5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Added support for "pipelining". Ciphers that have the
8 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
9 encryptions/decryptions simultaneously. There are currently no built-in
10 ciphers with this property but the expectation is that engines will be able
11 to offer it to significantly improve throughput. Support has been extended
12 into libssl so that multiple records for a single connection can be
13 processed in one go (for >=TLS 1.1).
16 *) Added the AFALG engine. This is an async capable engine which is able to
17 offload work to the Linux kernel. In this initial version it only supports
18 AES128-CBC. The kernel must be version 4.1.0 or greater.
21 *) OpenSSL now uses a new threading API. It is no longer necessary to
22 set locking callbacks to use OpenSSL in a multi-threaded environment. There
23 are two supported threading models: pthreads and windows threads. It is
24 also possible to configure OpenSSL at compile time for "no-threads". The
25 old threading API should no longer be used. The functions have been
26 replaced with "no-op" compatibility macros.
27 [Alessandro Ghedini, Matt Caswell]
29 *) Modify behavior of ALPN to invoke callback after SNI/servername
30 callback, such that updates to the SSL_CTX affect ALPN.
33 *) Add SSL_CIPHER queries for authentication and key-exchange.
35 *) Modify behavior of ALPN to invoke callback after SNI/servername
36 callback, such that updates to the SSL_CTX affect ALPN.
39 *) Changes to the DEFAULT cipherlist:
40 - Prefer (EC)DHE handshakes over plain RSA.
41 - Prefer AEAD ciphers over legacy ciphers.
42 - Prefer ECDSA over RSA when both certificates are available.
43 - Prefer TLSv1.2 ciphers/PRF.
44 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
48 *) Change the ECC default curve list to be this, in order: x25519,
49 secp256r1, secp521r1, secp384r1.
52 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
53 disabled by default. They can be re-enabled using the
54 enable-weak-ssl-ciphers option to Configure.
57 *) If the server has ALPN configured, but supports no protocols that the
58 client advertises, send a fatal "no_application_protocol" alert.
59 This behaviour is SHALL in RFC 7301, though it isn't universally
60 implemented by other servers.
63 *) Add X25519 support.
64 Integrate support for X25519 into EC library. This includes support
65 for public and private key encoding using the format documented in
66 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
67 OID from that draft, encodes public keys using little endian
68 format in the ECPoint structure and private keys using
69 little endian form in the privateKey field of the ECPrivateKey
70 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
73 Note: the current version supports key generation, public and
74 private key encoding and ECDH key agreement using the EC API.
75 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
79 *) Deprecate SRP_VBASE_get_by_user.
80 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
81 In order to fix an unavoidable memory leak (CVE-2016-0798),
82 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
83 seed, even if the seed is configured.
85 Users should use SRP_VBASE_get1_by_user instead. Note that in
86 SRP_VBASE_get1_by_user, caller must free the returned value. Note
87 also that even though configuring the SRP seed attempts to hide
88 invalid usernames by continuing the handshake with fake
89 credentials, this behaviour is not constant time and no strong
90 guarantees are made that the handshake is indistinguishable from
94 *) Configuration change; it's now possible to build dynamic engines
95 without having to build shared libraries and vice versa. This
96 only applies to the engines in engines/, those in crypto/engine/
97 will always be built into libcrypto (i.e. "static").
99 Building dynamic engines is enabled by default; to disable, use
100 the configuration option "disable-dynamic-engine".
102 The only requirements for building dynamic engines are the
103 presence of the DSO module and building with position independent
104 code, so they will also automatically be disabled if configuring
105 with "disable-dso" or "disable-pic".
107 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
108 are also taken away from openssl/opensslconf.h, as they are
112 *) Configuration change; if there is a known flag to compile
113 position independent code, it will always be applied on the
114 libcrypto and libssl object files, and never on the application
115 object files. This means other libraries that use routines from
116 libcrypto / libssl can be made into shared libraries regardless
117 of how OpenSSL was configured.
119 If this isn't desirable, the configuration options "disable-pic"
120 or "no-pic" can be used to disable the use of PIC. This will
121 also disable building shared libraries and dynamic engines.
124 *) Removed JPAKE code. It was experimental and has no wide use.
