. Since OpenSSL
5504 copies and compares OCSP nonces as opaque blobs without any attempt at
5505 parsing them this should not create any compatibility issues.
5508 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5509 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5510 this HMAC (and other) operations are several times slower than OpenSSL
5514 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5515 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5517 *) Use the correct content when signing type "other".
5520 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5522 *) Fix various bugs revealed by running the NISCC test suite:
5524 Stop out of bounds reads in the ASN1 code when presented with
5525 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5527 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5529 If verify callback ignores invalid public key errors don't try to check
5530 certificate signature with the NULL public key.
5534 *) New -ignore_err option in ocsp application to stop the server
5535 exiting on the first error in a request.
5538 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5539 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5543 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5544 extra data after the compression methods not only for TLS 1.0
5545 but also for SSL 3.0 (as required by the specification).
5546 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5548 *) Change X509_certificate_type() to mark the key as exported/exportable
5549 when it's 512 *bits* long, not 512 bytes.
5552 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5553 blocks during encryption.
5556 *) Various fixes to base64 BIO and non blocking I/O. On write
5557 flushes were not handled properly if the BIO retried. On read
5558 data was not being buffered properly and had various logic bugs.
5559 This also affects blocking I/O when the data being decoded is a
5563 *) Various S/MIME bugfixes and compatibility changes:
5564 output correct application/pkcs7 MIME type if
5565 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5566 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5567 of files as .eml work). Correctly handle very long lines in MIME
5571 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5573 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5574 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5575 a protocol version number mismatch like a decryption error
5576 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5579 *) Turn on RSA blinding by default in the default implementation
5580 to avoid a timing attack. Applications that don't want it can call
5581 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5582 They would be ill-advised to do so in most cases.
5583 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5585 *) Change RSA blinding code so that it works when the PRNG is not
5586 seeded (in this case, the secret RSA exponent is abused as
5587 an unpredictable seed -- if it is not unpredictable, there
5588 is no point in blinding anyway). Make RSA blinding thread-safe
5589 by remembering the creator's thread ID in rsa->blinding and
5590 having all other threads use local one-time blinding factors
5591 (this requires more computation than sharing rsa->blinding, but
5592 avoids excessive locking; and if an RSA object is not shared
5593 between threads, blinding will still be very fast).
5596 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5597 ENGINE as defaults for all supported algorithms irrespective of
5598 the 'flags' parameter. 'flags' is now honoured, so applications
5599 should make sure they are passing it correctly.
5602 *) Target "mingw" now allows native Windows code to be generated in
5603 the Cygwin environment as well as with the MinGW compiler.
5606 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5608 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5609 via timing by performing a MAC computation even if incorrect
5610 block cipher padding has been found. This is a countermeasure
5611 against active attacks where the attacker has to distinguish
5612 between bad padding and a MAC verification error. (CVE-2003-0078)
5614 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5615 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5616 Martin Vuagnoux (EPFL, Ilion)]
5618 *) Make the no-err option work as intended. The intention with no-err
5619 is not to have the whole error stack handling routines removed from
5620 libcrypto, it's only intended to remove all the function name and
5621 reason texts, thereby removing some of the footprint that may not
5622 be interesting if those errors aren't displayed anyway.
5624 NOTE: it's still possible for any application or module to have it's
5625 own set of error texts inserted. The routines are there, just not
5626 used by default when no-err is given.
5629 *) Add support for FreeBSD on IA64.
5630 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5632 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5633 Kerberos function mit_des_cbc_cksum(). Before this change,
5634 the value returned by DES_cbc_cksum() was like the one from
5635 mit_des_cbc_cksum(), except the bytes were swapped.
5636 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5638 *) Allow an application to disable the automatic SSL chain building.
5639 Before this a rather primitive chain build was always performed in
5640 ssl3_output_cert_chain(): an application had no way to send the
5641 correct chain if the automatic operation produced an incorrect result.
5643 Now the chain builder is disabled if either:
5645 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5647 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5649 The reasoning behind this is that an application would not want the
5650 auto chain building to take place if extra chain certificates are
5651 present and it might also want a means of sending no additional
5652 certificates (for example the chain has two certificates and the
5656 *) Add the possibility to build without the ENGINE framework.
5657 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5659 *) Under Win32 gmtime() can return NULL: check return value in
5660 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5663 *) DSA routines: under certain error conditions uninitialized BN objects
5664 could be freed. Solution: make sure initialization is performed early
5665 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5666 Nils Larsch <nla@trustcenter.de> via PR#459)
5669 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5670 checked on reconnect on the client side, therefore session resumption
5671 could still fail with a "ssl session id is different" error. This
5672 behaviour is masked when SSL_OP_ALL is used due to
5673 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5674 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5675 followup to PR #377.
5678 *) IA-32 assembler support enhancements: unified ELF targets, support
5679 for SCO/Caldera platforms, fix for Cygwin shared build.
5682 *) Add support for FreeBSD on sparc64. As a consequence, support for
5683 FreeBSD on non-x86 processors is separate from x86 processors on
5684 the config script, much like the NetBSD support.
5685 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5687 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5689 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5692 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5693 code (06) was taken as the first octet of the session ID and the last
5694 octet was ignored consequently. As a result SSLv2 client side session
5695 caching could not have worked due to the session ID mismatch between
5697 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5701 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5702 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5706 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5707 seems that in spite of existing for more than a year, many application
5708 author have done nothing to provide the necessary callbacks, which
5709 means that this particular engine will not work properly anywhere.
5710 This is a very unfortunate situation which forces us, in the name
5711 of usability, to give the hw_ncipher.c a static lock, which is part
5713 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5714 appear in 0.9.8 or later. We EXPECT application authors to have
5715 dealt properly with this when 0.9.8 is released (unless we actually
5716 make such changes in the libcrypto locking code that changes will
5717 have to be made anyway).
5720 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5721 octets have been read, EOF or an error occurs. Without this change
5722 some truncated ASN1 structures will not produce an error.
5725 *) Disable Heimdal support, since it hasn't been fully implemented.
5726 Still give the possibility to force the use of Heimdal, but with
5727 warnings and a request that patches get sent to openssl-dev.
5730 *) Add the VC-CE target, introduce the WINCE sysname, and add
5731 INSTALL.WCE and appropriate conditionals to make it build.
5732 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5734 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5735 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5736 edit numbers of the version.
5737 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5739 *) Introduce safe string copy and catenation functions
5740 (BUF_strlcpy() and BUF_strlcat()).
5741 [Ben Laurie (CHATS) and Richard Levitte]
5743 *) Avoid using fixed-size buffers for one-line DNs.
5744 [Ben Laurie (CHATS)]
5746 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5747 resizing buffers containing secrets, and use where appropriate.
5748 [Ben Laurie (CHATS)]
5750 *) Avoid using fixed size buffers for configuration file location.
5751 [Ben Laurie (CHATS)]
5753 *) Avoid filename truncation for various CA files.
5754 [Ben Laurie (CHATS)]
5756 *) Use sizeof in preference to magic numbers.
5757 [Ben Laurie (CHATS)]
5759 *) Avoid filename truncation in cert requests.
5760 [Ben Laurie (CHATS)]
5762 *) Add assertions to check for (supposedly impossible) buffer
5764 [Ben Laurie (CHATS)]
5766 *) Don't cache truncated DNS entries in the local cache (this could
5767 potentially lead to a spoofing attack).
5768 [Ben Laurie (CHATS)]
5770 *) Fix various buffers to be large enough for hex/decimal
5771 representations in a platform independent manner.
5772 [Ben Laurie (CHATS)]
5774 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5775 resizing buffers containing secrets, and use where appropriate.
5776 [Ben Laurie (CHATS)]
5778 *) Add BIO_indent() to avoid much slightly worrying code to do
5780 [Ben Laurie (CHATS)]
5782 *) Convert sprintf()/BIO_puts() to BIO_printf().
5783 [Ben Laurie (CHATS)]
5785 *) buffer_gets() could terminate with the buffer only half
5787 [Ben Laurie (CHATS)]
5789 *) Add assertions to prevent user-supplied crypto functions from
5790 overflowing internal buffers by having large block sizes, etc.
5791 [Ben Laurie (CHATS)]
5793 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5795 [Ben Laurie (CHATS)]
5797 *) Eliminate unused copy of key in RC4.
5798 [Ben Laurie (CHATS)]
5800 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5801 [Ben Laurie (CHATS)]
5803 *) Fix off-by-one error in EGD path.
5804 [Ben Laurie (CHATS)]
5806 *) If RANDFILE path is too long, ignore instead of truncating.
5807 [Ben Laurie (CHATS)]
5809 *) Eliminate unused and incorrectly sized X.509 structure
5811 [Ben Laurie (CHATS)]
5813 *) Eliminate unused and dangerous function knumber().
5814 [Ben Laurie (CHATS)]
5816 *) Eliminate unused and dangerous structure, KSSL_ERR.
5817 [Ben Laurie (CHATS)]
5819 *) Protect against overlong session ID context length in an encoded
5820 session object. Since these are local, this does not appear to be
5822 [Ben Laurie (CHATS)]
5824 *) Change from security patch (see 0.9.6e below) that did not affect
5825 the 0.9.6 release series:
5827 Remote buffer overflow in SSL3 protocol - an attacker could
5828 supply an oversized master key in Kerberos-enabled versions.
5830 [Ben Laurie (CHATS)]
5832 *) Change the SSL kerb5 codes to match RFC 2712.
5835 *) Make -nameopt work fully for req and add -reqopt switch.
5836 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5838 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5839 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5841 *) Make sure tests can be performed even if the corresponding algorithms
5842 have been removed entirely. This was also the last step to make
5843 OpenSSL compilable with DJGPP under all reasonable conditions.
5844 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5846 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5847 to allow version independent disabling of normally unselected ciphers,
5848 which may be activated as a side-effect of selecting a single cipher.
5850 (E.g., cipher list string "RSA" enables ciphersuites that are left
5851 out of "ALL" because they do not provide symmetric encryption.
5852 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5853 [Lutz Jaenicke, Bodo Moeller]
5855 *) Add appropriate support for separate platform-dependent build
5856 directories. The recommended way to make a platform-dependent
5857 build directory is the following (tested on Linux), maybe with
5860 # Place yourself outside of the OpenSSL source tree. In
5861 # this example, the environment variable OPENSSL_SOURCE
5862 # is assumed to contain the absolute OpenSSL source directory.
5863 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5864 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5865 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5866 mkdir -p `dirname $F`
5867 ln -s $OPENSSL_SOURCE/$F $F
5870 To be absolutely sure not to disturb the source tree, a "make clean"
5871 is a good thing. If it isn't successful, don't worry about it,
5872 it probably means the source directory is very clean.
5875 *) Make sure any ENGINE control commands make local copies of string
5876 pointers passed to them whenever necessary. Otherwise it is possible
5877 the caller may have overwritten (or deallocated) the original string
5878 data when a later ENGINE operation tries to use the stored values.
5879 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5881 *) Improve diagnostics in file reading and command-line digests.
5882 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5884 *) Add AES modes CFB and OFB to the object database. Correct an
5885 error in AES-CFB decryption.
5888 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5889 allows existing EVP_CIPHER_CTX structures to be reused after
5890 calling EVP_*Final(). This behaviour is used by encryption
5891 BIOs and some applications. This has the side effect that
5892 applications must explicitly clean up cipher contexts with
5893 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5896 *) Check the values of dna and dnb in bn_mul_recursive before calling
5897 bn_mul_comba (a non zero value means the a or b arrays do not contain
5898 n2 elements) and fallback to bn_mul_normal if either is not zero.
5901 *) Fix escaping of non-ASCII characters when using the -subj option
5902 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5905 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5906 form for "surname", serialNumber has no short form.
5907 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5908 therefore remove "mail" short name for "internet 7".
5909 The OID for unique identifiers in X509 certificates is
5910 x500UniqueIdentifier, not uniqueIdentifier.
5911 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5914 *) Add an "init" command to the ENGINE config module and auto initialize
5915 ENGINEs. Without any "init" command the ENGINE will be initialized
5916 after all ctrl commands have been executed on it. If init=1 the
5917 ENGINE is initailized at that point (ctrls before that point are run
5918 on the uninitialized ENGINE and after on the initialized one). If
5919 init=0 then the ENGINE will not be iniatialized at all.
5922 *) Fix the 'app_verify_callback' interface so that the user-defined
5923 argument is actually passed to the callback: In the
5924 SSL_CTX_set_cert_verify_callback() prototype, the callback
5925 declaration has been changed from
5928 int (*cb)(X509_STORE_CTX *,void *);
5929 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5930 i=s->ctx->app_verify_callback(&ctx)
5931 has been changed into
5932 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5934 To update applications using SSL_CTX_set_cert_verify_callback(),
5935 a dummy argument can be added to their callback functions.
5936 [D. K. Smetters <smetters@parc.xerox.com>]
5938 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5939 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5941 *) Add and OPENSSL_LOAD_CONF define which will cause
5942 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5943 This allows older applications to transparently support certain
5944 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5945 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5946 load the config file and OPENSSL_add_all_algorithms_conf() which will
5947 always load it have also been added.
5950 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5951 Adjust NIDs and EVP layer.
5952 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5954 *) Config modules support in openssl utility.
5956 Most commands now load modules from the config file,
5957 though in a few (such as version) this isn't done
5958 because it couldn't be used for anything.
5960 In the case of ca and req the config file used is
5961 the same as the utility itself: that is the -config
5962 command line option can be used to specify an
5966 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5967 use "openssl_conf" if filename is NULL use default openssl config file.
5970 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5971 config section name. Add a new flag to tolerate a missing config file
5972 and move code to CONF_modules_load_file().
5975 *) Support for crypto accelerator cards from Accelerated Encryption
5976 Processing, www.aep.ie. (Use engine 'aep')
5977 The support was copied from 0.9.6c [engine] and adapted/corrected
5978 to work with the new engine framework.
5979 [AEP Inc. and Richard Levitte]
5981 *) Support for SureWare crypto accelerator cards from Baltimore
5982 Technologies. (Use engine 'sureware')
5983 The support was copied from 0.9.6c [engine] and adapted
5984 to work with the new engine framework.
5987 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5988 make the newer ENGINE framework commands for the CHIL engine work.
5989 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5991 *) Make it possible to produce shared libraries on ReliantUNIX.
5992 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5994 *) Add the configuration target debug-linux-ppro.
5995 Make 'openssl rsa' use the general key loading routines
5996 implemented in apps.c, and make those routines able to
5997 handle the key format FORMAT_NETSCAPE and the variant
5999 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6001 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6002 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6004 *) Add -keyform to rsautl, and document -engine.
6005 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6007 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6008 BIO_R_NO_SUCH_FILE error code rather than the generic
6009 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6012 *) Add new functions
6014 ERR_peek_last_error_line
6015 ERR_peek_last_error_line_data.
6016 These are similar to
6019 ERR_peek_error_line_data,
6020 but report on the latest error recorded rather than the first one
6021 still in the error queue.
6022 [Ben Laurie, Bodo Moeller]
6024 *) default_algorithms option in ENGINE config module. This allows things
6026 default_algorithms = ALL
6027 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6030 *) Preliminary ENGINE config module.
6033 *) New experimental application configuration code.
6036 *) Change the AES code to follow the same name structure as all other
6037 symmetric ciphers, and behave the same way. Move everything to
6038 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6039 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6041 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6042 [Ben Laurie and Theo de Raadt]
6044 *) Add option to output public keys in req command.
6045 [Massimiliano Pala madwolf@openca.org]
6047 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6048 (up to about 10% better than before for P-192 and P-224).
6051 *) New functions/macros
6053 SSL_CTX_set_msg_callback(ctx, cb)
6054 SSL_CTX_set_msg_callback_arg(ctx, arg)
6055 SSL_set_msg_callback(ssl, cb)
6056 SSL_set_msg_callback_arg(ssl, arg)
6058 to request calling a callback function
6060 void cb(int write_p, int version, int content_type,
6061 const void *buf, size_t len, SSL *ssl, void *arg)
6063 whenever a protocol message has been completely received
6064 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6065 protocol version according to which the SSL library interprets
6066 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6067 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6068 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6069 specification (change_cipher_spec(20), alert(21), handshake(22)).
6070 'buf' and 'len' point to the actual message, 'ssl' to the
6071 SSL object, and 'arg' is the application-defined value set by
6072 SSL[_CTX]_set_msg_callback_arg().
6074 'openssl s_client' and 'openssl s_server' have new '-msg' options
6075 to enable a callback that displays all protocol messages.
6078 *) Change the shared library support so shared libraries are built as
6079 soon as the corresponding static library is finished, and thereby get
6080 openssl and the test programs linked against the shared library.
6081 This still only happens when the keyword "shard" has been given to
6082 the configuration scripts.
6084 NOTE: shared library support is still an experimental thing, and
6085 backward binary compatibility is still not guaranteed.
6086 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6088 *) Add support for Subject Information Access extension.
6089 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6091 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6092 additional bytes when new memory had to be allocated, not just
6093 when reusing an existing buffer.
6096 *) New command line and configuration option 'utf8' for the req command.
6097 This allows field values to be specified as UTF8 strings.
6100 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6101 runs for the former and machine-readable output for the latter.
6104 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6105 of the e-mail address in the DN (i.e., it will go into a certificate
6106 extension only). The new configuration file option 'email_in_dn = no'
6107 has the same effect.
6108 [Massimiliano Pala madwolf@openca.org]
6110 *) Change all functions with names starting with des_ to be starting
6111 with DES_ instead. Add wrappers that are compatible with libdes,
6112 but are named _ossl_old_des_*. Finally, add macros that map the
6113 des_* symbols to the corresponding _ossl_old_des_* if libdes
6114 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6115 desired, the des_* symbols will be mapped to DES_*, with one
6118 Since we provide two compatibility mappings, the user needs to
6119 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6120 compatibility is desired. The default (i.e., when that macro
6121 isn't defined) is OpenSSL 0.9.6c compatibility.
6123 There are also macros that enable and disable the support of old
6124 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6125 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6126 are defined, the default will apply: to support the old des routines.
6128 In either case, one must include openssl/des.h to get the correct
6129 definitions. Do not try to just include openssl/des_old.h, that
6132 NOTE: This is a major break of an old API into a new one. Software
6133 authors are encouraged to switch to the DES_ style functions. Some
6134 time in the future, des_old.h and the libdes compatibility functions
6135 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6136 default), and then completely removed.
6139 *) Test for certificates which contain unsupported critical extensions.
6140 If such a certificate is found during a verify operation it is
6141 rejected by default: this behaviour can be overridden by either
6142 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6143 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6144 X509_supported_extension() has also been added which returns 1 if a
6145 particular extension is supported.
6148 *) Modify the behaviour of EVP cipher functions in similar way to digests
6149 to retain compatibility with existing code.
6152 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6153 compatibility with existing code. In particular the 'ctx' parameter does
6154 not have to be to be initialized before the call to EVP_DigestInit() and
6155 it is tidied up after a call to EVP_DigestFinal(). New function
6156 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6157 EVP_MD_CTX_copy() changed to not require the destination to be
6158 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6159 requires the destination to be valid.
6161 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6162 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6165 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6166 so that complete 'Handshake' protocol structures are kept in memory
6167 instead of overwriting 'msg_type' and 'length' with 'body' data.
6170 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6171 [Massimo Santin via Richard Levitte]
6173 *) Major restructuring to the underlying ENGINE code. This includes
6174 reduction of linker bloat, separation of pure "ENGINE" manipulation
6175 (initialisation, etc) from functionality dealing with implementations
6176 of specific crypto iterfaces. This change also introduces integrated
6177 support for symmetric ciphers and digest implementations - so ENGINEs
6178 can now accelerate these by providing EVP_CIPHER and EVP_MD
6179 implementations of their own. This is detailed in crypto/engine/README
6180 as it couldn't be adequately described here. However, there are a few
6181 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6182 were changed in the original introduction of ENGINE code have now
6183 reverted back - the hooking from this code to ENGINE is now a good
6184 deal more passive and at run-time, operations deal directly with
6185 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6186 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6187 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6188 they were not being used by the framework as there is no concept of a
6189 BIGNUM_METHOD and they could not be generalised to the new
6190 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6191 ENGINE_cpy() has been removed as it cannot be consistently defined in
6195 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6198 *) Change mkdef.pl to sort symbols that get the same entry number,
6199 and make sure the automatically generated functions ERR_load_*
6200 become part of libeay.num as well.
6203 *) New function SSL_renegotiate_pending(). This returns true once
6204 renegotiation has been requested (either SSL_renegotiate() call
6205 or HelloRequest/ClientHello received from the peer) and becomes
6206 false once a handshake has been completed.
6207 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6208 sends a HelloRequest, but does not ensure that a handshake takes
6209 place. SSL_renegotiate_pending() is useful for checking if the
6210 client has followed the request.)
6213 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6214 By default, clients may request session resumption even during
6215 renegotiation (if session ID contexts permit); with this option,
6216 session resumption is possible only in the first handshake.
6218 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6219 more bits available for options that should not be part of
6220 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6223 *) Add some demos for certificate and certificate request creation.
6226 *) Make maximum certificate chain size accepted from the peer application
6227 settable (SSL*_get/set_max_cert_list()), as proposed by
6228 "Douglas E. Engert" <deengert@anl.gov>.
6231 *) Add support for shared libraries for Unixware-7
6232 (Boyd Lynn Gerber <gerberb@zenez.com>).
6235 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6236 be done prior to destruction. Use this to unload error strings from
6237 ENGINEs that load their own error strings. NB: This adds two new API
6238 functions to "get" and "set" this destroy handler in an ENGINE.
6241 *) Alter all existing ENGINE implementations (except "openssl" and
6242 "openbsd") to dynamically instantiate their own error strings. This
6243 makes them more flexible to be built both as statically-linked ENGINEs
6244 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6245 Also, add stub code to each that makes building them as self-contained
6246 shared-libraries easier (see README.ENGINE).
6249 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6250 implementations into applications that are completely implemented in
6251 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6252 commands that can be used to configure what shared-library to load and
6253 to control aspects of the way it is handled. Also, made an update to
6254 the README.ENGINE file that brings its information up-to-date and
6255 provides some information and instructions on the "dynamic" ENGINE
6256 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6259 *) Make it possible to unload ranges of ERR strings with a new
6260 "ERR_unload_strings" function.
6263 *) Add a copy() function to EVP_MD.
6266 *) Make EVP_MD routines take a context pointer instead of just the
6267 md_data void pointer.
6270 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6271 that the digest can only process a single chunk of data
6272 (typically because it is provided by a piece of
6273 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6274 is only going to provide a single chunk of data, and hence the
6275 framework needn't accumulate the data for oneshot drivers.
6278 *) As with "ERR", make it possible to replace the underlying "ex_data"
6279 functions. This change also alters the storage and management of global
6280 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6281 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6282 index counters. The API functions that use this state have been changed
6283 to take a "class_index" rather than pointers to the class's local STACK
6284 and counter, and there is now an API function to dynamically create new
6285 classes. This centralisation allows us to (a) plug a lot of the
6286 thread-safety problems that existed, and (b) makes it possible to clean
6287 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6288 such data would previously have always leaked in application code and
6289 workarounds were in place to make the memory debugging turn a blind eye
6290 to it. Application code that doesn't use this new function will still
6291 leak as before, but their memory debugging output will announce it now
6292 rather than letting it slide.
6294 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6295 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6296 has a return value to indicate success or failure.
6299 *) Make it possible to replace the underlying "ERR" functions such that the
6300 global state (2 LHASH tables and 2 locks) is only used by the "default"
6301 implementation. This change also adds two functions to "get" and "set"
6302 the implementation prior to it being automatically set the first time
6303 any other ERR function takes place. Ie. an application can call "get",
6304 pass the return value to a module it has just loaded, and that module
6305 can call its own "set" function using that value. This means the
6306 module's "ERR" operations will use (and modify) the error state in the
6307 application and not in its own statically linked copy of OpenSSL code.
6310 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6311 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6312 the operation, and provides a more encapsulated way for external code
6313 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6314 to use these functions rather than manually incrementing the counts.
6316 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6319 *) Add EVP test program.
6322 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6325 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6326 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6327 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6328 These allow a CRL to be built without having to access X509_CRL fields
6329 directly. Modify 'ca' application to use new functions.
6332 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6333 bug workarounds. Rollback attack detection is a security feature.
6334 The problem will only arise on OpenSSL servers when TLSv1 is not
6335 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6336 Software authors not wanting to support TLSv1 will have special reasons
6337 for their choice and can explicitly enable this option.
6338 [Bodo Moeller, Lutz Jaenicke]
6340 *) Rationalise EVP so it can be extended: don't include a union of
6341 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6342 (similar to those existing for EVP_CIPHER_CTX).
6347 EVP_MD_CTX_init(&md); /* new function call */
6348 EVP_DigestInit(&md, EVP_sha1());
6349 EVP_DigestUpdate(&md, in, len);
6350 EVP_DigestFinal(&md, out, NULL);
6351 EVP_MD_CTX_cleanup(&md); /* new function call */
6355 *) Make DES key schedule conform to the usual scheme, as well as
6356 correcting its structure. This means that calls to DES functions
6357 now have to pass a pointer to a des_key_schedule instead of a
6358 plain des_key_schedule (which was actually always a pointer
6361 des_key_schedule ks;
6363 des_set_key_checked(..., &ks);
6364 des_ncbc_encrypt(..., &ks, ...);
6366 (Note that a later change renames 'des_...' into 'DES_...'.)
6369 *) Initial reduction of linker bloat: the use of some functions, such as
6370 PEM causes large amounts of unused functions to be linked in due to
6371 poor organisation. For example pem_all.c contains every PEM function
6372 which has a knock on effect of linking in large amounts of (unused)
6373 ASN1 code. Grouping together similar functions and splitting unrelated
6374 functions prevents this.
6377 *) Cleanup of EVP macros.
6380 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6381 correct _ecb suffix.
6384 *) Add initial OCSP responder support to ocsp application. The
6385 revocation information is handled using the text based index
6386 use by the ca application. The responder can either handle
6387 requests generated internally, supplied in files (for example
6388 via a CGI script) or using an internal minimal server.
6391 *) Add configuration choices to get zlib compression for TLS.
6394 *) Changes to Kerberos SSL for RFC 2712 compliance:
6395 1. Implemented real KerberosWrapper, instead of just using
6396 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6397 2. Implemented optional authenticator field of KerberosWrapper.
6399 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6400 and authenticator structs; see crypto/krb5/.
6402 Generalized Kerberos calls to support multiple Kerberos libraries.
6403 [Vern Staats <staatsvr@asc.hpc.mil>,
6404 Jeffrey Altman <jaltman@columbia.edu>
6405 via Richard Levitte]
6407 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6408 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6409 values for each of the key sizes rather than having just
6410 parameters (and 'speed' generating keys each time).
6413 *) Speed up EVP routines.
6416 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6417 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6418 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6419 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6421 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6422 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6423 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6426 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6428 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6431 *) Added the OS2-EMX target.
6432 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6434 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6435 to support NCONF routines in extension code. New function CONF_set_nconf()
6436 to allow functions which take an NCONF to also handle the old LHASH
6437 structure: this means that the old CONF compatible routines can be
6438 retained (in particular wrt extensions) without having to duplicate the
6439 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6442 *) Enhance the general user interface with mechanisms for inner control
6443 and with possibilities to have yes/no kind of prompts.
6446 *) Change all calls to low level digest routines in the library and
6447 applications to use EVP. Add missing calls to HMAC_cleanup() and
6448 don't assume HMAC_CTX can be copied using memcpy().
6449 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6451 *) Add the possibility to control engines through control names but with
6452 arbitrary arguments instead of just a string.
6453 Change the key loaders to take a UI_METHOD instead of a callback
6454 function pointer. NOTE: this breaks binary compatibility with earlier
6455 versions of OpenSSL [engine].
6456 Adapt the nCipher code for these new conditions and add a card insertion
6460 *) Enhance the general user interface with mechanisms to better support
6461 dialog box interfaces, application-defined prompts, the possibility
6462 to use defaults (for example default passwords from somewhere else)
6463 and interrupts/cancellations.
6466 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6467 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6470 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6471 tidy up some unnecessarily weird code in 'sk_new()').
6472 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6474 *) Change the key loading routines for ENGINEs to use the same kind
6475 callback (pem_password_cb) as all other routines that need this
6479 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6480 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6481 than this minimum value is recommended.
6484 *) New random seeder for OpenVMS, using the system process statistics
6485 that are easily reachable.
6488 *) Windows apparently can't transparently handle global
6489 variables defined in DLLs. Initialisations such as:
6491 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6493 wont compile. This is used by the any applications that need to
6494 declare their own ASN1 modules. This was fixed by adding the option
6495 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6496 needed for static libraries under Win32.
6499 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6500 setting of purpose and trust fields. New X509_STORE trust and
6501 purpose functions and tidy up setting in other SSL functions.
6504 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6505 structure. These are inherited by X509_STORE_CTX when it is
6506 initialised. This allows various defaults to be set in the
6507 X509_STORE structure (such as flags for CRL checking and custom
6508 purpose or trust settings) for functions which only use X509_STORE_CTX
6509 internally such as S/MIME.
6511 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6512 trust settings if they are not set in X509_STORE. This allows X509_STORE
6513 purposes and trust (in S/MIME for example) to override any set by default.
6515 Add command line options for CRL checking to smime, s_client and s_server
6519 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6520 are set then the CRL is looked up in the X509_STORE structure and
6521 its validity and signature checked, then if the certificate is found
6522 in the CRL the verify fails with a revoked error.
6524 Various new CRL related callbacks added to X509_STORE_CTX structure.
6526 Command line options added to 'verify' application to support this.
6528 This needs some additional work, such as being able to handle multiple
6529 CRLs with different times, extension based lookup (rather than just
6530 by subject name) and ultimately more complete V2 CRL extension
6534 *) Add a general user interface API (crypto/ui/). This is designed
6535 to replace things like des_read_password and friends (backward
6536 compatibility functions using this new API are provided).
6537 The purpose is to remove prompting functions from the DES code
6538 section as well as provide for prompting through dialog boxes in
6539 a window system and the like.
6542 *) Add "ex_data" support to ENGINE so implementations can add state at a
6543 per-structure level rather than having to store it globally.
6546 *) Make it possible for ENGINE structures to be copied when retrieved by
6547 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6548 This causes the "original" ENGINE structure to act like a template,
6549 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6550 operational state can be localised to each ENGINE structure, despite the
6551 fact they all share the same "methods". New ENGINE structures returned in
6552 this case have no functional references and the return value is the single
6553 structural reference. This matches the single structural reference returned
6554 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6558 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6559 needs to match any other type at all we need to manually clear the
6563 *) Changes to the "openssl engine" utility to include;
6564 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6565 about an ENGINE's available control commands.
6566 - executing control commands from command line arguments using the
6567 '-pre' and '-post' switches. '-post' is only used if '-t' is
6568 specified and the ENGINE is successfully initialised. The syntax for
6569 the individual commands are colon-separated, for example;
6570 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6573 *) New dynamic control command support for ENGINEs. ENGINEs can now
6574 declare their own commands (numbers), names (strings), descriptions,
6575 and input types for run-time discovery by calling applications. A
6576 subset of these commands are implicitly classed as "executable"
6577 depending on their input type, and only these can be invoked through
6578 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6579 can be based on user input, config files, etc). The distinction is
6580 that "executable" commands cannot return anything other than a boolean
6581 result and can only support numeric or string input, whereas some
6582 discoverable commands may only be for direct use through
6583 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6584 pointers, or other custom uses. The "executable" commands are to
6585 support parameterisations of ENGINE behaviour that can be
6586 unambiguously defined by ENGINEs and used consistently across any
6587 OpenSSL-based application. Commands have been added to all the
6588 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6589 control over shared-library paths without source code alterations.
6592 *) Changed all ENGINE implementations to dynamically allocate their
6593 ENGINEs rather than declaring them statically. Apart from this being
6594 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6595 this also allows the implementations to compile without using the
6596 internal engine_int.h header.
6599 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6600 'const' value. Any code that should be able to modify a RAND_METHOD
6601 should already have non-const pointers to it (ie. they should only
6602 modify their own ones).
6605 *) Made a variety of little tweaks to the ENGINE code.
6606 - "atalla" and "ubsec" string definitions were moved from header files
6607 to C code. "nuron" string definitions were placed in variables
6608 rather than hard-coded - allowing parameterisation of these values
6609 later on via ctrl() commands.
6610 - Removed unused "#if 0"'d code.
6611 - Fixed engine list iteration code so it uses ENGINE_free() to release
6612 structural references.
6613 - Constified the RAND_METHOD element of ENGINE structures.
6614 - Constified various get/set functions as appropriate and added
6615 missing functions (including a catch-all ENGINE_cpy that duplicates
6616 all ENGINE values onto a new ENGINE except reference counts/state).
6617 - Removed NULL parameter checks in get/set functions. Setting a method
6618 or function to NULL is a way of cancelling out a previously set
6619 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6620 and doesn't justify the extra error symbols and code.
6621 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6622 flags from engine_int.h to engine.h.
6623 - Changed prototypes for ENGINE handler functions (init(), finish(),
6624 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6627 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6628 to the algorithm using long division. The binary algorithm can be
6629 used only if the modulus is odd. On 32-bit systems, it is faster
6630 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6631 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6632 up to 450 bits. In 64-bit environments, the binary algorithm
6633 appears to be advantageous for much longer moduli; here we use it
6634 for moduli up to 2048 bits.
6637 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6638 could not support the combine flag in choice fields.
6641 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6642 extensions from a certificate request to the certificate.
6645 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6646 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6647 file: this allows the display of the certificate about to be
6648 signed to be customised, to allow certain fields to be included
6649 or excluded and extension details. The old system didn't display
6650 multicharacter strings properly, omitted fields not in the policy
6651 and couldn't display additional details such as extensions.
6654 *) Function EC_POINTs_mul for multiple scalar multiplication
6655 of an arbitrary number of elliptic curve points
6656 \sum scalars[i]*points[i],
6657 optionally including the generator defined for the EC_GROUP:
6658 scalar*generator + \sum scalars[i]*points[i].
6660 EC_POINT_mul is a simple wrapper function for the typical case
6661 that the point list has just one item (besides the optional
6665 *) First EC_METHODs for curves over GF(p):
6667 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6668 operations and provides various method functions that can also
6669 operate with faster implementations of modular arithmetic.
6671 EC_GFp_mont_method() reuses most functions that are part of
6672 EC_GFp_simple_method, but uses Montgomery arithmetic.
6674 [Bodo Moeller; point addition and point doubling
6675 implementation directly derived from source code provided by
6676 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6678 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6679 crypto/ec/ec_lib.c):
6681 Curves are EC_GROUP objects (with an optional group generator)
6682 based on EC_METHODs that are built into the library.
6684 Points are EC_POINT objects based on EC_GROUP objects.
6686 Most of the framework would be able to handle curves over arbitrary
6687 finite fields, but as there are no obvious types for fields other
6688 than GF(p), some functions are limited to that for now.
6691 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6692 that the file contains a complete HTTP response.
6695 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6696 change the def and num file printf format specifier from "%-40sXXX"
6697 to "%-39s XXX". The latter will always guarantee a space after the
6698 field while the former will cause them to run together if the field
6699 is 40 of more characters long.
6702 *) Constify the cipher and digest 'method' functions and structures
6703 and modify related functions to take constant EVP_MD and EVP_CIPHER
6707 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6708 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6711 *) Modify EVP_Digest*() routines so they now return values. Although the
6712 internal software routines can never fail additional hardware versions
6716 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6718 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6719 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6722 ERR_R_NESTED_ASN1_ERROR
6724 ERR_R_MISSING_ASN1_EOS
6725 were 4 .. 9, conflicting with
6726 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6728 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6729 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6731 Add new error code 'ERR_R_INTERNAL_ERROR'.
6734 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6738 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6739 sets the subject name for a new request or supersedes the
6740 subject name in a given request. Formats that can be parsed are
6741 'CN=Some Name, OU=myOU, C=IT'
6743 'CN=Some Name/OU=myOU/C=IT'.
6745 Add options '-batch' and '-verbose' to 'openssl req'.
6746 [Massimiliano Pala <madwolf@hackmasters.net>]
6748 *) Introduce the possibility to access global variables through
6749 functions on platform were that's the best way to handle exporting
6750 global variables in shared libraries. To enable this functionality,
6751 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6752 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6753 is normally done by Configure or something similar).
6755 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6756 in the source file (foo.c) like this:
6758 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6759 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6761 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6762 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6764 OPENSSL_DECLARE_GLOBAL(int,foo);
6765 #define foo OPENSSL_GLOBAL_REF(foo)
6766 OPENSSL_DECLARE_GLOBAL(double,bar);
6767 #define bar OPENSSL_GLOBAL_REF(bar)
6769 The #defines are very important, and therefore so is including the
6770 header file everywhere where the defined globals are used.
6772 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6773 of ASN.1 items, but that structure is a bit different.
6775 The largest change is in util/mkdef.pl which has been enhanced with
6776 better and easier to understand logic to choose which symbols should
6777 go into the Windows .def files as well as a number of fixes and code
6778 cleanup (among others, algorithm keywords are now sorted
6779 lexicographically to avoid constant rewrites).
6782 *) In BN_div() keep a copy of the sign of 'num' before writing the
6783 result to 'rm' because if rm==num the value will be overwritten
6784 and produce the wrong result if 'num' is negative: this caused
6785 problems with BN_mod() and BN_nnmod().
6788 *) Function OCSP_request_verify(). This checks the signature on an
6789 OCSP request and verifies the signer certificate. The signer
6790 certificate is just checked for a generic purpose and OCSP request
6794 *) Add OCSP_check_validity() function to check the validity of OCSP
6795 responses. OCSP responses are prepared in real time and may only
6796 be a few seconds old. Simply checking that the current time lies
6797 between thisUpdate and nextUpdate max reject otherwise valid responses
6798 caused by either OCSP responder or client clock inaccuracy. Instead
6799 we allow thisUpdate and nextUpdate to fall within a certain period of
6800 the current time. The age of the response can also optionally be
6801 checked. Two new options -validity_period and -status_age added to
6805 *) If signature or public key algorithm is unrecognized print out its
6806 OID rather that just UNKNOWN.
6809 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6810 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6811 ID to be generated from the issuer certificate alone which can then be
6812 passed to OCSP_id_issuer_cmp().
6815 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6816 ASN1 modules to export functions returning ASN1_ITEM pointers
6817 instead of the ASN1_ITEM structures themselves. This adds several
6818 new macros which allow the underlying ASN1 function/structure to
6819 be accessed transparently. As a result code should not use ASN1_ITEM
6820 references directly (such as &X509_it) but instead use the relevant
6821 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6822 use of the new ASN1 code on platforms where exporting structures
6823 is problematical (for example in shared libraries) but exporting
6824 functions returning pointers to structures is not.
6827 *) Add support for overriding the generation of SSL/TLS session IDs.
6828 These callbacks can be registered either in an SSL_CTX or per SSL.
6829 The purpose of this is to allow applications to control, if they wish,
6830 the arbitrary values chosen for use as session IDs, particularly as it
6831 can be useful for session caching in multiple-server environments. A
6832 command-line switch for testing this (and any client code that wishes
6833 to use such a feature) has been added to "s_server".
6834 [Geoff Thorpe, Lutz Jaenicke]
6836 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6837 of the form '#if defined(...) || defined(...) || ...' and
6838 '#if !defined(...) && !defined(...) && ...'. This also avoids
6839 the growing number of special cases it was previously handling.
6842 *) Make all configuration macros available for application by making
6843 sure they are available in opensslconf.h, by giving them names starting
6844 with "OPENSSL_" to avoid conflicts with other packages and by making
6845 sure e_os2.h will cover all platform-specific cases together with
6847 Additionally, it is now possible to define configuration/platform-
6848 specific names (called "system identities"). In the C code, these
6849 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6850 macro with the name beginning with "OPENSSL_SYS_", which is determined
6851 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6855 *) New option -set_serial to 'req' and 'x509' this allows the serial
6856 number to use to be specified on the command line. Previously self
6857 signed certificates were hard coded with serial number 0 and the
6858 CA options of 'x509' had to use a serial number in a file which was
6862 *) New options to 'ca' utility to support V2 CRL entry extensions.
6863 Currently CRL reason, invalidity date and hold instruction are
6864 supported. Add new CRL extensions to V3 code and some new objects.
6867 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6868 disable standard block padding (aka PKCS#5 padding) in the EVP
6869 API, which was previously mandatory. This means that the data is
6870 not padded in any way and so the total length much be a multiple
6871 of the block size, otherwise an error occurs.
6874 *) Initial (incomplete) OCSP SSL support.
6877 *) New function OCSP_parse_url(). This splits up a URL into its host,
6878 port and path components: primarily to parse OCSP URLs. New -url
6879 option to ocsp utility.
6882 *) New nonce behavior. The return value of OCSP_check_nonce() now
6883 reflects the various checks performed. Applications can decide
6884 whether to tolerate certain situations such as an absent nonce
6885 in a response when one was present in a request: the ocsp application
6886 just prints out a warning. New function OCSP_add1_basic_nonce()
6887 this is to allow responders to include a nonce in a response even if
6888 the request is nonce-less.
6891 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6892 skipped when using openssl x509 multiple times on a single input file,
6893 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6896 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6897 set string type: to handle setting ASN1_TIME structures. Fix ca
6898 utility to correctly initialize revocation date of CRLs.
6901 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6902 the clients preferred ciphersuites and rather use its own preferences.
6903 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6904 Internet Explorer by ensuring unchanged hash method during stepup.
6905 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6908 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6909 to aes and add a new 'exist' option to print out symbols that don't
6913 *) Additional options to ocsp utility to allow flags to be set and
6914 additional certificates supplied.
6917 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6918 OCSP client a number of certificate to only verify the response
6922 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6923 handle the new API. Currently only ECB, CBC modes supported. Add new
6926 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6927 Encryption Standard (AES) Ciphersuites for Transport Layer
6928 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6929 not enabled by default and were not part of the "ALL" ciphersuite
6930 alias because they were not yet official; they could be
6931 explicitly requested by specifying the "AESdraft" ciphersuite
6932 group alias. In the final release of OpenSSL 0.9.7, the group
6933 alias is called "AES" and is part of "ALL".)
6934 [Ben Laurie, Steve Henson, Bodo Moeller]
6936 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6937 request to response.
6940 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6941 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6942 extract information from a certificate request. OCSP_response_create()
6943 creates a response and optionally adds a basic response structure.
6944 OCSP_basic_add1_status() adds a complete single response to a basic
6945 response and returns the OCSP_SINGLERESP structure just added (to allow
6946 extensions to be included for example). OCSP_basic_add1_cert() adds a
6947 certificate to a basic response and OCSP_basic_sign() signs a basic
6948 response with various flags. New helper functions ASN1_TIME_check()
6949 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6950 (converts ASN1_TIME to GeneralizedTime).
6953 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6954 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6955 structure from a certificate. X509_pubkey_digest() digests the public_key
6956 contents: this is used in various key identifiers.
6959 *) Make sk_sort() tolerate a NULL argument.
6960 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6962 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6963 passed by the function are trusted implicitly. If any of them signed the
6964 response then it is assumed to be valid and is not verified.
6967 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6968 to data. This was previously part of the PKCS7 ASN1 code. This
6969 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6970 [Steve Henson, reported by Kenneth R. Robinette
6971 <support@securenetterm.com>]
6973 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6974 routines: without these tracing memory leaks is very painful.
6975 Fix leaks in PKCS12 and PKCS7 routines.
6978 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6979 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6980 effectively meant GeneralizedTime would never be used. Now it
6981 is initialised to -1 but X509_time_adj() now has to check the value
6982 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6983 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6984 [Steve Henson, reported by Kenneth R. Robinette
6985 <support@securenetterm.com>]
6987 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6988 result in a zero length in the ASN1_INTEGER structure which was
6989 not consistent with the structure when d2i_ASN1_INTEGER() was used
6990 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6991 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6992 where it did not print out a minus for negative ASN1_INTEGER.
6995 *) Add summary printout to ocsp utility. The various functions which
6996 convert status values to strings have been renamed to:
6997 OCSP_response_status_str(), OCSP_cert_status_str() and
6998 OCSP_crl_reason_str() and are no longer static. New options
6999 to verify nonce values and to disable verification. OCSP response
7000 printout format cleaned up.
7003 *) Add additional OCSP certificate checks. These are those specified
7004 in RFC2560. This consists of two separate checks: the CA of the
7005 certificate being checked must either be the OCSP signer certificate
7006 or the issuer of the OCSP signer certificate. In the latter case the
7007 OCSP signer certificate must contain the OCSP signing extended key
7008 usage. This check is performed by attempting to match the OCSP
7009 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7010 in the OCSP_CERTID structures of the response.
7013 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7014 and related routines. This uses the standard OpenSSL certificate
7015 verify routines to perform initial checks (just CA validity) and
7016 to obtain the certificate chain. Then additional checks will be
7017 performed on the chain. Currently the root CA is checked to see
7018 if it is explicitly trusted for OCSP signing. This is used to set
7019 a root CA as a global signing root: that is any certificate that
7020 chains to that CA is an acceptable OCSP signing certificate.
7023 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7024 extensions from a separate configuration file.
7025 As when reading extensions from the main configuration file,
7026 the '-extensions ...' option may be used for specifying the
7028 [Massimiliano Pala <madwolf@comune.modena.it>]
7030 *) New OCSP utility. Allows OCSP requests to be generated or
7031 read. The request can be sent to a responder and the output
7032 parsed, outputed or printed in text form. Not complete yet:
7033 still needs to check the OCSP response validity.
7036 *) New subcommands for 'openssl ca':
7037 'openssl ca -status <serial>' prints the status of the cert with
7038 the given serial number (according to the index file).