127 *) The INSTALL_PREFIX Makefile variable has been renamed to
128 DESTDIR. That makes for less confusion on what this variable
129 is for. Also, the configuration option --install_prefix is
133 *) Heartbeat for TLS has been removed and is disabled by default
134 for DTLS; configure with enable-heartbeats. Code that uses the
135 old #define's might need to be updated.
136 [Emilia Käsper, Rich Salz]
138 *) Rename REF_CHECK to REF_DEBUG.
141 *) New "unified" build system
143 The "unified" build system is aimed to be a common system for all
144 platforms we support. With it comes new support for VMS.
146 This system builds supports building in a different directory tree
147 than the source tree. It produces one Makefile (for unix family
148 or lookalikes), or one descrip.mms (for VMS).
150 The source of information to make the Makefile / descrip.mms is
151 small files called 'build.info', holding the necessary
152 information for each directory with source to compile, and a
153 template in Configurations, like unix-Makefile.tmpl or
156 We rely heavily on the perl module Text::Template.
159 *) Added support for auto-initialisation and de-initialisation of the library.
160 OpenSSL no longer requires explicit init or deinit routines to be called,
161 except in certain circumstances. See the OPENSSL_init_crypto() and
162 OPENSSL_init_ssl() man pages for further information.
165 *) The arguments to the DTLSv1_listen function have changed. Specifically the
166 "peer" argument is now expected to be a BIO_ADDR object.
168 *) Rewrite of BIO networking library. The BIO library lacked consistent
169 support of IPv6, and adding it required some more extensive
170 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
171 which hold all types of addresses and chains of address information.
172 It also introduces a new API, with functions like BIO_socket,
173 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
174 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
175 have been adapted accordingly.
178 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
182 *) CRIME protection: disable compression by default, even if OpenSSL is
183 compiled with zlib enabled. Applications can still enable compression
184 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
185 using the SSL_CONF library to configure compression.
188 *) The signature of the session callback configured with
189 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
190 was explicitly marked as 'const unsigned char*' instead of
194 *) Always DPURIFY. Remove the use of uninitialized memory in the
195 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
198 *) Removed many obsolete configuration items, including
199 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
200 MD2_CHAR, MD2_INT, MD2_LONG
202 IDEA_SHORT, IDEA_LONG
203 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
204 [Rich Salz, with advice from Andy Polyakov]
206 *) Many BN internals have been moved to an internal header file.
207 [Rich Salz with help from Andy Polyakov]
209 *) Configuration and writing out the results from it has changed.
210 Files such as Makefile include/openssl/opensslconf.h and are now
211 produced through general templates, such as Makefile.in and
212 crypto/opensslconf.h.in and some help from the perl module
215 Also, the center of configuration information is no longer
216 Makefile. Instead, Configure produces a perl module in
217 configdata.pm which holds most of the config data (in the hash
218 table %config), the target data that comes from the target
219 configuration in one of the Configurations/*.conf files (in
223 *) To clarify their intended purposes, the Configure options
224 --prefix and --openssldir change their semantics, and become more
225 straightforward and less interdependent.
227 --prefix shall be used exclusively to give the location INSTALLTOP
228 where programs, scripts, libraries, include files and manuals are
229 going to be installed. The default is now /usr/local.
231 --openssldir shall be used exclusively to give the default
232 location OPENSSLDIR where certificates, private keys, CRLs are
233 managed. This is also where the default openssl.cnf gets
235 If the directory given with this option is a relative path, the
236 values of both the --prefix value and the --openssldir value will
237 be combined to become OPENSSLDIR.
238 The default for --openssldir is INSTALLTOP/ssl.
240 Anyone who uses --openssldir to specify where OpenSSL is to be
241 installed MUST change to use --prefix instead.
244 *) The GOST engine was out of date and therefore it has been removed. An up
245 to date GOST engine is now being maintained in an external repository.