7039 'openssl ca -updatedb' updates the expiry status of certificates
7041 [Massimiliano Pala <madwolf@comune.modena.it>]
7043 *) New '-newreq-nodes' command option to CA.pl. This is like
7044 '-newreq', but calls 'openssl req' with the '-nodes' option
7045 so that the resulting key is not encrypted.
7046 [Damien Miller <djm@mindrot.org>]
7048 *) New configuration for the GNU Hurd.
7049 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7051 *) Initial code to implement OCSP basic response verify. This
7052 is currently incomplete. Currently just finds the signer's
7053 certificate and verifies the signature on the response.
7056 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7057 value of OPENSSLDIR. This is available via the new '-d' option
7058 to 'openssl version', and is also included in 'openssl version -a'.
7061 *) Allowing defining memory allocation callbacks that will be given
7062 file name and line number information in additional arguments
7063 (a const char* and an int). The basic functionality remains, as
7064 well as the original possibility to just replace malloc(),
7065 realloc() and free() by functions that do not know about these
7066 additional arguments. To register and find out the current
7067 settings for extended allocation functions, the following
7068 functions are provided:
7070 CRYPTO_set_mem_ex_functions
7071 CRYPTO_set_locked_mem_ex_functions
7072 CRYPTO_get_mem_ex_functions
7073 CRYPTO_get_locked_mem_ex_functions
7075 These work the same way as CRYPTO_set_mem_functions and friends.
7076 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7077 extended allocation function is enabled.
7078 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7079 a conventional allocation function is enabled.
7080 [Richard Levitte, Bodo Moeller]
7082 *) Finish off removing the remaining LHASH function pointer casts.
7083 There should no longer be any prototype-casting required when using
7084 the LHASH abstraction, and any casts that remain are "bugs". See
7085 the callback types and macros at the head of lhash.h for details
7086 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7089 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7090 If /dev/[u]random devices are not available or do not return enough
7091 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7093 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7094 /etc/entropy will be queried once each in this sequence, quering stops
7095 when enough entropy was collected without querying more sockets.
7098 *) Change the Unix RAND_poll() variant to be able to poll several
7099 random devices, as specified by DEVRANDOM, until a sufficient amount
7100 of data has been collected. We spend at most 10 ms on each file
7101 (select timeout) and read in non-blocking mode. DEVRANDOM now
7102 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7103 (previously it was just the string "/dev/urandom"), so on typical
7104 platforms the 10 ms delay will never occur.
7105 Also separate out the Unix variant to its own file, rand_unix.c.
7106 For VMS, there's a currently-empty rand_vms.c.
7109 *) Move OCSP client related routines to ocsp_cl.c. These
7110 provide utility functions which an application needing
7111 to issue a request to an OCSP responder and analyse the
7112 response will typically need: as opposed to those which an
7113 OCSP responder itself would need which will be added later.
7115 OCSP_request_sign() signs an OCSP request with an API similar
7116 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7117 response. OCSP_response_get1_basic() extracts basic response
7118 from response. OCSP_resp_find_status(): finds and extracts status
7119 information from an OCSP_CERTID structure (which will be created
7120 when the request structure is built). These are built from lower
7121 level functions which work on OCSP_SINGLERESP structures but
7122 wont normally be used unless the application wishes to examine
7123 extensions in the OCSP response for example.
7125 Replace nonce routines with a pair of functions.
7126 OCSP_request_add1_nonce() adds a nonce value and optionally
7127 generates a random value. OCSP_check_nonce() checks the
7128 validity of the nonce in an OCSP response.
7131 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7132 This doesn't copy the supplied OCSP_CERTID and avoids the
7133 need to free up the newly created id. Change return type
7134 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7135 This can then be used to add extensions to the request.
7136 Deleted OCSP_request_new(), since most of its functionality
7137 is now in OCSP_REQUEST_new() (and the case insensitive name
7138 clash) apart from the ability to set the request name which
7139 will be added elsewhere.
7142 *) Update OCSP API. Remove obsolete extensions argument from
7143 various functions. Extensions are now handled using the new
7144 OCSP extension code. New simple OCSP HTTP function which
7145 can be used to send requests and parse the response.
7148 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7149 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7150 uses the special reorder version of SET OF to sort the attributes
7151 and reorder them to match the encoded order. This resolves a long
7152 standing problem: a verify on a PKCS7 structure just after signing
7153 it used to fail because the attribute order did not match the
7154 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7155 it uses the received order. This is necessary to tolerate some broken
7156 software that does not order SET OF. This is handled by encoding
7157 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7158 to produce the required SET OF.
7161 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7162 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7163 files to get correct declarations of the ASN.1 item variables.
7166 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7167 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7168 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7169 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7170 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7171 ASN1_ITEM and no wrapper functions.
7174 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7175 replace the old function pointer based I/O routines. Change most of
7176 the *_d2i_bio() and *_d2i_fp() functions to use these.
7179 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7180 lines, recognice more "algorithms" that can be deselected, and make
7181 it complain about algorithm deselection that isn't recognised.
7184 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7185 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7186 to use new functions. Add NO_ASN1_OLD which can be set to remove
7187 some old style ASN1 functions: this can be used to determine if old
7188 code will still work when these eventually go away.
7191 *) New extension functions for OCSP structures, these follow the
7192 same conventions as certificates and CRLs.
7195 *) New function X509V3_add1_i2d(). This automatically encodes and
7196 adds an extension. Its behaviour can be customised with various
7197 flags to append, replace or delete. Various wrappers added for
7198 certifcates and CRLs.
7201 *) Fix to avoid calling the underlying ASN1 print routine when
7202 an extension cannot be parsed. Correct a typo in the
7203 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7206 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7207 entries for variables.
7210 *) Add functionality to apps/openssl.c for detecting locking
7211 problems: As the program is single-threaded, all we have
7212 to do is register a locking callback using an array for
7213 storing which locks are currently held by the program.
7216 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7217 SSL_get_ex_data_X509_STORE_idx(), which is used in
7218 ssl_verify_cert_chain() and thus can be called at any time
7219 during TLS/SSL handshakes so that thread-safety is essential.
7220 Unfortunately, the ex_data design is not at all suited
7221 for multi-threaded use, so it probably should be abolished.
7224 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7225 [Broadcom, tweaked and integrated by Geoff Thorpe]
7227 *) Move common extension printing code to new function
7228 X509V3_print_extensions(). Reorganise OCSP print routines and
7229 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7232 *) New function X509_signature_print() to remove duplication in some
7236 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7237 set (this was treated exactly the same as SET OF previously). This
7238 is used to reorder the STACK representing the structure to match the
7239 encoding. This will be used to get round a problem where a PKCS7
7240 structure which was signed could not be verified because the STACK
7241 order did not reflect the encoded order.
7244 *) Reimplement the OCSP ASN1 module using the new code.
7247 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7248 for its ASN1 operations. The old style function pointers still exist
7249 for now but they will eventually go away.
7252 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7253 completely replaces the old ASN1 functionality with a table driven
7254 encoder and decoder which interprets an ASN1_ITEM structure describing
7255 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7256 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7257 has also been converted to the new form.
7260 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7261 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7262 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7263 for negative moduli.
7266 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7267 of not touching the result's sign bit.
7270 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7274 *) Changed the LHASH code to use prototypes for callbacks, and created
7275 macros to declare and implement thin (optionally static) functions
7276 that provide type-safety and avoid function pointer casting for the
7277 type-specific callbacks.
7280 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7282 [Veers Staats <staatsvr@asc.hpc.mil>,
7283 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7285 *) Reformat the FAQ so the different questions and answers can be divided
7286 in sections depending on the subject.
7289 *) Have the zlib compression code load ZLIB.DLL dynamically under
7293 *) New function BN_mod_sqrt for computing square roots modulo a prime
7294 (using the probabilistic Tonelli-Shanks algorithm unless
7295 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7296 be handled deterministically).
7297 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7299 *) Make BN_mod_inverse faster by explicitly handling small quotients
7300 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7301 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7304 *) New function BN_kronecker.
7307 *) Fix BN_gcd so that it works on negative inputs; the result is
7308 positive unless both parameters are zero.
7309 Previously something reasonably close to an infinite loop was
7310 possible because numbers could be growing instead of shrinking
7311 in the implementation of Euclid's algorithm.
7314 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7315 sign of the number in question.
7317 Fix BN_is_word(a,w) to work correctly for w == 0.
7319 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7320 because its test if the absolute value of 'a' equals 'w'.
7321 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7322 it exists mostly for use in the implementations of BN_is_zero(),
7323 BN_is_one(), and BN_is_word().
7326 *) New function BN_swap.
7329 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7330 the exponentiation functions are more likely to produce reasonable
7331 results on negative inputs.
7334 *) Change BN_mod_mul so that the result is always non-negative.
7335 Previously, it could be negative if one of the factors was negative;
7336 I don't think anyone really wanted that behaviour.
7339 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7340 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7341 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7342 and add new functions:
7351 BN_mod_lshift1_quick
7355 These functions always generate non-negative results.
7357 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7358 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7360 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7361 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7362 be reduced modulo m.
7363 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7366 The following entry accidentally appeared in the CHANGES file
7367 distributed with OpenSSL 0.9.7. The modifications described in
7368 it do *not* apply to OpenSSL 0.9.7.
7370 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7371 was actually never needed) and in BN_mul(). The removal in BN_mul()
7372 required a small change in bn_mul_part_recursive() and the addition
7373 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7374 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7375 bn_sub_words() and bn_add_words() except they take arrays with
7380 *) In 'openssl passwd', verify passwords read from the terminal
7381 unless the '-salt' option is used (which usually means that
7382 verification would just waste user's time since the resulting
7383 hash is going to be compared with some given password hash)
7384 or the new '-noverify' option is used.
7386 This is an incompatible change, but it does not affect
7387 non-interactive use of 'openssl passwd' (passwords on the command
7388 line, '-stdin' option, '-in ...' option) and thus should not
7392 *) Remove all references to RSAref, since there's no more need for it.
7395 *) Make DSO load along a path given through an environment variable
7396 (SHLIB_PATH) with shl_load().
7399 *) Constify the ENGINE code as a result of BIGNUM constification.
7400 Also constify the RSA code and most things related to it. In a
7401 few places, most notable in the depth of the ASN.1 code, ugly
7402 casts back to non-const were required (to be solved at a later
7406 *) Make it so the openssl application has all engines loaded by default.
7409 *) Constify the BIGNUM routines a little more.
7412 *) Add the following functions:
7414 ENGINE_load_cswift()
7416 ENGINE_load_atalla()
7418 ENGINE_load_builtin_engines()
7420 That way, an application can itself choose if external engines that
7421 are built-in in OpenSSL shall ever be used or not. The benefit is
7422 that applications won't have to be linked with libdl or other dso
7423 libraries unless it's really needed.
7425 Changed 'openssl engine' to load all engines on demand.
7426 Changed the engine header files to avoid the duplication of some
7427 declarations (they differed!).
7430 *) 'openssl engine' can now list capabilities.
7433 *) Better error reporting in 'openssl engine'.
7436 *) Never call load_dh_param(NULL) in s_server.
7439 *) Add engine application. It can currently list engines by name and
7440 identity, and test if they are actually available.
7443 *) Improve RPM specification file by forcing symbolic linking and making
7444 sure the installed documentation is also owned by root.root.
7445 [Damien Miller <djm@mindrot.org>]
7447 *) Give the OpenSSL applications more possibilities to make use of
7448 keys (public as well as private) handled by engines.
7451 *) Add OCSP code that comes from CertCo.
7454 *) Add VMS support for the Rijndael code.
7457 *) Added untested support for Nuron crypto accelerator.
7460 *) Add support for external cryptographic devices. This code was
7461 previously distributed separately as the "engine" branch.
7462 [Geoff Thorpe, Richard Levitte]
7464 *) Rework the filename-translation in the DSO code. It is now possible to
7465 have far greater control over how a "name" is turned into a filename
7466 depending on the operating environment and any oddities about the
7467 different shared library filenames on each system.
7470 *) Support threads on FreeBSD-elf in Configure.
7473 *) Fix for SHA1 assembly problem with MASM: it produces
7474 warnings about corrupt line number information when assembling
7475 with debugging information. This is caused by the overlapping
7477 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7480 NCONF_get_number() has no error checking at all. As a replacement,
7481 NCONF_get_number_e() is defined (_e for "error checking") and is
7482 promoted strongly. The old NCONF_get_number is kept around for
7483 binary backward compatibility.
7484 Make it possible for methods to load from something other than a BIO,
7485 by providing a function pointer that is given a name instead of a BIO.
7486 For example, this could be used to load configuration data from an
7490 *) Fix for non blocking accept BIOs. Added new I/O special reason
7491 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7492 with non blocking I/O was not possible because no retry code was
7493 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7497 *) Added the beginnings of Rijndael support.
7500 *) Fix for bug in DirectoryString mask setting. Add support for
7501 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7502 to allow certificate printing to more controllable, additional
7503 'certopt' option to 'x509' to allow new printing options to be
7507 *) Clean old EAY MD5 hack from e_os.h.
7510 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7512 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7513 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7514 [Joe Orton, Steve Henson]
7516 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7518 *) Fix additional bug revealed by the NISCC test suite:
7520 Stop bug triggering large recursion when presented with
7521 certain ASN.1 tags (CVE-2003-0851)
7524 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7526 *) Fix various bugs revealed by running the NISCC test suite:
7528 Stop out of bounds reads in the ASN1 code when presented with
7529 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7531 If verify callback ignores invalid public key errors don't try to check
7532 certificate signature with the NULL public key.
7536 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7537 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7541 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7542 extra data after the compression methods not only for TLS 1.0
7543 but also for SSL 3.0 (as required by the specification).
7544 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7546 *) Change X509_certificate_type() to mark the key as exported/exportable
7547 when it's 512 *bits* long, not 512 bytes.
7550 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7552 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7553 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7554 a protocol version number mismatch like a decryption error
7555 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7558 *) Turn on RSA blinding by default in the default implementation
7559 to avoid a timing attack. Applications that don't want it can call
7560 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7561 They would be ill-advised to do so in most cases.
7562 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7564 *) Change RSA blinding code so that it works when the PRNG is not
7565 seeded (in this case, the secret RSA exponent is abused as
7566 an unpredictable seed -- if it is not unpredictable, there
7567 is no point in blinding anyway). Make RSA blinding thread-safe
7568 by remembering the creator's thread ID in rsa->blinding and
7569 having all other threads use local one-time blinding factors
7570 (this requires more computation than sharing rsa->blinding, but
7571 avoids excessive locking; and if an RSA object is not shared
7572 between threads, blinding will still be very fast).
7575 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7577 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7578 via timing by performing a MAC computation even if incorrrect
7579 block cipher padding has been found. This is a countermeasure
7580 against active attacks where the attacker has to distinguish
7581 between bad padding and a MAC verification error. (CVE-2003-0078)
7583 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7584 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7585 Martin Vuagnoux (EPFL, Ilion)]
7587 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7589 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7590 memory from it's contents. This is done with a counter that will
7591 place alternating values in each byte. This can be used to solve
7592 two issues: 1) the removal of calls to memset() by highly optimizing
7593 compilers, and 2) cleansing with other values than 0, since those can
7594 be read through on certain media, for example a swap space on disk.
7597 *) Bugfix: client side session caching did not work with external caching,
7598 because the session->cipher setting was not restored when reloading
7599 from the external cache. This problem was masked, when
7600 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7601 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7604 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7605 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7606 [Zeev Lieber <zeev-l@yahoo.com>]
7608 *) Undo an undocumented change introduced in 0.9.6e which caused
7609 repeated calls to OpenSSL_add_all_ciphers() and
7610 OpenSSL_add_all_digests() to be ignored, even after calling
7614 *) Change the default configuration reader to deal with last line not
7615 being properly terminated.
7618 *) Change X509_NAME_cmp() so it applies the special rules on handling
7619 DN values that are of type PrintableString, as well as RDNs of type
7620 emailAddress where the value has the type ia5String.
7621 [stefank@valicert.com via Richard Levitte]
7623 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7624 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7625 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7626 the bitwise-OR of the two for use by the majority of applications
7627 wanting this behaviour, and update the docs. The documented
7628 behaviour and actual behaviour were inconsistent and had been
7629 changing anyway, so this is more a bug-fix than a behavioural
7631 [Geoff Thorpe, diagnosed by Nadav Har'El]
7633 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7634 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7637 *) Fix initialization code race conditions in
7638 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7639 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7640 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7641 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7642 ssl2_get_cipher_by_char(),
7643 ssl3_get_cipher_by_char().
7644 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7646 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7647 the cached sessions are flushed, as the remove_cb() might use ex_data
7648 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7649 (see [openssl.org #212]).
7650 [Geoff Thorpe, Lutz Jaenicke]
7652 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7653 length, instead of the encoding length to d2i_ASN1_OBJECT.
7656 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7658 *) [In 0.9.6g-engine release:]
7659 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7660 [Lynn Gazis <lgazis@rainbow.com>]
7662 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7664 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7665 and get fix the header length calculation.
7666 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7667 Alon Kantor <alonk@checkpoint.com> (and others),
7670 *) Use proper error handling instead of 'assertions' in buffer
7671 overflow checks added in 0.9.6e. This prevents DoS (the
7672 assertions could call abort()).
7673 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7675 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7677 *) Add various sanity checks to asn1_get_length() to reject
7678 the ASN1 length bytes if they exceed sizeof(long), will appear
7679 negative or the content length exceeds the length of the
7681 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7683 *) Fix cipher selection routines: ciphers without encryption had no flags
7684 for the cipher strength set and where therefore not handled correctly
7685 by the selection routines (PR #130).
7688 *) Fix EVP_dsa_sha macro.
7692 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7693 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7694 that was added in OpenSSL 0.9.6d.
7696 As the countermeasure turned out to be incompatible with some
7697 broken SSL implementations, the new option is part of SSL_OP_ALL.
7698 SSL_OP_ALL is usually employed when compatibility with weird SSL
7699 implementations is desired (e.g. '-bugs' option to 's_client' and
7700 's_server'), so the new option is automatically set in many
7704 *) Changes in security patch:
7706 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7707 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7708 Air Force Materiel Command, USAF, under agreement number
7711 *) Add various sanity checks to asn1_get_length() to reject
7712 the ASN1 length bytes if they exceed sizeof(long), will appear
7713 negative or the content length exceeds the length of the
7714 supplied buffer. (CVE-2002-0659)
7715 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7717 *) Assertions for various potential buffer overflows, not known to
7719 [Ben Laurie (CHATS)]
7721 *) Various temporary buffers to hold ASCII versions of integers were
7722 too small for 64 bit platforms. (CVE-2002-0655)
7723 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7725 *) Remote buffer overflow in SSL3 protocol - an attacker could
7726 supply an oversized session ID to a client. (CVE-2002-0656)
7727 [Ben Laurie (CHATS)]
7729 *) Remote buffer overflow in SSL2 protocol - an attacker could
7730 supply an oversized client master key. (CVE-2002-0656)
7731 [Ben Laurie (CHATS)]
7733 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7735 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7736 encoded as NULL) with id-dsa-with-sha1.
7737 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7739 *) Check various X509_...() return values in apps/req.c.
7740 [Nils Larsch <nla@trustcenter.de>]
7742 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7743 an end-of-file condition would erronously be flagged, when the CRLF
7744 was just at the end of a processed block. The bug was discovered when
7745 processing data through a buffering memory BIO handing the data to a
7746 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7747 <ptsekov@syntrex.com> and Nedelcho Stanev.
7750 *) Implement a countermeasure against a vulnerability recently found
7751 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7752 before application data chunks to avoid the use of known IVs
7753 with data potentially chosen by the attacker.
7756 *) Fix length checks in ssl3_get_client_hello().
7759 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7760 to prevent ssl3_read_internal() from incorrectly assuming that
7761 ssl3_read_bytes() found application data while handshake
7762 processing was enabled when in fact s->s3->in_read_app_data was
7763 merely automatically cleared during the initial handshake.
7764 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7766 *) Fix object definitions for Private and Enterprise: they were not
7767 recognized in their shortname (=lowercase) representation. Extend
7768 obj_dat.pl to issue an error when using undefined keywords instead
7769 of silently ignoring the problem (Svenning Sorensen
7770 <sss@sss.dnsalias.net>).
7773 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7774 generators, i.e. generators other than 2 and 5. (Previously, the
7775 code did not properly initialise the 'add' and 'rem' values to
7776 BN_generate_prime().)
7778 In the new general case, we do not insist that 'generator' is
7779 actually a primitive root: This requirement is rather pointless;
7780 a generator of the order-q subgroup is just as good, if not
7784 *) Map new X509 verification errors to alerts. Discovered and submitted by
7785 Tom Wu <tom@arcot.com>.
7788 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7789 returning non-zero before the data has been completely received
7790 when using non-blocking I/O.
7791 [Bodo Moeller; problem pointed out by John Hughes]
7793 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7794 [Ben Laurie, Lutz Jaenicke]
7796 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7797 Yoram Zahavi <YoramZ@gilian.com>).
7800 *) Add information about CygWin 1.3 and on, and preserve proper
7801 configuration for the versions before that.
7802 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7804 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7805 check whether we deal with a copy of a session and do not delete from
7806 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7807 <izhar@checkpoint.com>.
7810 *) Do not store session data into the internal session cache, if it
7811 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7812 flag is set). Proposed by Aslam <aslam@funk.com>.
7815 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7819 *) [In 0.9.6d-engine release:]
7820 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7821 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7823 *) Add the configuration target linux-s390x.
7824 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7826 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7827 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7828 variable as an indication that a ClientHello message has been
7829 received. As the flag value will be lost between multiple
7830 invocations of ssl3_accept when using non-blocking I/O, the
7831 function may not be aware that a handshake has actually taken
7832 place, thus preventing a new session from being added to the
7835 To avoid this problem, we now set s->new_session to 2 instead of
7836 using a local variable.
7837 [Lutz Jaenicke, Bodo Moeller]
7839 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7840 if the SSL_R_LENGTH_MISMATCH error is detected.
7841 [Geoff Thorpe, Bodo Moeller]
7843 *) New 'shared_ldflag' column in Configure platform table.
7846 *) Fix EVP_CIPHER_mode macro.
7847 ["Dan S. Camper" <dan@bti.net>]
7849 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7850 type, we must throw them away by setting rr->length to 0.
7851 [D P Chang <dpc@qualys.com>]
7853 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7855 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7856 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7857 worked incorrectly for those cases where range = 10..._2 and
7858 3*range is two bits longer than range.)
7861 *) Only add signing time to PKCS7 structures if it is not already
7865 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7866 OBJ_ld_ce should be OBJ_id_ce.
7867 Also some ip-pda OIDs in crypto/objects/objects.txt were
7868 incorrect (cf. RFC 3039).
7869 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7871 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7872 returns early because it has nothing to do.
7873 [Andy Schneider <andy.schneider@bjss.co.uk>]
7875 *) [In 0.9.6c-engine release:]
7876 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7877 [Andy Schneider <andy.schneider@bjss.co.uk>]
7879 *) [In 0.9.6c-engine release:]
7880 Add support for Cryptographic Appliance's keyserver technology.
7881 (Use engine 'keyclient')
7882 [Cryptographic Appliances and Geoff Thorpe]
7884 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7885 is called via tools/c89.sh because arguments have to be
7886 rearranged (all '-L' options must appear before the first object
7888 [Richard Shapiro <rshapiro@abinitio.com>]
7890 *) [In 0.9.6c-engine release:]
7891 Add support for Broadcom crypto accelerator cards, backported
7893 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7895 *) [In 0.9.6c-engine release:]
7896 Add support for SureWare crypto accelerator cards from
7897 Baltimore Technologies. (Use engine 'sureware')
7898 [Baltimore Technologies and Mark Cox]
7900 *) [In 0.9.6c-engine release:]
7901 Add support for crypto accelerator cards from Accelerated
7902 Encryption Processing, www.aep.ie. (Use engine 'aep')
7903 [AEP Inc. and Mark Cox]
7905 *) Add a configuration entry for gcc on UnixWare.
7906 [Gary Benson <gbenson@redhat.com>]
7908 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7909 messages are stored in a single piece (fixed-length part and
7910 variable-length part combined) and fix various bugs found on the way.
7913 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7914 instead. BIO_gethostbyname() does not know what timeouts are
7915 appropriate, so entries would stay in cache even when they have
7917 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7919 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7920 faced with a pathologically small ClientHello fragment that does
7921 not contain client_version: Instead of aborting with an error,
7922 simply choose the highest available protocol version (i.e.,
7923 TLS 1.0 unless it is disabled). In practice, ClientHello
7924 messages are never sent like this, but this change gives us
7925 strictly correct behaviour at least for TLS.
7928 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7929 never resets s->method to s->ctx->method when called from within
7930 one of the SSL handshake functions.
7931 [Bodo Moeller; problem pointed out by Niko Baric]
7933 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7934 (sent using the client's version number) if client_version is
7935 smaller than the protocol version in use. Also change
7936 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7937 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7938 the client will at least see that alert.
7941 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7945 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7946 client receives HelloRequest while in a handshake.
7947 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7949 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7950 should end in 'break', not 'goto end' which circumvents various
7951 cleanups done in state SSL_ST_OK. But session related stuff
7952 must be disabled for SSL_ST_OK in the case that we just sent a
7955 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7956 before just sending a HelloRequest.
7957 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7959 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7960 reveal whether illegal block cipher padding was found or a MAC
7961 verification error occurred. (Neither SSLerr() codes nor alerts
7962 are directly visible to potential attackers, but the information
7963 may leak via logfiles.)
7965 Similar changes are not required for the SSL 2.0 implementation
7966 because the number of padding bytes is sent in clear for SSL 2.0,
7967 and the extra bytes are just ignored. However ssl/s2_pkt.c
7968 failed to verify that the purported number of padding bytes is in
7972 *) Add OpenUNIX-8 support including shared libraries
7973 (Boyd Lynn Gerber <gerberb@zenez.com>).
7976 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7977 'wristwatch attack' using huge encoding parameters (cf.
7978 James H. Manger's CRYPTO 2001 paper). Note that the
7979 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7980 encoding parameters and hence was not vulnerable.
7983 *) BN_sqr() bug fix.
7984 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7986 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7987 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7988 followed by modular reduction.
7989 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7991 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7992 equivalent based on BN_pseudo_rand() instead of BN_rand().
7995 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7996 This function was broken, as the check for a new client hello message
7997 to handle SGC did not allow these large messages.
7998 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8001 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8004 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8005 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8008 *) Rework the configuration and shared library support for Tru64 Unix.
8009 The configuration part makes use of modern compiler features and
8010 still retains old compiler behavior for those that run older versions
8011 of the OS. The shared library support part includes a variant that
8012 uses the RPATH feature, and is available through the special
8013 configuration target "alpha-cc-rpath", which will never be selected
8015 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8017 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8018 with the same message size as in ssl3_get_certificate_request().
8019 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8020 messages might inadvertently be reject as too long.
8021 [Petr Lampa <lampa@fee.vutbr.cz>]
8023 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8026 *) Modified SSL library such that the verify_callback that has been set
8027 specificly for an SSL object with SSL_set_verify() is actually being
8028 used. Before the change, a verify_callback set with this function was
8029 ignored and the verify_callback() set in the SSL_CTX at the time of
8030 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8031 to allow the necessary settings.
8034 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8035 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8036 done automatically (in contradiction to the requirements of the C
8037 standard). This made problems when used from OpenSSH.
8040 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8041 dh->length and always used
8043 BN_rand_range(priv_key, dh->p).
8045 BN_rand_range() is not necessary for Diffie-Hellman, and this
8046 specific range makes Diffie-Hellman unnecessarily inefficient if
8047 dh->length (recommended exponent length) is much smaller than the
8048 length of dh->p. We could use BN_rand_range() if the order of
8049 the subgroup was stored in the DH structure, but we only have
8054 BN_rand(priv_key, l, ...)
8056 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8062 RSA_eay_public_encrypt
8063 RSA_eay_private_decrypt
8064 RSA_eay_private_encrypt (signing)
8065 RSA_eay_public_decrypt (signature verification)
8067 (default implementations for RSA_public_encrypt,
8068 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8069 always reject numbers >= n.
8072 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8073 to synchronize access to 'locking_thread'. This is necessary on
8074 systems where access to 'locking_thread' (an 'unsigned long'
8075 variable) is not atomic.
8078 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8079 *before* setting the 'crypto_lock_rand' flag. The previous code had
8080 a race condition if 0 is a valid thread ID.
8081 [Travis Vitek <vitek@roguewave.com>]
8083 *) Add support for shared libraries under Irix.
8084 [Albert Chin-A-Young <china@thewrittenword.com>]
8086 *) Add configuration option to build on Linux on both big-endian and
8088 [Ralf Baechle <ralf@uni-koblenz.de>]
8090 *) Add the possibility to create shared libraries on HP-UX.
8093 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8095 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8096 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8097 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8098 PRNG state recovery was possible based on the output of
8099 one PRNG request appropriately sized to gain knowledge on
8100 'md' followed by enough consecutive 1-byte PRNG requests
8101 to traverse all of 'state'.
8103 1. When updating 'md_local' (the current thread's copy of 'md')
8104 during PRNG output generation, hash all of the previous
8105 'md_local' value, not just the half used for PRNG output.
8107 2. Make the number of bytes from 'state' included into the hash
8108 independent from the number of PRNG bytes requested.
8110 The first measure alone would be sufficient to avoid
8111 Markku-Juhani's attack. (Actually it had never occurred
8112 to me that the half of 'md_local' used for chaining was the
8113 half from which PRNG output bytes were taken -- I had always
8114 assumed that the secret half would be used.) The second
8115 measure makes sure that additional data from 'state' is never
8116 mixed into 'md_local' in small portions; this heuristically
8117 further strengthens the PRNG.
8120 *) Fix crypto/bn/asm/mips3.s.
8123 *) When only the key is given to "enc", the IV is undefined. Print out
8124 an error message in this case.
8127 *) Handle special case when X509_NAME is empty in X509 printing routines.
8130 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8131 positive and less than q.
8134 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8135 used: it isn't thread safe and the add_lock_callback should handle
8137 [Paul Rose <Paul.Rose@bridge.com>]
8139 *) Verify that incoming data obeys the block size in
8140 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8144 [Ulf Möller, Bodo Möller]
8146 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8147 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8148 when fixing the server behaviour for backwards-compatible 'client
8149 hello' messages. (Note that the attack is impractical against
8150 SSL 3.0 and TLS 1.0 anyway because length and version checking
8151 means that the probability of guessing a valid ciphertext is
8152 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8155 Before 0.9.5, the countermeasure (hide the error by generating a
8156 random 'decryption result') did not work properly because
8157 ERR_clear_error() was missing, meaning that SSL_get_error() would
8158 detect the supposedly ignored error.
8160 Both problems are now fixed.
8163 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8164 (previously it was 1024).
8167 *) Fix for compatibility mode trust settings: ignore trust settings
8168 unless some valid trust or reject settings are present.
8171 *) Fix for blowfish EVP: its a variable length cipher.
8174 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8175 parameters in DSA public key structures and return an error in the
8176 DSA routines if parameters are absent.
8179 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8180 in the current directory if neither $RANDFILE nor $HOME was set.
8181 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8182 caused some confusion to Windows users who haven't defined $HOME.
8183 Thus RAND_file_name() is changed again: e_os.h can define a
8184 DEFAULT_HOME, which will be used if $HOME is not set.
8185 For Windows, we use "C:"; on other platforms, we still require
8186 environment variables.
8188 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8189 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8190 having multiple threads call RAND_poll() concurrently.
8193 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8194 combination of a flag and a thread ID variable.
8195 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8196 flag), *other* threads can enter ssleay_add_bytes without obeying
8197 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8198 that they do not hold after the first thread unsets add_do_not_lock).
8201 *) Change bctest again: '-x' expressions are not available in all
8205 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8207 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8208 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8210 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8211 the default extension for executables, if any. Also, make the perl
8212 scripts that use symlink() to test if it really exists and use "cp"
8213 if it doesn't. All this made OpenSSL compilable and installable in
8217 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8218 If SEQUENCE is length is indefinite just set c->slen to the total
8219 amount of data available.
8220 [Steve Henson, reported by shige@FreeBSD.org]
8221 [This change does not apply to 0.9.7.]
8223 *) Change bctest to avoid here-documents inside command substitution
8224 (workaround for FreeBSD /bin/sh bug).
8225 For compatibility with Ultrix, avoid shell functions (introduced
8226 in the bctest version that searches along $PATH).
8229 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8230 with des_encrypt() defined on some operating systems, like Solaris
8234 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8235 On the Importance of Eliminating Errors in Cryptographic
8236 Computations, J. Cryptology 14 (2001) 2, 101-119,
8237 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8240 *) MIPS assembler BIGNUM division bug fix.
8243 *) Disabled incorrect Alpha assembler code.
8246 *) Fix PKCS#7 decode routines so they correctly update the length
8247 after reading an EOC for the EXPLICIT tag.
8249 [This change does not apply to 0.9.7.]
8251 *) Fix bug in PKCS#12 key generation routines. This was triggered
8252 if a 3DES key was generated with a 0 initial byte. Include
8253 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8254 (but broken) behaviour.
8257 *) Enhance bctest to search for a working bc along $PATH and print
8259 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8261 *) Fix memory leaks in err.c: free err_data string if necessary;
8262 don't write to the wrong index in ERR_set_error_data.
8265 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8269 *) Replace rdtsc with _emit statements for VC++ version 5.
8270 [Jeremy Cooper <jeremy@baymoo.org>]
8272 *) Make it possible to reuse SSLv2 sessions.
8275 *) In copy_email() check for >= 0 as a return value for
8276 X509_NAME_get_index_by_NID() since 0 is a valid index.
8277 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8279 *) Avoid coredump with unsupported or invalid public keys by checking if
8280 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8281 PKCS7_verify() fails with non detached data.
8284 *) Don't use getenv in library functions when run as setuid/setgid.
8285 New function OPENSSL_issetugid().
8288 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8289 due to incorrect handling of multi-threading:
8291 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8293 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8295 3. Count how many times MemCheck_off() has been called so that
8296 nested use can be treated correctly. This also avoids
8297 inband-signalling in the previous code (which relied on the
8298 assumption that thread ID 0 is impossible).
8301 *) Add "-rand" option also to s_client and s_server.
8304 *) Fix CPU detection on Irix 6.x.
8305 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8306 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8308 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8311 [This change does not apply to 0.9.7.]
8313 *) Use the cached encoding of an X509_NAME structure rather than
8314 copying it. This is apparently the reason for the libsafe "errors"
8315 but the code is actually correct.
8318 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8319 Bleichenbacher's DSA attack.
8320 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8321 to be set and top=0 forces the highest bit to be set; top=-1 is new
8322 and leaves the highest bit random.
8323 [Ulf Moeller, Bodo Moeller]
8325 *) In the NCONF_...-based implementations for CONF_... queries
8326 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8327 a temporary CONF structure with the data component set to NULL
8328 (which gives segmentation faults in lh_retrieve).
8329 Instead, use NULL for the CONF pointer in CONF_get_string and
8330 CONF_get_number (which may use environment variables) and directly
8331 return NULL from CONF_get_section.
8334 *) Fix potential buffer overrun for EBCDIC.
8337 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8338 keyUsage if basicConstraints absent for a CA.
8341 *) Make SMIME_write_PKCS7() write mail header values with a format that
8342 is more generally accepted (no spaces before the semicolon), since
8343 some programs can't parse those values properly otherwise. Also make
8344 sure BIO's that break lines after each write do not create invalid
8348 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8349 macros previously used would not encode an empty SEQUENCE OF
8350 and break the signature.
8352 [This change does not apply to 0.9.7.]
8354 *) Zero the premaster secret after deriving the master secret in
8358 *) Add some EVP_add_digest_alias registrations (as found in
8359 OpenSSL_add_all_digests()) to SSL_library_init()
8360 aka OpenSSL_add_ssl_algorithms(). This provides improved
8361 compatibility with peers using X.509 certificates
8362 with unconventional AlgorithmIdentifier OIDs.
8365 *) Fix for Irix with NO_ASM.
8366 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8368 *) ./config script fixes.
8369 [Ulf Moeller, Richard Levitte]
8371 *) Fix 'openssl passwd -1'.
8374 *) Change PKCS12_key_gen_asc() so it can cope with non null
8375 terminated strings whose length is passed in the passlen
8376 parameter, for example from PEM callbacks. This was done
8377 by adding an extra length parameter to asc2uni().
8378 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8380 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8381 call failed, free the DSA structure.
8384 *) Fix to uni2asc() to cope with zero length Unicode strings.
8385 These are present in some PKCS#12 files.
8388 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8389 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8390 when writing a 32767 byte record.
8391 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8393 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8394 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8396 (RSA objects have a reference count access to which is protected
8397 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8398 so they are meant to be shared between threads.)
8399 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8400 "Reddie, Steven" <Steven.Reddie@ca.com>]
8402 *) Fix a deadlock in CRYPTO_mem_leaks().
8405 *) Use better test patterns in bntest.
8408 *) rand_win.c fix for Borland C.
8411 *) BN_rshift bugfix for n == 0.
8414 *) Add a 'bctest' script that checks for some known 'bc' bugs
8415 so that 'make test' does not abort just because 'bc' is broken.
8418 *) Store verify_result within SSL_SESSION also for client side to
8419 avoid potential security hole. (Re-used sessions on the client side
8420 always resulted in verify_result==X509_V_OK, not using the original
8421 result of the server certificate verification.)
8424 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8425 SSL3_RT_APPLICATION_DATA, return 0.
8426 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8430 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8431 releases, have been re-implemented by renaming the previous
8432 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8433 and ssl3_read_internal, respectively, and adding 'peek' parameters
8434 to them. The new ssl[23]_{read,peek} functions are calls to
8435 ssl[23]_read_internal with the 'peek' flag set appropriately.
8436 A 'peek' parameter has also been added to ssl3_read_bytes, which
8437 does the actual work for ssl3_read_internal.
8440 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8441 the method-specific "init()" handler. Also clean up ex_data after
8442 calling the method-specific "finish()" handler. Previously, this was
8443 happening the other way round.
8446 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8447 The previous value, 12, was not always sufficient for BN_mod_exp().
8450 *) Make sure that shared libraries get the internal name engine with
8451 the full version number and not just 0. This should mark the
8452 shared libraries as not backward compatible. Of course, this should
8453 be changed again when we can guarantee backward binary compatibility.
8456 *) Fix typo in get_cert_by_subject() in by_dir.c
8457 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8459 *) Rework the system to generate shared libraries:
8461 - Make note of the expected extension for the shared libraries and
8462 if there is a need for symbolic links from for example libcrypto.so.0
8463 to libcrypto.so.0.9.7. There is extended info in Configure for
8466 - Make as few rebuilds of the shared libraries as possible.
8468 - Still avoid linking the OpenSSL programs with the shared libraries.
8470 - When installing, install the shared libraries separately from the
8474 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8476 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8477 and not in SSL_clear because the latter is also used by the
8478 accept/connect functions; previously, the settings made by
8479 SSL_set_read_ahead would be lost during the handshake.
8480 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8482 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8483 Previously, it would create entries for disabled algorithms no
8487 *) Added several new manual pages for SSL_* function.
8490 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8492 *) In ssl23_get_client_hello, generate an error message when faced
8493 with an initial SSL 3.0/TLS record that is too small to contain the
8494 first two bytes of the ClientHello message, i.e. client_version.
8495 (Note that this is a pathologic case that probably has never happened
8496 in real life.) The previous approach was to use the version number
8497 from the record header as a substitute; but our protocol choice
8498 should not depend on that one because it is not authenticated
8499 by the Finished messages.
8502 *) More robust randomness gathering functions for Windows.
8503 [Jeffrey Altman <jaltman@columbia.edu>]
8505 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8506 not set then we don't setup the error code for issuer check errors
8507 to avoid possibly overwriting other errors which the callback does
8508 handle. If an application does set the flag then we assume it knows
8509 what it is doing and can handle the new informational codes
8513 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8514 a general "ANY" type, as such it should be able to decode anything
8515 including tagged types. However it didn't check the class so it would
8516 wrongly interpret tagged types in the same way as their universal
8517 counterpart and unknown types were just rejected. Changed so that the
8518 tagged and unknown types are handled in the same way as a SEQUENCE:
8519 that is the encoding is stored intact. There is also a new type
8520 "V_ASN1_OTHER" which is used when the class is not universal, in this
8521 case we have no idea what the actual type is so we just lump them all
8525 *) On VMS, stdout may very well lead to a file that is written to
8526 in a record-oriented fashion. That means that every write() will
8527 write a separate record, which will be read separately by the
8528 programs trying to read from it. This can be very confusing.
8530 The solution is to put a BIO filter in the way that will buffer
8531 text until a linefeed is reached, and then write everything a
8532 line at a time, so every record written will be an actual line,
8533 not chunks of lines and not (usually doesn't happen, but I've
8534 seen it once) several lines in one record. BIO_f_linebuffer() is
8537 Currently, it's a VMS-only method, because that's where it has
8538 been tested well enough.
8541 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8542 it can return incorrect results.
8543 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8544 but it was in 0.9.6-beta[12].)
8547 *) Disable the check for content being present when verifying detached
8548 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8549 include zero length content when signing messages.
8552 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8553 BIO_ctrl (for BIO pairs).
8556 *) Add DSO method for VMS.
8559 *) Bug fix: Montgomery multiplication could produce results with the
8563 *) Add RPM specification openssl.spec and modify it to build three
8564 packages. The default package contains applications, application
8565 documentation and run-time libraries. The devel package contains
8566 include files, static libraries and function documentation. The
8567 doc package contains the contents of the doc directory. The original
8568 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8571 *) Add a large number of documentation files for many SSL routines.
8572 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8574 *) Add a configuration entry for Sony News 4.
8575 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8577 *) Don't set the two most significant bits to one when generating a
8578 random number < q in the DSA library.
8581 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8582 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8583 the underlying transport is blocking) if a handshake took place.
8584 (The default behaviour is needed by applications such as s_client
8585 and s_server that use select() to determine when to use SSL_read;
8586 but for applications that know in advance when to expect data, it
8587 just makes things more complicated.)
8590 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8594 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8595 work better on such systems.
8596 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8598 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8599 Update PKCS12_parse() so it copies the friendlyName and the
8600 keyid to the certificates aux info.
8603 *) Fix bug in PKCS7_verify() which caused an infinite loop
8604 if there was more than one signature.
8605 [Sven Uszpelkat <su@celocom.de>]
8607 *) Major change in util/mkdef.pl to include extra information
8608 about each symbol, as well as presenting variables as well
8609 as functions. This change means that there's n more need
8610 to rebuild the .num files when some algorithms are excluded.
8613 *) Allow the verify time to be set by an application,
8614 rather than always using the current time.
8617 *) Phase 2 verify code reorganisation. The certificate
8618 verify code now looks up an issuer certificate by a
8619 number of criteria: subject name, authority key id
8620 and key usage. It also verifies self signed certificates
8621 by the same criteria. The main comparison function is
8622 X509_check_issued() which performs these checks.
8624 Lot of changes were necessary in order to support this
8625 without completely rewriting the lookup code.
8627 Authority and subject key identifier are now cached.
8629 The LHASH 'certs' is X509_STORE has now been replaced
8630 by a STACK_OF(X509_OBJECT). This is mainly because an
8631 LHASH can't store or retrieve multiple objects with
8632 the same hash value.
8634 As a result various functions (which were all internal
8635 use only) have changed to handle the new X509_STORE
8636 structure. This will break anything that messed round
8637 with X509_STORE internally.
8639 The functions X509_STORE_add_cert() now checks for an
8640 exact match, rather than just subject name.
8642 The X509_STORE API doesn't directly support the retrieval
8643 of multiple certificates matching a given criteria, however
8644 this can be worked round by performing a lookup first
8645 (which will fill the cache with candidate certificates)
8646 and then examining the cache for matches. This is probably
8647 the best we can do without throwing out X509_LOOKUP
8648 entirely (maybe later...).
8650 The X509_VERIFY_CTX structure has been enhanced considerably.
8652 All certificate lookup operations now go via a get_issuer()
8653 callback. Although this currently uses an X509_STORE it
8654 can be replaced by custom lookups. This is a simple way
8655 to bypass the X509_STORE hackery necessary to make this
8656 work and makes it possible to use more efficient techniques
8657 in future. A very simple version which uses a simple
8658 STACK for its trusted certificate store is also provided
8659 using X509_STORE_CTX_trusted_stack().
8661 The verify_cb() and verify() callbacks now have equivalents
8662 in the X509_STORE_CTX structure.
8664 X509_STORE_CTX also has a 'flags' field which can be used
8665 to customise the verify behaviour.
8668 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8669 excludes S/MIME capabilities.
8672 *) When a certificate request is read in keep a copy of the
8673 original encoding of the signed data and use it when outputing
8674 again. Signatures then use the original encoding rather than
8675 a decoded, encoded version which may cause problems if the
8676 request is improperly encoded.
8679 *) For consistency with other BIO_puts implementations, call
8680 buffer_write(b, ...) directly in buffer_puts instead of calling
8683 In BIO_puts, increment b->num_write as in BIO_write.
8684 [Peter.Sylvester@EdelWeb.fr]
8686 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8687 BN_zero, we may not return a BIGNUM with an array consisting of
8691 *) Avoid calling abort() from within the library when problems are
8692 detected, except if preprocessor symbols have been defined
8693 (such as REF_CHECK, BN_DEBUG etc.).
8696 *) New openssl application 'rsautl'. This utility can be
8697 used for low level RSA operations. DER public key
8698 BIO/fp routines also added.
8701 *) New Configure entry and patches for compiling on QNX 4.