246 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
247 support for GOST ciphersuites (these are only activated if a GOST engine
251 *) EGD is no longer supported by default; use enable-egd when
253 [Ben Kaduk and Rich Salz]
255 *) The distribution now has Makefile.in files, which are used to
256 create Makefile's when Configure is run. *Configure must be run
257 before trying to build now.*
260 *) The return value for SSL_CIPHER_description() for error conditions
264 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
266 Obtaining and performing DNSSEC validation of TLSA records is
267 the application's responsibility. The application provides
268 the TLSA records of its choice to OpenSSL, and these are then
269 used to authenticate the peer.
271 The TLSA records need not even come from DNS. They can, for
272 example, be used to implement local end-entity certificate or
273 trust-anchor "pinning", where the "pin" data takes the form
274 of TLSA records, which can augment or replace verification
275 based on the usual WebPKI public certification authorities.
278 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
279 continues to support deprecated interfaces in default builds.
280 However, applications are strongly advised to compile their
281 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
282 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
283 or the 1.1.0 releases.
285 In environments in which all applications have been ported to
286 not use any deprecated interfaces OpenSSL's Configure script
287 should be used with the --api=1.1.0 option to entirely remove
288 support for the deprecated features from the library and
289 unconditionally disable them in the installed headers.
290 Essentially the same effect can be achieved with the "no-deprecated"
291 argument to Configure, except that this will always restrict
292 the build to just the latest API, rather than a fixed API
295 As applications are ported to future revisions of the API,
296 they should update their compile-time OPENSSL_API_COMPAT define
297 accordingly, but in most cases should be able to continue to
298 compile with later releases.
300 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
301 0x10000000L and 0x00908000L, respectively. However those
302 versions did not support the OPENSSL_API_COMPAT feature, and
303 so applications are not typically tested for explicit support
304 of just the undeprecated features of either release.
307 *) Add support for setting the minimum and maximum supported protocol.
308 It can bet set via the SSL_set_min_proto_version() and
309 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
310 MaxProtcol. It's recommended to use the new APIs to disable
311 protocols instead of disabling individual protocols using
312 SSL_set_options() or SSL_CONF's Protocol. This change also
313 removes support for disabling TLS 1.2 in the OpenSSL TLS
314 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
317 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
320 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
321 and integrates ECDSA and ECDH functionality into EC. Implementations can
322 now redirect key generation and no longer need to convert to or from
325 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
326 include the ec.h header file instead.
329 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
330 ciphers who are no longer supported and drops support the ephemeral RSA key
331 exchange. The LOW ciphers currently doesn't have any ciphers in it.
334 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
335 opaque. For HMAC_CTX, the following constructors and destructors
338 HMAC_CTX *HMAC_CTX_new(void);
339 void HMAC_CTX_free(HMAC_CTX *ctx);
341 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
342 destroy such methods has been added. See EVP_MD_meth_new(3) and
343 EVP_CIPHER_meth_new(3) for documentation.
346 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
347 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
348 EVP_MD_CTX_reset() should be called instead to reinitialise
349 an already created structure.
350 2) For consistency with the majority of our object creators and
351 destructors, EVP_MD_CTX_(create|destroy) were renamed to
352 EVP_MD_CTX_(new|free). The old names are retained as macros
353 for deprecated builds.
356 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
357 cryptographic operations to be performed asynchronously as long as an
358 asynchronous capable engine is used. See the ASYNC_start_job() man page for
359 further details. Libssl has also had this capability integrated with the
360 introduction of the new mode SSL_MODE_ASYNC and associated error
361 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
362 pages. This work was developed in partnership with Intel Corp.
365 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
366 always enabled now. If you want to disable the support you should
367 exclude it using the list of supported ciphers. This also means that the
368 "-no_ecdhe" option has been removed from s_server.
371 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
372 SSL_{CTX_}set1_curves() which can set a list.
375 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
376 curve you want to support using SSL_{CTX_}set1_curves().
379 *) State machine rewrite. The state machine code has been significantly
380 refactored in order to remove much duplication of code and solve issues
381 with the old code (see ssl/statem/README for further details). This change
382 does have some associated API changes. Notably the SSL_state() function
383 has been removed and replaced by SSL_get_state which now returns an
384 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
385 altogether. The previous handshake states defined in ssl.h and ssl3.h have
389 *) All instances of the string "ssleay" in the public API were replaced
390 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
391 Some error codes related to internal RSA_eay API's were renamed.
394 *) The demo files in crypto/threads were moved to demo/threads.
397 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
399 [Matt Caswell, Rich Salz]
401 *) New ASN.1 embed macro.
403 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
404 structure is not allocated: it is part of the parent. That is instead of
412 This reduces memory fragmentation and make it impossible to accidentally
413 set a mandatory field to NULL.
415 This currently only works for some fields specifically a SEQUENCE, CHOICE,
416 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
417 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
421 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
424 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
425 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
426 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
427 DES and RC4 ciphersuites.
430 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
431 This changes the decoding behaviour for some invalid messages,
432 though the change is mostly in the more lenient direction, and
433 legacy behaviour is preserved as much as possible.
436 *) Fix no-stdio build.
437 [ David Woodhouse <David.Woodhouse@intel.com> and also
438 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
440 *) New testing framework
441 The testing framework has been largely rewritten and is now using
442 perl and the perl modules Test::Harness and an extended variant of
443 Test::More called OpenSSL::Test to do its work. All test scripts in
444 test/ have been rewritten into test recipes, and all direct calls to
445 executables in test/Makefile have become individual recipes using the
446 simplified testing OpenSSL::Test::Simple.
448 For documentation on our testing modules, do:
450 perldoc test/testlib/OpenSSL/Test/Simple.pm
451 perldoc test/testlib/OpenSSL/Test.pm
455 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
456 are used; the latter aborts on memory leaks (usually checked on exit).
457 Some undocumented "set malloc, etc., hooks" functions were removed
458 and others were changed. All are now documented.
461 *) In DSA_generate_parameters_ex, if the provided seed is too short,
463 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
465 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
466 from RFC4279, RFC4785, RFC5487, RFC5489.
468 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
469 original RSA_PSK patch.
472 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
473 era flag was never set throughout the codebase (only read). Also removed
474 SSL3_FLAGS_POP_BUFFER which was only used if
475 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
478 *) Changed the default name options in the "ca", "crl", "req" and "x509"
479 to be "oneline" instead of "compat".
482 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
483 not aware of clients that still exhibit this bug, and the workaround
484 hasn't been working properly for a while.
487 *) The return type of BIO_number_read() and BIO_number_written() as well as
488 the corresponding num_read and num_write members in the BIO structure has
489 changed from unsigned long to uint64_t. On platforms where an unsigned
490 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
494 *) Given the pervasive nature of TLS extensions it is inadvisable to run
495 OpenSSL without support for them. It also means that maintaining
496 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
497 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
500 *) Removed support for the two export grade static DH ciphersuites
501 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
502 were newly added (along with a number of other static DH ciphersuites) to
503 1.0.2. However the two export ones have *never* worked since they were
504 introduced. It seems strange in any case to be adding new export
505 ciphersuites, and given "logjam" it also does not seem correct to fix them.
508 *) Version negotiation has been rewritten. In particular SSLv23_method(),
509 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
510 and turned into macros which simply call the new preferred function names
511 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
512 should use the new names instead. Also as part of this change the ssl23.h
513 header file has been removed.
516 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
517 code and the associated standard is no longer considered fit-for-purpose.
520 *) RT2547 was closed. When generating a private key, try to make the
521 output file readable only by the owner. This behavior change might
522 be noticeable when interacting with other software.
524 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
528 *) Added HTTP GET support to the ocsp command.
531 *) Changed default digest for the dgst and enc commands from MD5 to
535 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
538 *) Added support for TLS extended master secret from
539 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
540 initial patch which was a great help during development.
543 *) All libssl internal structures have been removed from the public header
544 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
545 now redundant). Users should not attempt to access internal structures
546 directly. Instead they should use the provided API functions.
549 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
550 Access to deprecated functions can be re-enabled by running config with
551 "enable-deprecated". In addition applications wishing to use deprecated
552 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
553 will, by default, disable some transitive includes that previously existed
554 in the header files (e.g. ec.h will no longer, by default, include bn.h)
557 *) Added support for OCB mode. OpenSSL has been granted a patent license
558 compatible with the OpenSSL license for use of OCB. Details are available
559 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
560 for OCB can be removed by calling config with no-ocb.
563 *) SSLv2 support has been removed. It still supports receiving a SSLv2
564 compatible client hello.