8702 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8704 *) A demo state-machine implementation was sponsored by
8705 Nuron (http://www.nuron.com/) and is now available in
8706 demos/state_machine.
8709 *) New options added to the 'dgst' utility for signature
8710 generation and verification.
8713 *) Unrecognized PKCS#7 content types are now handled via a
8714 catch all ASN1_TYPE structure. This allows unsupported
8715 types to be stored as a "blob" and an application can
8716 encode and decode it manually.
8719 *) Fix various signed/unsigned issues to make a_strex.c
8721 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8723 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8724 length if passed a buffer. ASN1_INTEGER_to_BN failed
8725 if passed a NULL BN and its argument was negative.
8726 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8728 *) Modification to PKCS#7 encoding routines to output definite
8729 length encoding. Since currently the whole structures are in
8730 memory there's not real point in using indefinite length
8731 constructed encoding. However if OpenSSL is compiled with
8732 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8735 *) Added BIO_vprintf() and BIO_vsnprintf().
8738 *) Added more prefixes to parse for in the the strings written
8739 through a logging bio, to cover all the levels that are available
8740 through syslog. The prefixes are now:
8742 PANIC, EMERG, EMR => LOG_EMERG
8743 ALERT, ALR => LOG_ALERT
8744 CRIT, CRI => LOG_CRIT
8745 ERROR, ERR => LOG_ERR
8746 WARNING, WARN, WAR => LOG_WARNING
8747 NOTICE, NOTE, NOT => LOG_NOTICE
8748 INFO, INF => LOG_INFO
8749 DEBUG, DBG => LOG_DEBUG
8751 and as before, if none of those prefixes are present at the
8752 beginning of the string, LOG_ERR is chosen.
8754 On Win32, the LOG_* levels are mapped according to this:
8756 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8757 LOG_WARNING => EVENTLOG_WARNING_TYPE
8758 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8762 *) Made it possible to reconfigure with just the configuration
8763 argument "reconf" or "reconfigure". The command line arguments
8764 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8765 and are retrieved from there when reconfiguring.
8769 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8771 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8774 *) The obj_dat.pl script was messing up the sorting of object
8775 names. The reason was that it compared the quoted version
8776 of strings as a result "OCSP" > "OCSP Signing" because
8777 " > SPACE. Changed script to store unquoted versions of
8778 names and add quotes on output. It was also omitting some
8779 names from the lookup table if they were given a default
8780 value (that is if SN is missing it is given the same
8781 value as LN and vice versa), these are now added on the
8782 grounds that if an object has a name we should be able to
8783 look it up. Finally added warning output when duplicate
8784 short or long names are found.
8787 *) Changes needed for Tandem NSK.
8788 [Scott Uroff <scott@xypro.com>]
8790 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8791 RSA_padding_check_SSLv23(), special padding was never detected
8792 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8793 version rollback attacks was not effective.
8795 In s23_clnt.c, don't use special rollback-attack detection padding
8796 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8797 client; similarly, in s23_srvr.c, don't do the rollback check if
8798 SSL 2.0 is the only protocol enabled in the server.
8801 *) Make it possible to get hexdumps of unprintable data with 'openssl
8802 asn1parse'. By implication, the functions ASN1_parse_dump() and
8803 BIO_dump_indent() are added.
8806 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8807 these print out strings and name structures based on various
8808 flags including RFC2253 support and proper handling of
8809 multibyte characters. Added options to the 'x509' utility
8810 to allow the various flags to be set.
8813 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8814 Also change the functions X509_cmp_current_time() and
8815 X509_gmtime_adj() work with an ASN1_TIME structure,
8816 this will enable certificates using GeneralizedTime in validity
8817 dates to be checked.
8820 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8821 negative public key encodings) on by default,
8822 NO_NEG_PUBKEY_BUG can be set to disable it.
8825 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8826 content octets. An i2c_ASN1_OBJECT is unnecessary because
8827 the encoding can be trivially obtained from the structure.
8830 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8831 not read locks (CRYPTO_r_[un]lock).
8834 *) A first attempt at creating official support for shared
8835 libraries through configuration. I've kept it so the
8836 default is static libraries only, and the OpenSSL programs
8837 are always statically linked for now, but there are
8838 preparations for dynamic linking in place.
8839 This has been tested on Linux and Tru64.
8842 *) Randomness polling function for Win9x, as described in:
8843 Peter Gutmann, Software Generation of Practically Strong
8847 *) Fix so PRNG is seeded in req if using an already existing
8851 *) New options to smime application. -inform and -outform
8852 allow alternative formats for the S/MIME message including
8853 PEM and DER. The -content option allows the content to be
8854 specified separately. This should allow things like Netscape
8855 form signing output easier to verify.
8858 *) Fix the ASN1 encoding of tags using the 'long form'.
8861 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8862 STRING types. These convert content octets to and from the
8863 underlying type. The actual tag and length octets are
8864 already assumed to have been read in and checked. These
8865 are needed because all other string types have virtually
8866 identical handling apart from the tag. By having versions
8867 of the ASN1 functions that just operate on content octets
8868 IMPLICIT tagging can be handled properly. It also allows
8869 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8870 and ASN1_INTEGER are identical apart from the tag.
8873 *) Change the handling of OID objects as follows:
8875 - New object identifiers are inserted in objects.txt, following
8876 the syntax given in objects.README.
8877 - objects.pl is used to process obj_mac.num and create a new
8879 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8882 This is currently kind of a hack, and the perl code in objects.pl
8883 isn't very elegant, but it works as I intended. The simplest way
8884 to check that it worked correctly is to look in obj_dat.h and
8885 check the array nid_objs and make sure the objects haven't moved
8886 around (this is important!). Additions are OK, as well as
8887 consistent name changes.
8890 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8893 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8894 The given file adds to whatever has already been seeded into the
8895 random pool through the RANDFILE configuration file option or
8896 environment variable, or the default random state file.
8899 *) mkstack.pl now sorts each macro group into lexical order.
8900 Previously the output order depended on the order the files
8901 appeared in the directory, resulting in needless rewriting
8905 *) Patches to make OpenSSL compile under Win32 again. Mostly
8906 work arounds for the VC++ problem that it treats func() as
8907 func(void). Also stripped out the parts of mkdef.pl that
8908 added extra typesafe functions: these no longer exist.
8911 *) Reorganisation of the stack code. The macros are now all
8912 collected in safestack.h . Each macro is defined in terms of
8913 a "stack macro" of the form SKM_<name>(type, a, b). The
8914 DEBUG_SAFESTACK is now handled in terms of function casts,
8915 this has the advantage of retaining type safety without the
8916 use of additional functions. If DEBUG_SAFESTACK is not defined
8917 then the non typesafe macros are used instead. Also modified the
8918 mkstack.pl script to handle the new form. Needs testing to see
8919 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8920 the default if no major problems. Similar behaviour for ASN1_SET_OF
8921 and PKCS12_STACK_OF.
8924 *) When some versions of IIS use the 'NET' form of private key the
8925 key derivation algorithm is different. Normally MD5(password) is
8926 used as a 128 bit RC4 key. In the modified case
8927 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8928 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8929 as the old Netscape_RSA functions except they have an additional
8930 'sgckey' parameter which uses the modified algorithm. Also added
8931 an -sgckey command line option to the rsa utility. Thanks to
8932 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8933 algorithm to openssl-dev.
8936 *) The evp_local.h macros were using 'c.##kname' which resulted in
8937 invalid expansion on some systems (SCO 5.0.5 for example).
8938 Corrected to 'c.kname'.
8939 [Phillip Porch <root@theporch.com>]
8941 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8942 a STACK of email addresses from a certificate or request, these look
8943 in the subject name and the subject alternative name extensions and
8944 omit any duplicate addresses.
8947 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8948 This makes DSA verification about 2 % faster.
8951 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8952 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8953 plus overhead for 1024 bit moduli).
8954 This makes exponentiations about 0.5 % faster for 1024 bit
8955 exponents (as measured by "openssl speed rsa2048").
8958 *) Rename memory handling macros to avoid conflicts with other
8960 Malloc => OPENSSL_malloc
8961 Malloc_locked => OPENSSL_malloc_locked
8962 Realloc => OPENSSL_realloc
8963 Free => OPENSSL_free
8966 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8967 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8970 *) CygWin32 support.
8971 [John Jarvie <jjarvie@newsguy.com>]
8973 *) The type-safe stack code has been rejigged. It is now only compiled
8974 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8975 by default all type-specific stack functions are "#define"d back to
8976 standard stack functions. This results in more streamlined output
8977 but retains the type-safety checking possibilities of the original
8981 *) The STACK code has been cleaned up, and certain type declarations
8982 that didn't make a lot of sense have been brought in line. This has
8983 also involved a cleanup of sorts in safestack.h to more correctly
8984 map type-safe stack functions onto their plain stack counterparts.
8985 This work has also resulted in a variety of "const"ifications of
8986 lots of the code, especially "_cmp" operations which should normally
8987 be prototyped with "const" parameters anyway.
8990 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8991 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8992 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8993 where all of 'md' is used each time the PRNG is used, but 'state'
8994 is used only indexed by a cyclic counter. As entropy may not be
8995 well distributed from the beginning, 'md' is important as a
8996 chaining variable. However, the output function chains only half
8997 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8998 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8999 in all of 'state' being rewritten, with the new values depending
9000 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9003 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9004 the handshake is continued after ssl_verify_cert_chain();
9005 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9006 can lead to 'unexplainable' connection aborts later.
9007 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9009 *) Major EVP API cipher revision.
9010 Add hooks for extra EVP features. This allows various cipher
9011 parameters to be set in the EVP interface. Support added for variable
9012 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9013 setting of RC2 and RC5 parameters.
9015 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9018 Remove lots of duplicated code from the EVP library. For example *every*
9019 cipher init() function handles the 'iv' in the same way according to the
9020 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9021 for CFB and OFB modes they zero ctx->num.
9023 New functionality allows removal of S/MIME code RC2 hack.
9025 Most of the routines have the same form and so can be declared in terms
9028 By shifting this to the top level EVP_CipherInit() it can be removed from
9029 all individual ciphers. If the cipher wants to handle IVs or keys
9030 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9033 Change lots of functions like EVP_EncryptUpdate() to now return a
9034 value: although software versions of the algorithms cannot fail
9035 any installed hardware versions can.
9038 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9039 this option is set, tolerate broken clients that send the negotiated
9040 protocol version number instead of the requested protocol version
9044 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9045 i.e. non-zero for export ciphersuites, zero otherwise.
9046 Previous versions had this flag inverted, inconsistent with
9047 rsa_tmp_cb (..._TMP_RSA_CB).
9048 [Bodo Moeller; problem reported by Amit Chopra]
9050 *) Add missing DSA library text string. Work around for some IIS
9051 key files with invalid SEQUENCE encoding.
9054 *) Add a document (doc/standards.txt) that list all kinds of standards
9055 and so on that are implemented in OpenSSL.
9058 *) Enhance c_rehash script. Old version would mishandle certificates
9059 with the same subject name hash and wouldn't handle CRLs at all.
9060 Added -fingerprint option to crl utility, to support new c_rehash
9064 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9067 *) Fix for SSL server purpose checking. Server checking was
9068 rejecting certificates which had extended key usage present
9069 but no ssl client purpose.
9070 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9072 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9073 is a little unclear about how a blank password is handled.
9074 Since the password in encoded as a BMPString with terminating
9075 double NULL a zero length password would end up as just the
9076 double NULL. However no password at all is different and is
9077 handled differently in the PKCS#12 key generation code. NS
9078 treats a blank password as zero length. MSIE treats it as no
9079 password on export: but it will try both on import. We now do
9080 the same: PKCS12_parse() tries zero length and no password if
9081 the password is set to "" or NULL (NULL is now a valid password:
9082 it wasn't before) as does the pkcs12 application.
9085 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9086 perror when PEM_read_bio_X509_REQ fails, the error message must
9087 be obtained from the error queue.
9090 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9091 it in ERR_remove_state if appropriate, and change ERR_get_state
9092 accordingly to avoid race conditions (this is necessary because
9093 thread_hash is no longer constant once set).
9096 *) Bugfix for linux-elf makefile.one.
9099 *) RSA_get_default_method() will now cause a default
9100 RSA_METHOD to be chosen if one doesn't exist already.
9101 Previously this was only set during a call to RSA_new()
9102 or RSA_new_method(NULL) meaning it was possible for
9103 RSA_get_default_method() to return NULL.
9106 *) Added native name translation to the existing DSO code
9107 that will convert (if the flag to do so is set) filenames
9108 that are sufficiently small and have no path information
9109 into a canonical native form. Eg. "blah" converted to
9110 "libblah.so" or "blah.dll" etc.
9113 *) New function ERR_error_string_n(e, buf, len) which is like
9114 ERR_error_string(e, buf), but writes at most 'len' bytes
9115 including the 0 terminator. For ERR_error_string_n, 'buf'
9117 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9119 *) CONF library reworked to become more general. A new CONF
9120 configuration file reader "class" is implemented as well as a
9121 new functions (NCONF_*, for "New CONF") to handle it. The now
9122 old CONF_* functions are still there, but are reimplemented to
9123 work in terms of the new functions. Also, a set of functions
9124 to handle the internal storage of the configuration data is
9125 provided to make it easier to write new configuration file
9126 reader "classes" (I can definitely see something reading a
9127 configuration file in XML format, for example), called _CONF_*,
9128 or "the configuration storage API"...
9130 The new configuration file reading functions are:
9132 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9133 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9135 NCONF_default, NCONF_WIN32
9137 NCONF_dump_fp, NCONF_dump_bio
9139 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9140 NCONF_new creates a new CONF object. This works in the same way
9141 as other interfaces in OpenSSL, like the BIO interface.
9142 NCONF_dump_* dump the internal storage of the configuration file,
9143 which is useful for debugging. All other functions take the same
9144 arguments as the old CONF_* functions wth the exception of the
9145 first that must be a `CONF *' instead of a `LHASH *'.
9147 To make it easer to use the new classes with the old CONF_* functions,
9148 the function CONF_set_default_method is provided.
9151 *) Add '-tls1' option to 'openssl ciphers', which was already
9152 mentioned in the documentation but had not been implemented.
9153 (This option is not yet really useful because even the additional
9154 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9157 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9158 OpenSSL-based applications) load shared libraries and bind to
9159 them in a portable way.
9160 [Geoff Thorpe, with contributions from Richard Levitte]
9162 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9164 *) Make sure _lrotl and _lrotr are only used with MSVC.
9166 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9167 (the default implementation of RAND_status).
9169 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9170 to '-clrext' (= clear extensions), as intended and documented.
9171 [Bodo Moeller; inconsistency pointed out by Michael Attili
9174 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9175 was larger than the MD block size.
9176 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9178 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9179 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9180 using the passed key: if the passed key was a private key the result
9181 of X509_print(), for example, would be to print out all the private key
9185 *) des_quad_cksum() byte order bug fix.
9186 [Ulf Möller, using the problem description in krb4-0.9.7, where
9187 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9189 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9191 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9193 *) For easily testing in shell scripts whether some command
9194 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9195 returns with exit code 0 iff no command of the given name is available.
9196 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9197 the output goes to stdout and nothing is printed to stderr.
9198 Additional arguments are always ignored.
9200 Since for each cipher there is a command of the same name,
9201 the 'no-cipher' compilation switches can be tested this way.
9203 ('openssl no-XXX' is not able to detect pseudo-commands such
9204 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9207 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9210 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9211 is set; it will be thrown away anyway because each handshake creates
9213 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9214 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9215 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9216 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9219 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9220 'Q' and 'R' lose their special meanings (quit/renegotiate).
9221 This is part of what -quiet does; unlike -quiet, -ign_eof
9222 does not suppress any output.
9225 *) Add compatibility options to the purpose and trust code. The
9226 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9227 accepts a certificate or CA, this was the previous behaviour,
9228 with all the associated security issues.
9230 X509_TRUST_COMPAT is the old trust behaviour: only and
9231 automatically trust self signed roots in certificate store. A
9232 new trust setting X509_TRUST_DEFAULT is used to specify that
9233 a purpose has no associated trust setting and it should instead
9234 use the value in the default purpose.
9237 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9238 and fix a memory leak.
9241 *) In util/mkerr.pl (which implements 'make errors'), preserve
9242 reason strings from the previous version of the .c file, as
9243 the default to have only downcase letters (and digits) in
9244 automatically generated reasons codes is not always appropriate.
9247 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9248 using strerror. Previously, ERR_reason_error_string() returned
9249 library names as reason strings for SYSerr; but SYSerr is a special
9250 case where small numbers are errno values, not library numbers.
9253 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9254 converts DSA parameters into DH parameters. (When creating parameters,
9255 DSA_generate_parameters is used.)
9258 *) Include 'length' (recommended exponent length) in C code generated
9259 by 'openssl dhparam -C'.
9262 *) The second argument to set_label in perlasm was already being used
9263 so couldn't be used as a "file scope" flag. Moved to third argument
9267 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9268 instead of RAND_bytes for encryption IVs and salts.
9271 *) Include RAND_status() into RAND_METHOD instead of implementing
9272 it only for md_rand.c Otherwise replacing the PRNG by calling
9273 RAND_set_rand_method would be impossible.
9276 *) Don't let DSA_generate_key() enter an infinite loop if the random
9277 number generation fails.
9280 *) New 'rand' application for creating pseudo-random output.
9283 *) Added configuration support for Linux/IA64
9284 [Rolf Haberrecker <rolf@suse.de>]
9286 *) Assembler module support for Mingw32.
9289 *) Shared library support for HPUX (in shlib/).
9290 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9292 *) Shared library support for Solaris gcc.
9293 [Lutz Behnke <behnke@trustcenter.de>]
9295 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9297 *) PKCS7_encrypt() was adding text MIME headers twice because they
9298 were added manually and by SMIME_crlf_copy().
9301 *) In bntest.c don't call BN_rand with zero bits argument.
9302 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9304 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9305 case was implemented. This caused BN_div_recp() to fail occasionally.
9308 *) Add an optional second argument to the set_label() in the perl
9309 assembly language builder. If this argument exists and is set
9310 to 1 it signals that the assembler should use a symbol whose
9311 scope is the entire file, not just the current function. This
9312 is needed with MASM which uses the format label:: for this scope.
9313 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9315 *) Change the ASN1 types so they are typedefs by default. Before
9316 almost all types were #define'd to ASN1_STRING which was causing
9317 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9321 *) Change names of new functions to the new get1/get0 naming
9322 convention: After 'get1', the caller owns a reference count
9323 and has to call ..._free; 'get0' returns a pointer to some
9324 data structure without incrementing reference counters.
9325 (Some of the existing 'get' functions increment a reference
9326 counter, some don't.)
9327 Similarly, 'set1' and 'add1' functions increase reference
9328 counters or duplicate objects.
9331 *) Allow for the possibility of temp RSA key generation failure:
9332 the code used to assume it always worked and crashed on failure.
9335 *) Fix potential buffer overrun problem in BIO_printf().
9336 [Ulf Möller, using public domain code by Patrick Powell; problem
9337 pointed out by David Sacerdote <das33@cornell.edu>]
9339 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9340 RAND_egd() and RAND_status(). In the command line application,
9341 the EGD socket can be specified like a seed file using RANDFILE
9345 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9346 Some CAs (e.g. Verisign) distribute certificates in this form.
9349 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9350 list to exclude them. This means that no special compilation option
9351 is needed to use anonymous DH: it just needs to be included in the
9355 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9356 EVP_MD_type. The old functionality is available in a new macro called
9357 EVP_MD_md(). Change code that uses it and update docs.
9360 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9361 where the 'void *' argument is replaced by a function pointer argument.
9362 Previously 'void *' was abused to point to functions, which works on
9363 many platforms, but is not correct. As these functions are usually
9364 called by macros defined in OpenSSL header files, most source code
9365 should work without changes.
9368 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9369 sections with information on -D... compiler switches used for
9370 compiling the library so that applications can see them. To enable
9371 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9372 must be defined. E.g.,
9373 #define OPENSSL_ALGORITHM_DEFINES
9374 #include <openssl/opensslconf.h>
9375 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9376 [Richard Levitte, Ulf and Bodo Möller]
9378 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9382 *) Change the 'other' type in certificate aux info to a STACK_OF
9383 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9384 the required ASN1 format: arbitrary types determined by an OID.
9387 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9388 argument to 'req'. This is not because the function is newer or
9389 better than others it just uses the work 'NEW' in the certificate
9390 request header lines. Some software needs this.
9393 *) Reorganise password command line arguments: now passwords can be
9394 obtained from various sources. Delete the PEM_cb function and make
9395 it the default behaviour: i.e. if the callback is NULL and the
9396 usrdata argument is not NULL interpret it as a null terminated pass
9397 phrase. If usrdata and the callback are NULL then the pass phrase
9398 is prompted for as usual.
9401 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9402 the support is automatically enabled. The resulting binaries will
9403 autodetect the card and use it if present.
9404 [Ben Laurie and Compaq Inc.]
9406 *) Work around for Netscape hang bug. This sends certificate request
9407 and server done in one record. Since this is perfectly legal in the
9408 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9409 the bugs/SSLv3 entry for more info.
9412 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9415 *) Add -rand argument to smime and pkcs12 applications and read/write
9419 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9422 *) Add command line password options to the remaining applications.
9425 *) Bug fix for BN_div_recp() for numerators with an even number of
9429 *) More tests in bntest.c, and changed test_bn output.
9432 *) ./config recognizes MacOS X now.
9435 *) Bug fix for BN_div() when the first words of num and divsor are
9436 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9439 *) Add support for various broken PKCS#8 formats, and command line
9440 options to produce them.
9443 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9444 get temporary BIGNUMs from a BN_CTX.
9447 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9451 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9452 include a #define from the old name to the new. The original intent
9453 was that statically linked binaries could for example just call
9454 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9455 link with digests. This never worked because SSLeay_add_all_digests()
9456 and SSLeay_add_all_ciphers() were in the same source file so calling
9457 one would link with the other. They are now in separate source files.
9460 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9463 *) Use a less unusual form of the Miller-Rabin primality test (it used
9464 a binary algorithm for exponentiation integrated into the Miller-Rabin
9465 loop, our standard modexp algorithms are faster).
9468 *) Support for the EBCDIC character set completed.
9469 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9471 *) Source code cleanups: use const where appropriate, eliminate casts,
9472 use void * instead of char * in lhash.
9475 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9476 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9477 this the server could overwrite ephemeral keys that the client
9481 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9482 using 50 iterations of the Rabin-Miller test.
9484 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9485 iterations of the Rabin-Miller test as required by the appendix
9486 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9487 As BN_is_prime_fasttest includes trial division, DSA parameter
9488 generation becomes much faster.
9490 This implies a change for the callback functions in DSA_is_prime
9491 and DSA_generate_parameters: The callback function is called once
9492 for each positive witness in the Rabin-Miller test, not just
9493 occasionally in the inner loop; and the parameters to the
9494 callback function now provide an iteration count for the outer
9495 loop rather than for the current invocation of the inner loop.
9496 DSA_generate_parameters additionally can call the callback
9497 function with an 'iteration count' of -1, meaning that a
9498 candidate has passed the trial division test (when q is generated
9499 from an application-provided seed, trial division is skipped).
9502 *) New function BN_is_prime_fasttest that optionally does trial
9503 division before starting the Rabin-Miller test and has
9504 an additional BN_CTX * argument (whereas BN_is_prime always
9505 has to allocate at least one BN_CTX).
9506 'callback(1, -1, cb_arg)' is called when a number has passed the
9507 trial division stage.
9510 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9514 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9517 *) New function BN_pseudo_rand().
9520 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9521 bignum version of BN_from_montgomery() with the working code from
9522 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9526 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9527 made it impossible to use the same SSL_SESSION data structure in
9528 SSL2 clients in multiple threads.
9531 *) The return value of RAND_load_file() no longer counts bytes obtained
9532 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9533 to seed the PRNG (previously an explicit byte count was required).
9534 [Ulf Möller, Bodo Möller]
9536 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9537 used (char *) instead of (void *) and had casts all over the place.
9540 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9543 *) Retain source code compatibility for BN_prime_checks macro:
9544 BN_is_prime(..., BN_prime_checks, ...) now uses
9545 BN_prime_checks_for_size to determine the appropriate number of
9546 Rabin-Miller iterations.
9549 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9550 DH_CHECK_P_NOT_SAFE_PRIME.
9551 (Check if this is true? OpenPGP calls them "strong".)
9554 *) Merge the functionality of "dh" and "gendh" programs into a new program
9555 "dhparam". The old programs are retained for now but will handle DH keys
9556 (instead of parameters) in future.
9559 *) Make the ciphers, s_server and s_client programs check the return values
9560 when a new cipher list is set.
9563 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9564 ciphers. Before when the 56bit ciphers were enabled the sorting was
9567 The syntax for the cipher sorting has been extended to support sorting by
9568 cipher-strength (using the strength_bits hard coded in the tables).
9569 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9571 Fix a bug in the cipher-command parser: when supplying a cipher command
9572 string with an "undefined" symbol (neither command nor alphanumeric
9573 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9574 an error is flagged.
9576 Due to the strength-sorting extension, the code of the
9577 ssl_create_cipher_list() function was completely rearranged. I hope that
9578 the readability was also increased :-)
9579 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9581 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9582 for the first serial number and places 2 in the serial number file. This
9583 avoids problems when the root CA is created with serial number zero and
9584 the first user certificate has the same issuer name and serial number
9588 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9589 the new code. Add documentation for this stuff.
9592 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9593 X509_*() to X509at_*() on the grounds that they don't handle X509
9594 structures and behave in an analogous way to the X509v3 functions:
9595 they shouldn't be called directly but wrapper functions should be used
9598 So we also now have some wrapper functions that call the X509at functions
9599 when passed certificate requests. (TO DO: similar things can be done with
9600 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9601 things. Some of these need some d2i or i2d and print functionality
9602 because they handle more complex structures.)
9605 *) Add missing #ifndefs that caused missing symbols when building libssl
9606 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9607 NO_RSA in ssl/s2*.c.
9608 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9610 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9611 has a return value which indicates the quality of the random data
9612 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9613 error queue. New function RAND_pseudo_bytes() generates output that is
9614 guaranteed to be unique but not unpredictable. RAND_add is like
9615 RAND_seed, but takes an extra argument for an entropy estimate
9616 (RAND_seed always assumes full entropy).
9619 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9620 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9621 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9622 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9623 false-positive rate of at most 2^-80 for random input.
9626 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9629 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9630 in the 0.9.5 release), this returns the chain
9631 from an X509_CTX structure with a dup of the stack and all
9632 the X509 reference counts upped: so the stack will exist
9633 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9636 Also make SSL_SESSION_print() print out the verify return
9640 *) Add manpage for the pkcs12 command. Also change the default
9641 behaviour so MAC iteration counts are used unless the new
9642 -nomaciter option is used. This improves file security and
9643 only older versions of MSIE (4.0 for example) need it.
9646 *) Honor the no-xxx Configure options when creating .DEF files.
9649 *) Add PKCS#10 attributes to field table: challengePassword,
9650 unstructuredName and unstructuredAddress. These are taken from
9651 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9652 international characters are used.
9654 More changes to X509_ATTRIBUTE code: allow the setting of types
9655 based on strings. Remove the 'loc' parameter when adding
9656 attributes because these will be a SET OF encoding which is sorted
9660 *) Initial changes to the 'req' utility to allow request generation
9661 automation. This will allow an application to just generate a template
9662 file containing all the field values and have req construct the
9665 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9666 used all over the place including certificate requests and PKCS#7
9667 structures. They are currently handled manually where necessary with
9668 some primitive wrappers for PKCS#7. The new functions behave in a
9669 manner analogous to the X509 extension functions: they allow
9670 attributes to be looked up by NID and added.
9672 Later something similar to the X509V3 code would be desirable to
9673 automatically handle the encoding, decoding and printing of the
9674 more complex types. The string types like challengePassword can
9675 be handled by the string table functions.
9677 Also modified the multi byte string table handling. Now there is
9678 a 'global mask' which masks out certain types. The table itself
9679 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9680 is useful when for example there is only one permissible type
9681 (as in countryName) and using the mask might result in no valid
9685 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9686 SSL_get_peer_finished to allow applications to obtain the latest
9687 Finished messages sent to the peer or expected from the peer,
9688 respectively. (SSL_get_peer_finished is usually the Finished message
9689 actually received from the peer, otherwise the protocol will be aborted.)
9691 As the Finished message are message digests of the complete handshake
9692 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9693 be used for external authentication procedures when the authentication
9694 provided by SSL/TLS is not desired or is not enough.
9697 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9698 the host supports BWX extension and if Compaq C is present on the
9699 $PATH. Just exploiting of the BWX extension results in 20-30%
9700 performance kick for some algorithms, e.g. DES and RC4 to mention
9701 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9705 *) Add support for MS "fast SGC". This is arguably a violation of the
9706 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9707 weak crypto and after checking the certificate is SGC a second one
9708 with strong crypto. MS SGC stops the first handshake after receiving
9709 the server certificate message and sends a second client hello. Since
9710 a server will typically do all the time consuming operations before
9711 expecting any further messages from the client (server key exchange
9712 is the most expensive) there is little difference between the two.
9714 To get OpenSSL to support MS SGC we have to permit a second client
9715 hello message after we have sent server done. In addition we have to
9716 reset the MAC if we do get this second client hello.
9719 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9720 if a DER encoded private key is RSA or DSA traditional format. Changed
9721 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9722 format DER encoded private key. Newer code should use PKCS#8 format which
9723 has the key type encoded in the ASN1 structure. Added DER private key
9724 support to pkcs8 application.
9727 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9728 ciphersuites has been selected (as required by the SSL 3/TLS 1
9729 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9730 is set, we interpret this as a request to violate the specification
9731 (the worst that can happen is a handshake failure, and 'correct'
9732 behaviour would result in a handshake failure anyway).
9735 *) In SSL_CTX_add_session, take into account that there might be multiple
9736 SSL_SESSION structures with the same session ID (e.g. when two threads
9737 concurrently obtain them from an external cache).
9738 The internal cache can handle only one SSL_SESSION with a given ID,
9739 so if there's a conflict, we now throw out the old one to achieve
9743 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9744 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9745 some routines that use cipher OIDs: some ciphers do not have OIDs
9746 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9750 *) Simplify the trust setting structure and code. Now we just have
9751 two sequences of OIDs for trusted and rejected settings. These will
9752 typically have values the same as the extended key usage extension
9753 and any application specific purposes.
9755 The trust checking code now has a default behaviour: it will just
9756 check for an object with the same NID as the passed id. Functions can
9757 be provided to override either the default behaviour or the behaviour
9758 for a given id. SSL client, server and email already have functions
9759 in place for compatibility: they check the NID and also return "trusted"
9760 if the certificate is self signed.
9763 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9764 traditional format into an EVP_PKEY structure.
9767 *) Add a password callback function PEM_cb() which either prompts for
9768 a password if usr_data is NULL or otherwise assumes it is a null
9769 terminated password. Allow passwords to be passed on command line
9770 environment or config files in a few more utilities.
9773 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9774 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9775 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9776 Update documentation.
9779 *) Support for ASN1 "NULL" type. This could be handled before by using
9780 ASN1_TYPE but there wasn't any function that would try to read a NULL
9781 and produce an error if it couldn't. For compatibility we also have
9782 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9783 don't allocate anything because they don't need to.
9786 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9788 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9790 *) Rebuild of the memory allocation routines used by OpenSSL code and
9791 possibly others as well. The purpose is to make an interface that
9792 provide hooks so anyone can build a separate set of allocation and
9793 deallocation routines to be used by OpenSSL, for example memory
9794 pool implementations, or something else, which was previously hard
9795 since Malloc(), Realloc() and Free() were defined as macros having
9796 the values malloc, realloc and free, respectively (except for Win32
9797 compilations). The same is provided for memory debugging code.
9798 OpenSSL already comes with functionality to find memory leaks, but
9799 this gives people a chance to debug other memory problems.
9801 With these changes, a new set of functions and macros have appeared:
9803 CRYPTO_set_mem_debug_functions() [F]
9804 CRYPTO_get_mem_debug_functions() [F]
9805 CRYPTO_dbg_set_options() [F]
9806 CRYPTO_dbg_get_options() [F]
9807 CRYPTO_malloc_debug_init() [M]
9809 The memory debug functions are NULL by default, unless the library
9810 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9811 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9812 gives the standard debugging functions that come with OpenSSL) or
9813 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9814 provided by the library user) must be used. When the standard
9815 debugging functions are used, CRYPTO_dbg_set_options can be used to
9816 request additional information:
9817 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9818 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9820 Also, things like CRYPTO_set_mem_functions will always give the
9821 expected result (the new set of functions is used for allocation
9822 and deallocation) at all times, regardless of platform and compiler
9825 To finish it up, some functions that were never use in any other
9826 way than through macros have a new API and new semantic:
9829 CRYPTO_dbg_realloc()
9832 All macros of value have retained their old syntax.
9833 [Richard Levitte and Bodo Moeller]
9835 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9836 ordering of SMIMECapabilities wasn't in "strength order" and there
9837 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9841 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9842 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9843 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9845 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9846 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9847 functionality to handle multipart/signed properly) and a utility
9848 called 'smime' to call all this stuff. This is based on code I
9849 originally wrote for Celo who have kindly allowed it to be
9850 included in OpenSSL.
9853 *) Add variants des_set_key_checked and des_set_key_unchecked of
9854 des_set_key (aka des_key_sched). Global variable des_check_key
9855 decides which of these is called by des_set_key; this way
9856 des_check_key behaves as it always did, but applications and
9857 the library itself, which was buggy for des_check_key == 1,
9858 have a cleaner way to pick the version they need.
9861 *) New function PKCS12_newpass() which changes the password of a
9865 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9866 dynamic mix. In both cases the ids can be used as an index into the
9867 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9868 functions so they accept a list of the field values and the
9869 application doesn't need to directly manipulate the X509_TRUST
9873 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9877 *) Modify the way the V3 extension code looks up extensions. This now
9878 works in a similar way to the object code: we have some "standard"
9879 extensions in a static table which is searched with OBJ_bsearch()
9880 and the application can add dynamic ones if needed. The file
9881 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9882 updated whenever a new extension is added to the core code and kept
9883 in ext_nid order. There is a simple program 'tabtest.c' which checks
9884 this. New extensions are not added too often so this file can readily
9885 be maintained manually.
9887 There are two big advantages in doing things this way. The extensions
9888 can be looked up immediately and no longer need to be "added" using
9889 X509V3_add_standard_extensions(): this function now does nothing.
9890 [Side note: I get *lots* of email saying the extension code doesn't
9891 work because people forget to call this function]
9892 Also no dynamic allocation is done unless new extensions are added:
9893 so if we don't add custom extensions there is no need to call
9894 X509V3_EXT_cleanup().
9897 *) Modify enc utility's salting as follows: make salting the default. Add a
9898 magic header, so unsalted files fail gracefully instead of just decrypting
9899 to garbage. This is because not salting is a big security hole, so people
9900 should be discouraged from doing it.
9903 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9904 digest to be passed on the command line but it only used this
9905 parameter when signing a certificate. Modified so all relevant
9906 operations are affected by the digest parameter including the
9907 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9908 DSA key was used because it didn't fix the digest.
9911 *) Initial certificate chain verify code. Currently tests the untrusted
9912 certificates for consistency with the verify purpose (which is set
9913 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9915 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9916 this is because it will reject chains with invalid extensions whereas
9917 every previous version of OpenSSL and SSLeay made no checks at all.
9919 Trust code: checks the root CA for the relevant trust settings. Trust
9920 settings have an initial value consistent with the verify purpose: e.g.
9921 if the verify purpose is for SSL client use it expects the CA to be
9922 trusted for SSL client use. However the default value can be changed to
9923 permit custom trust settings: one example of this would be to only trust
9924 certificates from a specific "secure" set of CAs.
9926 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9927 which should be used for version portability: especially since the
9928 verify structure is likely to change more often now.
9930 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9931 to set them. If not set then assume SSL clients will verify SSL servers
9934 Two new options to the verify program: -untrusted allows a set of
9935 untrusted certificates to be passed in and -purpose which sets the
9936 intended purpose of the certificate. If a purpose is set then the
9937 new chain verify code is used to check extension consistency.
9940 *) Support for the authority information access extension.
9943 *) Modify RSA and DSA PEM read routines to transparently handle
9944 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9945 public keys in a format compatible with certificate
9946 SubjectPublicKeyInfo structures. Unfortunately there were already
9947 functions called *_PublicKey_* which used various odd formats so
9948 these are retained for compatibility: however the DSA variants were
9949 never in a public release so they have been deleted. Changed dsa/rsa
9950 utilities to handle the new format: note no releases ever handled public
9951 keys so we should be OK.
9953 The primary motivation for this change is to avoid the same fiasco
9954 that dogs private keys: there are several incompatible private key
9955 formats some of which are standard and some OpenSSL specific and
9956 require various evil hacks to allow partial transparent handling and
9957 even then it doesn't work with DER formats. Given the option anything
9958 other than PKCS#8 should be dumped: but the other formats have to
9959 stay in the name of compatibility.
9961 With public keys and the benefit of hindsight one standard format
9962 is used which works with EVP_PKEY, RSA or DSA structures: though
9963 it clearly returns an error if you try to read the wrong kind of key.
9965 Added a -pubkey option to the 'x509' utility to output the public key.
9966 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9967 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9968 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9969 that do the same as the EVP_PKEY_assign_*() except they up the
9970 reference count of the added key (they don't "swallow" the
9974 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9975 CRLs would fail if the file contained no certificates or no CRLs:
9976 added a new function to read in both types and return the number
9977 read: this means that if none are read it will be an error. The
9978 DER versions of the certificate and CRL reader would always fail
9979 because it isn't possible to mix certificates and CRLs in DER format
9980 without choking one or the other routine. Changed this to just read
9981 a certificate: this is the best we can do. Also modified the code
9982 in apps/verify.c to take notice of return codes: it was previously
9983 attempting to read in certificates from NULL pointers and ignoring
9984 any errors: this is one reason why the cert and CRL reader seemed
9985 to work. It doesn't check return codes from the default certificate
9986 routines: these may well fail if the certificates aren't installed.
9989 *) Code to support otherName option in GeneralName.
9992 *) First update to verify code. Change the verify utility
9993 so it warns if it is passed a self signed certificate:
9994 for consistency with the normal behaviour. X509_verify
9995 has been modified to it will now verify a self signed
9996 certificate if *exactly* the same certificate appears
9997 in the store: it was previously impossible to trust a
9998 single self signed certificate. This means that:
9999 openssl verify ss.pem
10000 now gives a warning about a self signed certificate but
10001 openssl verify -CAfile ss.pem ss.pem
10005 *) For servers, store verify_result in SSL_SESSION data structure
10006 (and add it to external session representation).
10007 This is needed when client certificate verifications fails,
10008 but an application-provided verification callback (set by
10009 SSL_CTX_set_cert_verify_callback) allows accepting the session
10010 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10011 but returns 1): When the session is reused, we have to set
10012 ssl->verify_result to the appropriate error code to avoid
10014 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10016 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10017 case in PKCS7_dataInit() where the signed PKCS7 structure
10018 didn't contain any existing data because it was being created.
10019 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10021 *) Add a salt to the key derivation routines in enc.c. This
10022 forms the first 8 bytes of the encrypted file. Also add a
10023 -S option to allow a salt to be input on the command line.
10026 *) New function X509_cmp(). Oddly enough there wasn't a function
10027 to compare two certificates. We do this by working out the SHA1
10028 hash and comparing that. X509_cmp() will be needed by the trust
10032 *) SSL_get1_session() is like SSL_get_session(), but increments
10033 the reference count in the SSL_SESSION returned.
10034 [Geoff Thorpe <geoff@eu.c2.net>]
10036 *) Fix for 'req': it was adding a null to request attributes.
10037 Also change the X509_LOOKUP and X509_INFO code to handle
10038 certificate auxiliary information.
10041 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10045 *) Add the possibility to add extra information to the memory leak
10046 detecting output, to form tracebacks, showing from where each
10047 allocation was originated: CRYPTO_push_info("constant string") adds
10048 the string plus current file name and line number to a per-thread
10049 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10050 is like calling CYRPTO_pop_info() until the stack is empty.
10051 Also updated memory leak detection code to be multi-thread-safe.
10054 *) Add options -text and -noout to pkcs7 utility and delete the
10055 encryption options which never did anything. Update docs.
10058 *) Add options to some of the utilities to allow the pass phrase
10059 to be included on either the command line (not recommended on
10060 OSes like Unix) or read from the environment. Update the
10061 manpages and fix a few bugs.
10064 *) Add a few manpages for some of the openssl commands.
10067 *) Fix the -revoke option in ca. It was freeing up memory twice,
10068 leaking and not finding already revoked certificates.
10071 *) Extensive changes to support certificate auxiliary information.
10072 This involves the use of X509_CERT_AUX structure and X509_AUX
10073 functions. An X509_AUX function such as PEM_read_X509_AUX()
10074 can still read in a certificate file in the usual way but it
10075 will also read in any additional "auxiliary information". By
10076 doing things this way a fair degree of compatibility can be
10077 retained: existing certificates can have this information added
10078 using the new 'x509' options.
10080 Current auxiliary information includes an "alias" and some trust
10081 settings. The trust settings will ultimately be used in enhanced
10082 certificate chain verification routines: currently a certificate
10083 can only be trusted if it is self signed and then it is trusted
10087 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10088 The problem was that one of the replacement routines had not been working
10089 since SSLeay releases. For now the offending routine has been replaced
10090 with non-optimised assembler. Even so, this now gives around 95%
10091 performance improvement for 1024 bit RSA signs.
10094 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10095 handling. Most clients have the effective key size in bits equal to
10096 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10097 A few however don't do this and instead use the size of the decrypted key
10098 to determine the RC2 key length and the AlgorithmIdentifier to determine
10099 the effective key length. In this case the effective key length can still
10100 be 40 bits but the key length can be 168 bits for example. This is fixed
10101 by manually forcing an RC2 key into the EVP_PKEY structure because the
10102 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10103 the key length and effective key length are equal.
10106 *) Add a bunch of functions that should simplify the creation of
10107 X509_NAME structures. Now you should be able to do:
10108 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10109 and have it automatically work out the correct field type and fill in
10110 the structures. The more adventurous can try:
10111 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10112 and it will (hopefully) work out the correct multibyte encoding.
10115 *) Change the 'req' utility to use the new field handling and multibyte
10116 copy routines. Before the DN field creation was handled in an ad hoc
10117 way in req, ca, and x509 which was rather broken and didn't support
10118 BMPStrings or UTF8Strings. Since some software doesn't implement
10119 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10120 using the dirstring_type option. See the new comment in the default
10121 openssl.cnf for more info.
10124 *) Make crypto/rand/md_rand.c more robust:
10125 - Assure unique random numbers after fork().
10126 - Make sure that concurrent threads access the global counter and
10127 md serializably so that we never lose entropy in them
10128 or use exactly the same state in multiple threads.
10129 Access to the large state is not always serializable because
10130 the additional locking could be a performance killer, and
10131 md should be large enough anyway.
10134 *) New file apps/app_rand.c with commonly needed functionality
10135 for handling the random seed file.
10137 Use the random seed file in some applications that previously did not:
10139 dsaparam -genkey (which also ignored its '-rand' option),
10142 x509 (when signing).
10143 Except on systems with /dev/urandom, it is crucial to have a random
10144 seed file at least for key creation, DSA signing, and for DH exchanges;
10145 for RSA signatures we could do without one.
10147 gendh and gendsa (unlike genrsa) used to read only the first byte
10148 of each file listed in the '-rand' option. The function as previously
10149 found in genrsa is now in app_rand.c and is used by all programs
10150 that support '-rand'.
10153 *) In RAND_write_file, use mode 0600 for creating files;
10154 don't just chmod when it may be too late.
10157 *) Report an error from X509_STORE_load_locations
10158 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10161 *) New function ASN1_mbstring_copy() this copies a string in either
10162 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10163 into an ASN1_STRING type. A mask of permissible types is passed
10164 and it chooses the "minimal" type to use or an error if not type
10168 *) Add function equivalents to the various macros in asn1.h. The old
10169 macros are retained with an M_ prefix. Code inside the library can
10170 use the M_ macros. External code (including the openssl utility)
10171 should *NOT* in order to be "shared library friendly".
10174 *) Add various functions that can check a certificate's extensions
10175 to see if it usable for various purposes such as SSL client,
10176 server or S/MIME and CAs of these types. This is currently
10177 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10178 verification. Also added a -purpose flag to x509 utility to
10179 print out all the purposes.
10182 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10186 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10187 for, obtain and decode and extension and obtain its critical flag.
10188 This allows all the necessary extension code to be handled in a
10189 single function call.
10192 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10193 platforms. See crypto/rc4/rc4_enc.c for further details.
10196 *) New -noout option to asn1parse. This causes no output to be produced
10197 its main use is when combined with -strparse and -out to extract data
10198 from a file (which may not be in ASN.1 format).
10201 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10202 when producing the local key id.
10203 [Richard Levitte <levitte@stacken.kth.se>]
10205 *) New option -dhparam in s_server. This allows a DH parameter file to be
10206 stated explicitly. If it is not stated then it tries the first server
10207 certificate file. The previous behaviour hard coded the filename
10211 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10212 a public key to be input or output. For example:
10213 openssl rsa -in key.pem -pubout -out pubkey.pem
10214 Also added necessary DSA public key functions to handle this.
10217 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10218 in the message. This was handled by allowing
10219 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10220 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10222 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10223 to the end of the strings whereas this didn't. This would cause problems
10224 if strings read with d2i_ASN1_bytes() were later modified.