567 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
568 done while fixing the error code for the key-too-small case.
569 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
571 *) CA.sh has been removmed; use CA.pl instead.
574 *) Removed old DES API.
577 *) Remove various unsupported platforms:
583 Sinix/ReliantUNIX RM400
588 16-bit platforms such as WIN16
591 *) Clean up OPENSSL_NO_xxx #define's
592 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
593 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
594 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
595 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
596 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
597 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
598 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
599 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
600 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
601 Remove MS_STATIC; it's a relic from platforms <32 bits.
604 *) Cleaned up dead code
605 Remove all but one '#ifdef undef' which is to be looked at.
608 *) Clean up calling of xxx_free routines.
609 Just like free(), fix most of the xxx_free routines to accept
610 NULL. Remove the non-null checks from callers. Save much code.
613 *) Add secure heap for storage of private keys (when possible).
614 Add BIO_s_secmem(), CBIGNUM, etc.
615 Contributed by Akamai Technologies under our Corporate CLA.
618 *) Experimental support for a new, fast, unbiased prime candidate generator,
619 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
620 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
622 *) New output format NSS in the sess_id command line tool. This allows
623 exporting the session id and the master key in NSS keylog format.
624 [Martin Kaiser <martin@kaiser.cx>]
626 *) Harmonize version and its documentation. -f flag is used to display
628 [mancha <mancha1@zoho.com>]
630 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
631 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
632 [mancha <mancha1@zoho.com>]
634 *) Fix some double frees. These are not thought to be exploitable.
635 [mancha <mancha1@zoho.com>]
637 *) A missing bounds check in the handling of the TLS heartbeat extension
638 can be used to reveal up to 64k of memory to a connected client or
641 Thanks for Neel Mehta of Google Security for discovering this bug and to
642 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
643 preparing the fix (CVE-2014-0160)
644 [Adam Langley, Bodo Moeller]
646 *) Fix for the attack described in the paper "Recovering OpenSSL
647 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
648 by Yuval Yarom and Naomi Benger. Details can be obtained from:
649 http://eprint.iacr.org/2014/140
651 Thanks to Yuval Yarom and Naomi Benger for discovering this
652 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
653 [Yuval Yarom and Naomi Benger]
655 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
656 this fixes a limitation in previous versions of OpenSSL.
659 *) Experimental encrypt-then-mac support.
661 Experimental support for encrypt then mac from
662 draft-gutmann-tls-encrypt-then-mac-02.txt
664 To enable it set the appropriate extension number (0x42 for the test
665 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
667 For non-compliant peers (i.e. just about everything) this should have no
670 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
674 *) Add EVP support for key wrapping algorithms, to avoid problems with
675 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
676 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
677 algorithms and include tests cases.
680 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
684 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
685 MGF1 digest and OAEP label.
688 *) Make openssl verify return errors.
689 [Chris Palmer <palmer@google.com> and Ben Laurie]
691 *) New function ASN1_TIME_diff to calculate the difference between two
692 ASN1_TIME structures or one structure and the current time.
695 *) Update fips_test_suite to support multiple command line options. New
696 test to induce all self test errors in sequence and check expected
700 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
701 sign or verify all in one operation.
704 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
705 test programs and fips_test_suite. Includes functionality to parse
706 the minimal script output of fipsalgest.pl directly.
709 *) Add authorisation parameter to FIPS_module_mode_set().
712 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
715 *) Use separate DRBG fields for internal and external flags. New function
716 FIPS_drbg_health_check() to perform on demand health checking. Add
717 generation tests to fips_test_suite with reduced health check interval to
718 demonstrate periodic health checking. Add "nodh" option to
719 fips_test_suite to skip very slow DH test.
722 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
726 *) More extensive health check for DRBG checking many more failure modes.
727 New function FIPS_selftest_drbg_all() to handle every possible DRBG
728 combination: call this in fips_test_suite.
731 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
732 and POST to handle Dual EC cases.
735 *) Add support for canonical generation of DSA parameter 'g'. See
738 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
739 POST to handle HMAC cases.
742 *) Add functions FIPS_module_version() and FIPS_module_version_text()
743 to return numerical and string versions of the FIPS module number.