10225 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10227 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10228 data and it contains EOF it will end up returning an error. This is
10229 caused by input 46 bytes long. The cause is due to the way base64
10230 BIOs find the start of base64 encoded data. They do this by trying a
10231 trial decode on each line until they find one that works. When they
10232 do a flag is set and it starts again knowing it can pass all the
10233 data directly through the decoder. Unfortunately it doesn't reset
10234 the context it uses. This means that if EOF is reached an attempt
10235 is made to pass two EOFs through the context and this causes the
10236 resulting error. This can also cause other problems as well. As is
10237 usual with these problems it takes *ages* to find and the fix is
10238 trivial: move one line.
10239 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10241 *) Ugly workaround to get s_client and s_server working under Windows. The
10242 old code wouldn't work because it needed to select() on sockets and the
10243 tty (for keypresses and to see if data could be written). Win32 only
10244 supports select() on sockets so we select() with a 1s timeout on the
10245 sockets and then see if any characters are waiting to be read, if none
10246 are present then we retry, we also assume we can always write data to
10247 the tty. This isn't nice because the code then blocks until we've
10248 received a complete line of data and it is effectively polling the
10249 keyboard at 1s intervals: however it's quite a bit better than not
10250 working at all :-) A dedicated Windows application might handle this
10251 with an event loop for example.
10254 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10255 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10256 will be called when RSA_sign() and RSA_verify() are used. This is useful
10257 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10258 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10259 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10260 This necessitated the support of an extra signature type NID_md5_sha1
10261 for SSL signatures and modifications to the SSL library to use it instead
10262 of calling RSA_public_decrypt() and RSA_private_encrypt().
10265 *) Add new -verify -CAfile and -CApath options to the crl program, these
10266 will lookup a CRL issuers certificate and verify the signature in a
10267 similar way to the verify program. Tidy up the crl program so it
10268 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10269 less strict. It will now permit CRL extensions even if it is not
10270 a V2 CRL: this will allow it to tolerate some broken CRLs.
10273 *) Initialize all non-automatic variables each time one of the openssl
10274 sub-programs is started (this is necessary as they may be started
10275 multiple times from the "OpenSSL>" prompt).
10276 [Lennart Bang, Bodo Moeller]
10278 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10279 removing all other RSA functionality (this is what NO_RSA does). This
10280 is so (for example) those in the US can disable those operations covered
10281 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10285 *) Non-copying interface to BIO pairs.
10286 (still largely untested)
10289 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10290 ASCII string. This was handled independently in various places before.
10293 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10294 UTF8 strings a character at a time.
10297 *) Use client_version from client hello to select the protocol
10298 (s23_srvr.c) and for RSA client key exchange verification
10299 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10302 *) Add various utility functions to handle SPKACs, these were previously
10303 handled by poking round in the structure internals. Added new function
10304 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10305 print, verify and generate SPKACs. Based on an original idea from
10306 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10309 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10312 *) Allow the config file extension section to be overwritten on the
10313 command line. Based on an original idea from Massimiliano Pala
10314 <madwolf@comune.modena.it>. The new option is called -extensions
10315 and can be applied to ca, req and x509. Also -reqexts to override
10316 the request extensions in req and -crlexts to override the crl extensions
10320 *) Add new feature to the SPKAC handling in ca. Now you can include
10321 the same field multiple times by preceding it by "XXXX." for example:
10324 this is the same syntax as used in the req config file.
10327 *) Allow certificate extensions to be added to certificate requests. These
10328 are specified in a 'req_extensions' option of the req section of the
10329 config file. They can be printed out with the -text option to req but
10330 are otherwise ignored at present.
10333 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10334 data read consists of only the final block it would not decrypted because
10335 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10336 A misplaced 'break' also meant the decrypted final block might not be
10337 copied until the next read.
10340 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10341 a few extra parameters to the DH structure: these will be useful if
10342 for example we want the value of 'q' or implement X9.42 DH.
10345 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10346 provides hooks that allow the default DSA functions or functions on a
10347 "per key" basis to be replaced. This allows hardware acceleration and
10348 hardware key storage to be handled without major modification to the
10349 library. Also added low level modexp hooks and CRYPTO_EX structure and
10350 associated functions.
10353 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10354 as "read only": it can't be written to and the buffer it points to will
10355 not be freed. Reading from a read only BIO is much more efficient than
10356 a normal memory BIO. This was added because there are several times when
10357 an area of memory needs to be read from a BIO. The previous method was
10358 to create a memory BIO and write the data to it, this results in two
10359 copies of the data and an O(n^2) reading algorithm. There is a new
10360 function BIO_new_mem_buf() which creates a read only memory BIO from
10361 an area of memory. Also modified the PKCS#7 routines to use read only
10365 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10366 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10367 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10368 but a retry condition occurred while trying to read the rest.
10371 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10372 NID_pkcs7_encrypted by default: this was wrong since this should almost
10373 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10374 the encrypted data type: this is a more sensible place to put it and it
10375 allows the PKCS#12 code to be tidied up that duplicated this
10379 *) Changed obj_dat.pl script so it takes its input and output files on
10380 the command line. This should avoid shell escape redirection problems
10384 *) Initial support for certificate extension requests, these are included
10385 in things like Xenroll certificate requests. Included functions to allow
10386 extensions to be obtained and added.
10389 *) -crlf option to s_client and s_server for sending newlines as
10390 CRLF (as required by many protocols).
10393 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10395 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10396 [Ralf S. Engelschall]
10398 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10399 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10401 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10405 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10406 DH parameters/keys (q is lost during that conversion, but the resulting
10407 DH parameters contain its length).
10409 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10410 much faster than DH_generate_parameters (which creates parameters
10411 where p = 2*q + 1), and also the smaller q makes DH computations
10412 much more efficient (160-bit exponentiation instead of 1024-bit
10413 exponentiation); so this provides a convenient way to support DHE
10414 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10415 utter importance to use
10416 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10418 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10419 when such DH parameters are used, because otherwise small subgroup
10420 attacks may become possible!
10423 *) Avoid memory leak in i2d_DHparams.
10426 *) Allow the -k option to be used more than once in the enc program:
10427 this allows the same encrypted message to be read by multiple recipients.
10430 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10431 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10432 it will always use the numerical form of the OID, even if it has a short
10436 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10437 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10438 otherwise bn_mod_exp was called. In the case of hardware keys for example
10439 no private key components need be present and it might store extra data
10440 in the RSA structure, which cannot be accessed from bn_mod_exp.
10441 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10442 private key operations.
10445 *) Added support for SPARC Linux.
10448 *) pem_password_cb function type incompatibly changed from
10449 typedef int pem_password_cb(char *buf, int size, int rwflag);
10451 ....(char *buf, int size, int rwflag, void *userdata);
10452 so that applications can pass data to their callbacks:
10453 The PEM[_ASN1]_{read,write}... functions and macros now take an
10454 additional void * argument, which is just handed through whenever
10455 the password callback is called.
10456 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10458 New function SSL_CTX_set_default_passwd_cb_userdata.
10460 Compatibility note: As many C implementations push function arguments
10461 onto the stack in reverse order, the new library version is likely to
10462 interoperate with programs that have been compiled with the old
10463 pem_password_cb definition (PEM_whatever takes some data that
10464 happens to be on the stack as its last argument, and the callback
10465 just ignores this garbage); but there is no guarantee whatsoever that
10468 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10469 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10470 problems not only on Windows, but also on some Unix platforms.
10471 To avoid problematic command lines, these definitions are now in an
10472 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10473 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10476 *) MIPS III/IV assembler module is reimplemented.
10479 *) More DES library cleanups: remove references to srand/rand and
10480 delete an unused file.
10483 *) Add support for the the free Netwide assembler (NASM) under Win32,
10484 since not many people have MASM (ml) and it can be hard to obtain.
10485 This is currently experimental but it seems to work OK and pass all
10486 the tests. Check out INSTALL.W32 for info.
10489 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10490 without temporary keys kept an extra copy of the server key,
10491 and connections with temporary keys did not free everything in case
10495 *) New function RSA_check_key and new openssl rsa option -check
10496 for verifying the consistency of RSA keys.
10497 [Ulf Moeller, Bodo Moeller]
10499 *) Various changes to make Win32 compile work:
10500 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10501 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10502 comparison" warnings.
10503 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10506 *) Add a debugging option to PKCS#5 v2 key generation function: when
10507 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10508 derived keys are printed to stderr.
10511 *) Copy the flags in ASN1_STRING_dup().
10512 [Roman E. Pavlov <pre@mo.msk.ru>]
10514 *) The x509 application mishandled signing requests containing DSA
10515 keys when the signing key was also DSA and the parameters didn't match.
10517 It was supposed to omit the parameters when they matched the signing key:
10518 the verifying software was then supposed to automatically use the CA's
10519 parameters if they were absent from the end user certificate.
10521 Omitting parameters is no longer recommended. The test was also
10522 the wrong way round! This was probably due to unusual behaviour in
10523 EVP_cmp_parameters() which returns 1 if the parameters match.
10524 This meant that parameters were omitted when they *didn't* match and
10525 the certificate was useless. Certificates signed with 'ca' didn't have
10527 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10529 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10530 The interface is as follows:
10531 Applications can use
10532 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10533 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10534 "off" is now the default.
10535 The library internally uses
10536 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10537 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10538 to disable memory-checking temporarily.
10540 Some inconsistent states that previously were possible (and were
10541 even the default) are now avoided.
10543 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10544 with each memory chunk allocated; this is occasionally more helpful
10545 than just having a counter.
10547 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10549 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10553 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10554 which largely parallels "options", but is for changing API behaviour,
10555 whereas "options" are about protocol behaviour.
10556 Initial "mode" flags are:
10558 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10559 a single record has been written.
10560 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10561 retries use the same buffer location.
10562 (But all of the contents must be
10566 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10569 *) Fix problems with no-hmac etc.
10570 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10572 *) New functions RSA_get_default_method(), RSA_set_method() and
10573 RSA_get_method(). These allows replacement of RSA_METHODs without having
10574 to mess around with the internals of an RSA structure.
10577 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10578 Also really enable memory leak checks in openssl.c and in some
10580 [Chad C. Mulligan, Bodo Moeller]
10582 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10583 up the length of negative integers. This has now been simplified to just
10584 store the length when it is first determined and use it later, rather
10585 than trying to keep track of where data is copied and updating it to
10587 [Steve Henson, reported by Brien Wheeler
10588 <bwheeler@authentica-security.com>]
10590 *) Add a new function PKCS7_signatureVerify. This allows the verification
10591 of a PKCS#7 signature but with the signing certificate passed to the
10592 function itself. This contrasts with PKCS7_dataVerify which assumes the
10593 certificate is present in the PKCS#7 structure. This isn't always the
10594 case: certificates can be omitted from a PKCS#7 structure and be
10595 distributed by "out of band" means (such as a certificate database).
10598 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10599 function prototypes in pem.h, also change util/mkdef.pl to add the
10600 necessary function names.
10603 *) mk1mf.pl (used by Windows builds) did not properly read the
10604 options set by Configure in the top level Makefile, and Configure
10605 was not even able to write more than one option correctly.
10606 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10609 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10610 file to be loaded from a BIO or FILE pointer. The BIO version will
10611 for example allow memory BIOs to contain config info.
10614 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10615 Whoever hopes to achieve shared-library compatibility across versions
10616 must use this, not the compile-time macro.
10617 (Exercise 0.9.4: Which is the minimum library version required by
10619 Note: All this applies only to multi-threaded programs, others don't
10623 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10624 through a BIO pair triggered the default case, i.e.
10625 SSLerr(...,SSL_R_UNKNOWN_STATE).
10628 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10629 can use the SSL library even if none of the specific BIOs is
10633 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10634 for the encoded length.
10635 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10637 *) Add initial documentation of the X509V3 functions.
10640 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10641 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10642 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10643 secure PKCS#8 private key format with a high iteration count.
10646 *) Fix determination of Perl interpreter: A perl or perl5
10647 _directory_ in $PATH was also accepted as the interpreter.
10648 [Ralf S. Engelschall]
10650 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10651 wrong with it but it was very old and did things like calling
10652 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10653 unusual formatting.
10656 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10657 to use the new extension code.
10660 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10661 with macros. This should make it easier to change their form, add extra
10662 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10666 *) Add to configuration table a new entry that can specify an alternative
10667 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10668 according to Mark Crispin <MRC@Panda.COM>.
10672 *) DES CBC did not update the IV. Weird.
10675 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10676 Changing the behaviour of the former might break existing programs --
10677 where IV updating is needed, des_ncbc_encrypt can be used.
10680 *) When bntest is run from "make test" it drives bc to check its
10681 calculations, as well as internally checking them. If an internal check
10682 fails, it needs to cause bc to give a non-zero result or make test carries
10683 on without noticing the failure. Fixed.
10686 *) DES library cleanups.
10689 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10690 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10691 ciphers. NOTE: although the key derivation function has been verified
10692 against some published test vectors it has not been extensively tested
10693 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10697 *) Instead of "mkdir -p", which is not fully portable, use new
10698 Perl script "util/mkdir-p.pl".
10701 *) Rewrite the way password based encryption (PBE) is handled. It used to
10702 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10703 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10704 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10705 the 'parameter' field of the AlgorithmIdentifier is passed to the
10706 underlying key generation function so it must do its own ASN1 parsing.
10707 This has also changed the EVP_PBE_CipherInit() function which now has a
10708 'parameter' argument instead of literal salt and iteration count values
10709 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10712 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10713 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10714 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10715 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10716 value was just used as a "magic string" and not used directly its
10717 value doesn't matter.
10720 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10724 *) "linux-sparc64" configuration (ultrapenguin).
10725 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10726 "linux-sparc" configuration.
10727 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10729 *) config now generates no-xxx options for missing ciphers.
10732 *) Support the EBCDIC character set (work in progress).
10733 File ebcdic.c not yet included because it has a different license.
10734 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10736 *) Support BS2000/OSD-POSIX.
10737 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10739 *) Make callbacks for key generation use void * instead of char *.
10742 *) Make S/MIME samples compile (not yet tested).
10745 *) Additional typesafe stacks.
10748 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10752 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10754 *) New configuration variant "sco5-gcc".
10756 *) Updated some demos.
10757 [Sean O Riordain, Wade Scholine]
10759 *) Add missing BIO_free at exit of pkcs12 application.
10762 *) Fix memory leak in conf.c.
10765 *) Updates for Win32 to assembler version of MD5.
10768 *) Set #! path to perl in apps/der_chop to where we found it
10769 instead of using a fixed path.
10772 *) SHA library changes for irix64-mips4-cc.
10775 *) Improvements for VMS support.
10779 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10781 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10782 This also avoids the problems with SC4.2 and unpatched SC5.
10783 [Andy Polyakov <appro@fy.chalmers.se>]
10785 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10786 These are required because of the typesafe stack would otherwise break
10787 existing code. If old code used a structure member which used to be STACK
10788 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10789 sk_num or sk_value it would produce an error because the num, data members
10790 are not present in STACK_OF. Now it just produces a warning. sk_set
10791 replaces the old method of assigning a value to sk_value
10792 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10793 that does this will no longer work (and should use sk_set instead) but
10794 this could be regarded as a "questionable" behaviour anyway.
10797 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10798 correctly handle encrypted S/MIME data.
10801 *) Change type of various DES function arguments from des_cblock
10802 (which means, in function argument declarations, pointer to char)
10803 to des_cblock * (meaning pointer to array with 8 char elements),
10804 which allows the compiler to do more typechecking; it was like
10805 that back in SSLeay, but with lots of ugly casts.
10807 Introduce new type const_des_cblock.
10810 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10811 problems: find RecipientInfo structure that matches recipient certificate
10812 and initialise the ASN1 structures properly based on passed cipher.
10815 *) Belatedly make the BN tests actually check the results.
10818 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10819 to and from BNs: it was completely broken. New compilation option
10820 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10821 key elements as negative integers.
10824 *) Reorganize and speed up MD5.
10825 [Andy Polyakov <appro@fy.chalmers.se>]
10828 [Richard Levitte <richard@levitte.org>]
10830 *) New option -out to asn1parse to allow the parsed structure to be
10831 output to a file. This is most useful when combined with the -strparse
10832 option to examine the output of things like OCTET STRINGS.
10835 *) Make SSL library a little more fool-proof by not requiring any longer
10836 that SSL_set_{accept,connect}_state be called before
10837 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10838 in many applications because usually everything *appeared* to work as
10839 intended anyway -- now it really works as intended).
10842 *) Move openssl.cnf out of lib/.
10845 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10846 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10847 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10848 [Ralf S. Engelschall]
10850 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10851 handle PKCS#7 enveloped data properly.
10852 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10854 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10855 copying pointers. The cert_st handling is changed by this in
10856 various ways (and thus what used to be known as ctx->default_cert
10857 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10858 any longer when s->cert does not give us what we need).
10859 ssl_cert_instantiate becomes obsolete by this change.
10860 As soon as we've got the new code right (possibly it already is?),
10861 we have solved a couple of bugs of the earlier code where s->cert
10862 was used as if it could not have been shared with other SSL structures.
10864 Note that using the SSL API in certain dirty ways now will result
10865 in different behaviour than observed with earlier library versions:
10866 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10867 does not influence s as it used to.
10869 In order to clean up things more thoroughly, inside SSL_SESSION
10870 we don't use CERT any longer, but a new structure SESS_CERT
10871 that holds per-session data (if available); currently, this is
10872 the peer's certificate chain and, for clients, the server's certificate
10873 and temporary key. CERT holds only those values that can have
10874 meaningful defaults in an SSL_CTX.
10877 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10878 from the internal representation. Various PKCS#7 fixes: remove some
10879 evil casts and set the enc_dig_alg field properly based on the signing
10883 *) Allow PKCS#12 password to be set from the command line or the
10884 environment. Let 'ca' get its config file name from the environment
10885 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10889 *) Allow certificate policies extension to use an IA5STRING for the
10890 organization field. This is contrary to the PKIX definition but
10891 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10895 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10896 without disallowing inline assembler and the like for non-pedantic builds.
10899 *) Support Borland C++ builder.
10900 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10902 *) Support Mingw32.
10905 *) SHA-1 cleanups and performance enhancements.
10906 [Andy Polyakov <appro@fy.chalmers.se>]
10908 *) Sparc v8plus assembler for the bignum library.
10909 [Andy Polyakov <appro@fy.chalmers.se>]
10911 *) Accept any -xxx and +xxx compiler options in Configure.
10914 *) Update HPUX configuration.
10917 *) Add missing sk_<type>_unshift() function to safestack.h
10918 [Ralf S. Engelschall]
10920 *) New function SSL_CTX_use_certificate_chain_file that sets the
10921 "extra_cert"s in addition to the certificate. (This makes sense
10922 only for "PEM" format files, as chains as a whole are not
10926 *) Support verify_depth from the SSL API.
10927 x509_vfy.c had what can be considered an off-by-one-error:
10928 Its depth (which was not part of the external interface)
10929 was actually counting the number of certificates in a chain;
10930 now it really counts the depth.
10933 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10934 instead of X509err, which often resulted in confusing error
10935 messages since the error codes are not globally unique
10936 (e.g. an alleged error in ssl3_accept when a certificate
10937 didn't match the private key).
10939 *) New function SSL_CTX_set_session_id_context that allows to set a default
10940 value (so that you don't need SSL_set_session_id_context for each
10941 connection using the SSL_CTX).
10944 *) OAEP decoding bug fix.
10947 *) Support INSTALL_PREFIX for package builders, as proposed by
10951 *) New Configure options "threads" and "no-threads". For systems
10952 where the proper compiler options are known (currently Solaris
10953 and Linux), "threads" is the default.
10956 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10959 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10960 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10961 such as /usr/local/bin.
10964 *) "make linux-shared" to build shared libraries.
10965 [Niels Poppe <niels@netbox.org>]
10967 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10970 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10971 extension adding in x509 utility.
10974 *) Remove NOPROTO sections and error code comments.
10977 *) Partial rewrite of the DEF file generator to now parse the ANSI
10981 *) New Configure options --prefix=DIR and --openssldir=DIR.
10984 *) Complete rewrite of the error code script(s). It is all now handled
10985 by one script at the top level which handles error code gathering,
10986 header rewriting and C source file generation. It should be much better
10987 than the old method: it now uses a modified version of Ulf's parser to
10988 read the ANSI prototypes in all header files (thus the old K&R definitions
10989 aren't needed for error creation any more) and do a better job of
10990 translating function codes into names. The old 'ASN1 error code imbedded
10991 in a comment' is no longer necessary and it doesn't use .err files which
10992 have now been deleted. Also the error code call doesn't have to appear all
10993 on one line (which resulted in some large lines...).
10996 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10999 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11000 0 (which usually indicates a closed connection), but continue reading.
11003 *) Fix some race conditions.
11006 *) Add support for CRL distribution points extension. Add Certificate
11007 Policies and CRL distribution points documentation.
11010 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11013 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11014 8 of keying material. Merlin has also confirmed interop with this fix
11015 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11016 [Merlin Hughes <merlin@baltimore.ie>]
11018 *) Fix lots of warnings.
11019 [Richard Levitte <levitte@stacken.kth.se>]
11021 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11022 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11023 [Richard Levitte <levitte@stacken.kth.se>]
11025 *) Fix problems with sizeof(long) == 8.
11026 [Andy Polyakov <appro@fy.chalmers.se>]
11028 *) Change functions to ANSI C.
11031 *) Fix typos in error codes.
11032 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11034 *) Remove defunct assembler files from Configure.
11037 *) SPARC v8 assembler BIGNUM implementation.
11038 [Andy Polyakov <appro@fy.chalmers.se>]
11040 *) Support for Certificate Policies extension: both print and set.
11041 Various additions to support the r2i method this uses.
11044 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11045 return a const string when you are expecting an allocated buffer.
11048 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11049 types DirectoryString and DisplayText.
11052 *) Add code to allow r2i extensions to access the configuration database,
11053 add an LHASH database driver and add several ctx helper functions.
11056 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11057 fail when they extended the size of a BIGNUM.
11060 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11061 support typesafe stack.
11064 *) Fix typo in SSL_[gs]et_options().
11065 [Nils Frostberg <nils@medcom.se>]
11067 *) Delete various functions and files that belonged to the (now obsolete)
11068 old X509V3 handling code.
11071 *) New Configure option "rsaref".
11074 *) Don't auto-generate pem.h.
11077 *) Introduce type-safe ASN.1 SETs.
11080 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11081 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11083 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11084 that links with OpenSSL (well at least cause lots of warnings), but fear
11085 not: the conversion is trivial, and it eliminates loads of evil casts. A
11086 few STACKed things have been converted already. Feel free to convert more.
11087 In the fullness of time, I'll do away with the STACK type altogether.
11090 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11091 specified in <certfile> by updating the entry in the index.txt file.
11092 This way one no longer has to edit the index.txt file manually for
11093 revoking a certificate. The -revoke option does the gory details now.
11094 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11096 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11097 `-text' option at all and this way the `-noout -text' combination was
11098 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11099 [Ralf S. Engelschall]
11101 *) Make sure a corresponding plain text error message exists for the
11102 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11103 verify callback function determined that a certificate was revoked.
11104 [Ralf S. Engelschall]
11106 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11107 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11108 all available cipers including rc5, which was forgotten until now.
11109 In order to let the testing shell script know which algorithms
11110 are available, a new (up to now undocumented) command
11111 "openssl list-cipher-commands" is used.
11114 *) Bugfix: s_client occasionally would sleep in select() when
11115 it should have checked SSL_pending() first.
11118 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11119 the raw DSA values prior to ASN.1 encoding.
11122 *) Tweaks to Configure
11123 [Niels Poppe <niels@netbox.org>]
11125 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11129 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11132 *) New config option to avoid instructions that are illegal on the 80386.
11133 The default code is faster, but requires at least a 486.
11136 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11137 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11138 same as SSL2_VERSION anyway.
11141 *) New "-showcerts" option for s_client.
11144 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11145 application. Various cleanups and fixes.
11148 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11149 modify error routines to work internally. Add error codes and PBE init
11150 to library startup routines.
11153 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11154 packing functions to asn1 and evp. Changed function names and error
11155 codes along the way.
11158 *) PKCS12 integration: and so it begins... First of several patches to
11159 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11160 objects to objects.h
11163 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11164 and display support for Thawte strong extranet extension.
11167 *) Add LinuxPPC support.
11168 [Jeff Dubrule <igor@pobox.org>]
11170 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11171 bn_div_words in alpha.s.
11172 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11174 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11175 OAEP isn't supported when OpenSSL is built with RSAref.
11176 [Ulf Moeller <ulf@fitug.de>]
11178 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11179 so they no longer are missing under -DNOPROTO.
11180 [Soren S. Jorvang <soren@t.dk>]
11183 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11185 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11186 doesn't work when the session is reused. Coming soon!
11189 *) Fix a security hole, that allows sessions to be reused in the wrong
11190 context thus bypassing client cert protection! All software that uses
11191 client certs and session caches in multiple contexts NEEDS PATCHING to
11192 allow session reuse! A fuller solution is in the works.
11193 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11195 *) Some more source tree cleanups (removed obsolete files
11196 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11197 permission on "config" script to be executable) and a fix for the INSTALL
11199 [Ulf Moeller <ulf@fitug.de>]
11201 *) Remove some legacy and erroneous uses of malloc, free instead of
11203 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11205 *) Make rsa_oaep_test return non-zero on error.
11206 [Ulf Moeller <ulf@fitug.de>]
11208 *) Add support for native Solaris shared libraries. Configure
11209 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11210 if someone would make that last step automatic.
11211 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11213 *) ctx_size was not built with the right compiler during "make links". Fixed.
11216 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11217 except NULL ciphers". This means the default cipher list will no longer
11218 enable NULL ciphers. They need to be specifically enabled e.g. with
11219 the string "DEFAULT:eNULL".
11222 *) Fix to RSA private encryption routines: if p < q then it would
11223 occasionally produce an invalid result. This will only happen with
11224 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11227 *) Be less restrictive and allow also `perl util/perlpath.pl
11228 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11229 because this way one can also use an interpreter named `perl5' (which is
11230 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11231 installed as `perl').
11232 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11234 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11235 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11237 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11238 advapi32.lib to Win32 build and change the pem test comparison
11239 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11240 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11241 and crypto/des/ede_cbcm_enc.c.
11244 *) DES quad checksum was broken on big-endian architectures. Fixed.
11247 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11248 Win32 test batch file so it (might) work again. The Win32 test batch file
11249 is horrible: I feel ill....
11252 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11253 in e_os.h. Audit of header files to check ANSI and non ANSI
11254 sections: 10 functions were absent from non ANSI section and not exported
11255 from Windows DLLs. Fixed up libeay.num for new functions.
11258 *) Make `openssl version' output lines consistent.
11259 [Ralf S. Engelschall]
11261 *) Fix Win32 symbol export lists for BIO functions: Added
11262 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11263 to ms/libeay{16,32}.def.
11264 [Ralf S. Engelschall]
11266 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11267 fine under Unix and passes some trivial tests I've now added. But the
11268 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11269 added to make sure no one expects that this stuff really works in the
11270 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11271 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11273 [Ralf S. Engelschall]
11275 *) Fix the generation of two part addresses in perl.
11276 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11278 *) Add config entry for Linux on MIPS.
11279 [John Tobey <jtobey@channel1.com>]
11281 *) Make links whenever Configure is run, unless we are on Windoze.
11284 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11285 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11289 *) Add a useful kludge to allow package maintainers to specify compiler and
11290 other platforms details on the command line without having to patch the
11291 Configure script everytime: One now can use ``perl Configure
11292 <id>:<details>'', i.e. platform ids are allowed to have details appended
11293 to them (separated by colons). This is treated as there would be a static
11294 pre-configured entry in Configure's %table under key <id> with value
11295 <details> and ``perl Configure <id>'' is called. So, when you want to
11296 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11297 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11298 now, which overrides the FreeBSD-elf entry on-the-fly.
11299 [Ralf S. Engelschall]
11301 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11304 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11305 on the `perl Configure ...' command line. This way one can compile
11306 OpenSSL libraries with Position Independent Code (PIC) which is needed
11307 for linking it into DSOs.
11308 [Ralf S. Engelschall]
11310 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11314 *) Cleaned up the LICENSE document: The official contact for any license
11315 questions now is the OpenSSL core team under openssl-core@openssl.org.
11316 And add a paragraph about the dual-license situation to make sure people
11317 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11318 to the OpenSSL toolkit.
11319 [Ralf S. Engelschall]
11321 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11322 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11323 Additionally cleaned up the `make links' target: Remove unnecessary
11324 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11325 to speed processing and no longer clutter the display with confusing
11326 stuff. Instead only the actually done links are displayed.
11327 [Ralf S. Engelschall]
11329 *) Permit null encryption ciphersuites, used for authentication only. It used
11330 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11331 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11335 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11336 signed attributes when verifying signatures (this would break them),
11337 the detached data encoding was wrong and public keys obtained using
11338 X509_get_pubkey() weren't freed.
11341 *) Add text documentation for the BUFFER functions. Also added a work around
11342 to a Win95 console bug. This was triggered by the password read stuff: the
11343 last character typed gets carried over to the next fread(). If you were
11344 generating a new cert request using 'req' for example then the last
11345 character of the passphrase would be CR which would then enter the first
11349 *) Added the new `Includes OpenSSL Cryptography Software' button as
11350 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11351 button and can be used by applications based on OpenSSL to show the
11352 relationship to the OpenSSL project.
11353 [Ralf S. Engelschall]
11355 *) Remove confusing variables in function signatures in files
11356 ssl/ssl_lib.c and ssl/ssl.h.
11357 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11359 *) Don't install bss_file.c under PREFIX/include/
11360 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11362 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11363 functions that return function pointers and has support for NT specific
11364 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11365 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11366 unsigned to signed types: this was killing the Win32 compile.
11369 *) Add new certificate file to stack functions,
11370 SSL_add_dir_cert_subjects_to_stack() and
11371 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11372 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11373 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11374 This means that Apache-SSL and similar packages don't have to mess around
11375 to add as many CAs as they want to the preferred list.
11378 *) Experiment with doxygen documentation. Currently only partially applied to
11380 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11381 openssl.doxy as the configuration file.
11384 *) Get rid of remaining C++-style comments which strict C compilers hate.
11385 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11387 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11388 compiled in by default: it has problems with large keys.
11391 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11392 DH private keys and/or callback functions which directly correspond to
11393 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11394 is needed for applications which have to configure certificates on a
11395 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11397 For the RSA certificate situation is makes no difference, but
11398 for the DSA certificate situation this fixes the "no shared cipher"
11399 problem where the OpenSSL cipher selection procedure failed because the
11400 temporary keys were not overtaken from the context and the API provided
11401 no way to reconfigure them.
11402 The new functions now let applications reconfigure the stuff and they
11403 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11404 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11405 non-public-API function ssl_cert_instantiate() is used as a helper
11406 function and also to reduce code redundancy inside ssl_rsa.c.
11407 [Ralf S. Engelschall]
11409 *) Move s_server -dcert and -dkey options out of the undocumented feature
11410 area because they are useful for the DSA situation and should be
11411 recognized by the users.
11412 [Ralf S. Engelschall]
11414 *) Fix the cipher decision scheme for export ciphers: the export bits are
11415 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11416 SSL_EXP_MASK. So, the original variable has to be used instead of the
11417 already masked variable.
11418 [Richard Levitte <levitte@stacken.kth.se>]
11420 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11421 [Richard Levitte <levitte@stacken.kth.se>]
11423 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11424 from `int' to `unsigned int' because it's a length and initialized by
11425 EVP_DigestFinal() which expects an `unsigned int *'.
11426 [Richard Levitte <levitte@stacken.kth.se>]
11428 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11429 script. Instead use the usual Shell->Perl transition trick.
11430 [Ralf S. Engelschall]
11432 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11433 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11434 -noout -modulus' as it's already the case for `openssl rsa -noout
11435 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11436 currently the public key is printed (a decision which was already done by
11437 `openssl dsa -modulus' in the past) which serves a similar purpose.
11438 Additionally the NO_RSA no longer completely removes the whole -modulus
11439 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11441 [Ralf S. Engelschall]
11443 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11444 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11445 [Arne Ansper <arne@ats.cyber.ee>]
11447 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11448 to be added. Now both 'req' and 'ca' can use new objects defined in the
11452 *) Add cool BIO that does syslog (or event log on NT).
11453 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11455 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11456 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11457 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11458 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11461 *) Add preliminary config info for new extension code.
11464 *) Make RSA_NO_PADDING really use no padding.
11465 [Ulf Moeller <ulf@fitug.de>]
11467 *) Generate errors when private/public key check is done.
11470 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11471 for some CRL extensions and new objects added.
11474 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11475 key usage extension and fuller support for authority key id.
11478 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11479 padding method for RSA, which is recommended for new applications in PKCS
11480 #1 v2.0 (RFC 2437, October 1998).
11481 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11482 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11483 against Bleichbacher's attack on RSA.
11484 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11487 *) Updates to the new SSL compression code
11488 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11490 *) Fix so that the version number in the master secret, when passed
11491 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11492 (because the server will not accept higher), that the version number
11493 is 0x03,0x01, not 0x03,0x00
11494 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11496 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11497 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11498 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11501 *) Support for RAW extensions where an arbitrary extension can be
11502 created by including its DER encoding. See apps/openssl.cnf for
11506 *) Make sure latest Perl versions don't interpret some generated C array
11507 code as Perl array code in the crypto/err/err_genc.pl script.
11508 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11510 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11511 not many people have the assembler. Various Win32 compilation fixes and
11512 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11513 build instructions.
11516 *) Modify configure script 'Configure' to automatically create crypto/date.h
11517 file under Win32 and also build pem.h from pem.org. New script
11518 util/mkfiles.pl to create the MINFO file on environments that can't do a
11519 'make files': perl util/mkfiles.pl >MINFO should work.
11522 *) Major rework of DES function declarations, in the pursuit of correctness
11523 and purity. As a result, many evil casts evaporated, and some weirdness,
11524 too. You may find this causes warnings in your code. Zapping your evil
11525 casts will probably fix them. Mostly.
11528 *) Fix for a typo in asn1.h. Bug fix to object creation script
11529 obj_dat.pl. It considered a zero in an object definition to mean
11530 "end of object": none of the objects in objects.h have any zeros
11531 so it wasn't spotted.
11532 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11534 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11535 Masking (CBCM). In the absence of test vectors, the best I have been able
11536 to do is check that the decrypt undoes the encrypt, so far. Send me test
11537 vectors if you have them.
11540 *) Correct calculation of key length for export ciphers (too much space was
11541 allocated for null ciphers). This has not been tested!
11544 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11545 message is now correct (it understands "crypto" and "ssl" on its
11546 command line). There is also now an "update" option. This will update
11547 the util/ssleay.num and util/libeay.num files with any new functions.
11549 perl util/mkdef.pl crypto ssl update
11550 it will update them.
11553 *) Overhauled the Perl interface (perl/*):
11554 - ported BN stuff to OpenSSL's different BN library
11555 - made the perl/ source tree CVS-aware
11556 - renamed the package from SSLeay to OpenSSL (the files still contain
11557 their history because I've copied them in the repository)
11558 - removed obsolete files (the test scripts will be replaced
11559 by better Test::Harness variants in the future)
11560 [Ralf S. Engelschall]
11562 *) First cut for a very conservative source tree cleanup:
11563 1. merge various obsolete readme texts into doc/ssleay.txt
11564 where we collect the old documents and readme texts.
11565 2. remove the first part of files where I'm already sure that we no
11566 longer need them because of three reasons: either they are just temporary
11567 files which were left by Eric or they are preserved original files where
11568 I've verified that the diff is also available in the CVS via "cvs diff
11569 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11570 the crypto/md/ stuff).
11571 [Ralf S. Engelschall]
11573 *) More extension code. Incomplete support for subject and issuer alt
11574 name, issuer and authority key id. Change the i2v function parameters
11575 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11576 what that's for :-) Fix to ASN1 macro which messed up
11577 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11580 *) Preliminary support for ENUMERATED type. This is largely copied from the
11584 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11585 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11587 *) Make sure `make rehash' target really finds the `openssl' program.
11588 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11590 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11591 like to hear about it if this slows down other processors.
11594 *) Add CygWin32 platform information to Configure script.
11595 [Alan Batie <batie@aahz.jf.intel.com>]
11597 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11598 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11600 *) New program nseq to manipulate netscape certificate sequences
11603 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11607 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11608 but the BN code had some problems that would cause failures when
11609 doing certificate verification and some other functions.
11610 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11612 *) Add ASN1 and PEM code to support netscape certificate sequences.
11615 *) Add ASN1 and PEM code to support netscape certificate sequences.
11618 *) Add several PKIX and private extended key usage OIDs.
11621 *) Modify the 'ca' program to handle the new extension code. Modify
11622 openssl.cnf for new extension format, add comments.
11625 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11626 and add a sample to openssl.cnf so req -x509 now adds appropriate
11630 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11631 error code, add initial support to X509_print() and x509 application.
11634 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11635 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11636 stuff is currently isolated and isn't even compiled yet.
11639 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11640 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11641 Removed the versions check from X509 routines when loading extensions:
11642 this allows certain broken certificates that don't set the version
11643 properly to be processed.
11646 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11647 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11648 can still be regenerated with "make depend".
11651 *) Spelling mistake in C version of CAST-128.
11652 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11654 *) Changes to the error generation code. The perl script err-code.pl
11655 now reads in the old error codes and retains the old numbers, only
11656 adding new ones if necessary. It also only changes the .err files if new
11657 codes are added. The makefiles have been modified to only insert errors
11658 when needed (to avoid needlessly modifying header files). This is done
11659 by only inserting errors if the .err file is newer than the auto generated
11660 C file. To rebuild all the error codes from scratch (the old behaviour)
11661 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11662 or delete all the .err files.
11665 *) CAST-128 was incorrectly implemented for short keys. The C version has
11666 been fixed, but is untested. The assembler versions are also fixed, but
11667 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11668 to regenerate it if needed.
11669 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11670 Hagino <itojun@kame.net>]
11672 *) File was opened incorrectly in randfile.c.
11673 [Ulf Möller <ulf@fitug.de>]
11675 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11676 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11677 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11678 al: it's just almost always a UTCTime. Note this patch adds new error
11679 codes so do a "make errors" if there are problems.
11682 *) Correct Linux 1 recognition in config.
11683 [Ulf Möller <ulf@fitug.de>]
11685 *) Remove pointless MD5 hash when using DSA keys in ca.
11686 [Anonymous <nobody@replay.com>]
11688 *) Generate an error if given an empty string as a cert directory. Also
11689 generate an error if handed NULL (previously returned 0 to indicate an
11690 error, but didn't set one).
11691 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11693 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11696 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11697 parameters. This was causing a warning which killed off the Win32 compile.
11700 *) Remove C++ style comments from crypto/bn/bn_local.h.
11701 [Neil Costigan <neil.costigan@celocom.com>]
11703 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11704 based on a text string, looking up short and long names and finally
11705 "dot" format. The "dot" format stuff didn't work. Added new function
11706 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11707 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11708 OID is not part of the table.
11711 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11712 X509_LOOKUP_by_alias().
11715 *) Sort openssl functions by name.
11718 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11719 encryption from sample DSA keys (in case anyone is interested the password
11723 *) Make _all_ *_free functions accept a NULL pointer.
11724 [Frans Heymans <fheymans@isaserver.be>]
11726 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11728 [Anonymous <nobody@replay.com>]
11730 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11731 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11733 *) Don't blow it for numeric -newkey arguments to apps/req.
11734 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11736 *) Temp key "for export" tests were wrong in s3_srvr.c.
11737 [Anonymous <nobody@replay.com>]
11739 *) Add prototype for temp key callback functions
11740 SSL_CTX_set_tmp_{rsa,dh}_callback().
11743 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11744 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11747 *) X509_name_add_entry() freed the wrong thing after an error.
11748 [Arne Ansper <arne@ats.cyber.ee>]
11750 *) rsa_eay.c would attempt to free a NULL context.
11751 [Arne Ansper <arne@ats.cyber.ee>]
11753 *) BIO_s_socket() had a broken should_retry() on Windoze.
11754 [Arne Ansper <arne@ats.cyber.ee>]
11756 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11757 [Arne Ansper <arne@ats.cyber.ee>]
11759 *) Make sure the already existing X509_STORE->depth variable is initialized
11760 in X509_STORE_new(), but document the fact that this variable is still
11761 unused in the certificate verification process.
11762 [Ralf S. Engelschall]
11764 *) Fix the various library and apps files to free up pkeys obtained from
11765 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11768 *) Fix reference counting in X509_PUBKEY_get(). This makes
11769 demos/maurice/example2.c work, amongst others, probably.
11770 [Steve Henson and Ben Laurie]
11772 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11773 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11774 are no longer created. This way we have a single and consistent command
11775 line interface `openssl <command>', similar to `cvs <command>'.
11776 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11778 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11779 BIT STRING wrapper always have zero unused bits.
11782 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11785 *) Make the top-level INSTALL documentation easier to understand.
11788 *) Makefiles updated to exit if an error occurs in a sub-directory
11789 make (including if user presses ^C) [Paul Sutton]
11791 *) Make Montgomery context stuff explicit in RSA data structure.
11794 *) Fix build order of pem and err to allow for generated pem.h.
11797 *) Fix renumbering bug in X509_NAME_delete_entry().
11800 *) Enhanced the err-ins.pl script so it makes the error library number
11801 global and can add a library name. This is needed for external ASN1 and
11802 other error libraries.
11805 *) Fixed sk_insert which never worked properly.
11808 *) Fix ASN1 macros so they can handle indefinite length constructed
11809 EXPLICIT tags. Some non standard certificates use these: they can now
11813 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11814 into a single doc/ssleay.txt bundle. This way the information is still
11815 preserved but no longer messes up this directory. Now it's new room for
11816 the new set of documentation files.
11817 [Ralf S. Engelschall]
11819 *) SETs were incorrectly DER encoded. This was a major pain, because they
11820 shared code with SEQUENCEs, which aren't coded the same. This means that
11821 almost everything to do with SETs or SEQUENCEs has either changed name or
11822 number of arguments.
11823 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11825 *) Fix test data to work with the above.
11828 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11829 was already fixed by Eric for 0.9.1 it seems.
11830 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11832 *) Autodetect FreeBSD3.
11835 *) Fix various bugs in Configure. This affects the following platforms:
11839 unixware-2.0-pentium
11843 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11844 before they are needed.
11847 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11851 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11853 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11854 changed SSLeay to OpenSSL in version strings.
11855 [Ralf S. Engelschall]
11857 *) Some fixups to the top-level documents.
11860 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11861 because the symlink to include/ was missing.
11862 [Ralf S. Engelschall]
11864 *) Incorporated the popular no-RSA/DSA-only patches
11865 which allow to compile a RSA-free SSLeay.
11866 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11868 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11869 when "ssleay" is still not found.
11870 [Ralf S. Engelschall]
11872 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11873 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11875 *) Updated the README file.
11876 [Ralf S. Engelschall]
11878 *) Added various .cvsignore files in the CVS repository subdirs
11879 to make a "cvs update" really silent.
11880 [Ralf S. Engelschall]
11882 *) Recompiled the error-definition header files and added
11883 missing symbols to the Win32 linker tables.
11884 [Ralf S. Engelschall]
11886 *) Cleaned up the top-level documents;
11887 o new files: CHANGES and LICENSE
11888 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11889 o merged COPYRIGHT into LICENSE
11890 o removed obsolete TODO file
11891 o renamed MICROSOFT to INSTALL.W32
11892 [Ralf S. Engelschall]
11894 *) Removed dummy files from the 0.9.1b source tree:
11895 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11896 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11897 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11898 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11899 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11900 [Ralf S. Engelschall]
11902 *) Added various platform portability fixes.
11905 *) The Genesis of the OpenSSL rpject:
11906 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11907 Young and Tim J. Hudson created while they were working for C2Net until
11909 [The OpenSSL Project]
11912 Changes between 0.9.0b and 0.9.1b [not released]
11914 *) Updated a few CA certificates under certs/
11917 *) Changed some BIGNUM api stuff.
11920 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11921 DGUX x86, Linux Alpha, etc.
11924 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11925 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11929 *) Add -strparse option to asn1pars program which parses nested
11931 [Dr Stephen Henson <shenson@bigfoot.com>]
11933 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11936 *) DSA fix for "ca" program.
11939 *) Added "-genkey" option to "dsaparam" program.
11942 *) Added RIPE MD160 (rmd160) message digest.
11945 *) Added -a (all) option to "ssleay version" command.
11948 *) Added PLATFORM define which is the id given to Configure.
11951 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11954 *) Extended the ASN.1 parser routines.
11957 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11960 *) Added a BN_CTX to the BN library.
11963 *) Fixed the weak key values in DES library
11966 *) Changed API in EVP library for cipher aliases.
11969 *) Added support for RC2/64bit cipher.
11972 *) Converted the lhash library to the crypto/mem.c functions.
11975 *) Added more recognized ASN.1 object ids.
11978 *) Added more RSA padding checks for SSL/TLS.
11981 *) Added BIO proxy/filter functionality.
11984 *) Added extra_certs to SSL_CTX which can be used
11985 send extra CA certificates to the client in the CA cert chain sending
11986 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11989 *) Now Fortezza is denied in the authentication phase because
11990 this is key exchange mechanism is not supported by SSLeay at all.
11993 *) Additional PKCS1 checks.
11996 *) Support the string "TLSv1" for all TLS v1 ciphers.
11999 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12000 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12003 *) Fixed a few memory leaks.
12006 *) Fixed various code and comment typos.
12009 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12010 bytes sent in the client random.
12011 [Edward Bishop <ebishop@spyglass.com>]