746 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
747 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
748 outside the validated module in the FIPS capable OpenSSL.
751 *) Minor change to DRBG entropy callback semantics. In some cases
752 there is no multiple of the block length between min_len and
753 max_len. Allow the callback to return more than max_len bytes
754 of entropy but discard any extra: it is the callback's responsibility
755 to ensure that the extra data discarded does not impact the
756 requested amount of entropy.
759 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
760 information in FIPS186-3, SP800-57 and SP800-131A.
763 *) CCM support via EVP. Interface is very similar to GCM case except we
764 must supply all data in one chunk (i.e. no update, final) and the
765 message length must be supplied if AAD is used. Add algorithm test
769 *) Initial version of POST overhaul. Add POST callback to allow the status
770 of POST to be monitored and/or failures induced. Modify fips_test_suite
771 to use callback. Always run all selftests even if one fails.
774 *) XTS support including algorithm test driver in the fips_gcmtest program.
775 Note: this does increase the maximum key length from 32 to 64 bytes but
776 there should be no binary compatibility issues as existing applications
777 will never use XTS mode.
780 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
781 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
782 performs algorithm blocking for unapproved PRNG types. Also do not
783 set PRNG type in FIPS_mode_set(): leave this to the application.
784 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
785 the standard OpenSSL PRNG: set additional data to a date time vector.
788 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
789 This shouldn't present any incompatibility problems because applications
790 shouldn't be using these directly and any that are will need to rethink
791 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
794 *) Extensive self tests and health checking required by SP800-90 DRBG.
795 Remove strength parameter from FIPS_drbg_instantiate and always
796 instantiate at maximum supported strength.
799 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
802 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
805 *) New function DH_compute_key_padded() to compute a DH key and pad with
806 leading zeroes if needed: this complies with SP800-56A et al.
809 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
810 anything, incomplete, subject to change and largely untested at present.
813 *) Modify fipscanisteronly build option to only build the necessary object
814 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
817 *) Add experimental option FIPSSYMS to give all symbols in
818 fipscanister.o and FIPS or fips prefix. This will avoid
819 conflicts with future versions of OpenSSL. Add perl script
820 util/fipsas.pl to preprocess assembly language source files
821 and rename any affected symbols.
824 *) Add selftest checks and algorithm block of non-fips algorithms in
825 FIPS mode. Remove DES2 from selftests.
828 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
829 return internal method without any ENGINE dependencies. Add new
830 tiny fips sign and verify functions.
833 *) New build option no-ec2m to disable characteristic 2 code.
836 *) New build option "fipscanisteronly". This only builds fipscanister.o
837 and (currently) associated fips utilities. Uses the file Makefile.fips
838 instead of Makefile.org as the prototype.
841 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
842 Update fips_gcmtest to use IV generator.
845 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
846 setting output buffer to NULL. The *Final function must be
847 called although it will not retrieve any additional data. The tag
848 can be set or retrieved with a ctrl. The IV length is by default 12
849 bytes (96 bits) but can be set to an alternative value. If the IV
850 length exceeds the maximum IV length (currently 16 bytes) it cannot be
854 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
855 underlying do_cipher function handles all cipher semantics itself
856 including padding and finalisation. This is useful if (for example)
857 an ENGINE cipher handles block padding itself. The behaviour of
858 do_cipher is subtly changed if this flag is set: the return value
859 is the number of characters written to the output buffer (zero is
860 no longer an error code) or a negative error code. Also if the
861 input buffer is NULL and length 0 finalisation should be performed.
864 *) If a candidate issuer certificate is already part of the constructed
865 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
868 *) Improve forward-security support: add functions
870 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
871 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
873 for use by SSL/TLS servers; the callback function will be called whenever a
874 new session is created, and gets to decide whether the session may be
875 cached to make it resumable (return 0) or not (return 1). (As by the
876 SSL/TLS protocol specifications, the session_id sent by the server will be
877 empty to indicate that the session is not resumable; also, the server will
878 not generate RFC 4507 (RFC 5077) session tickets.)
880 A simple reasonable callback implementation is to return is_forward_secure.
881 This parameter will be set to 1 or 0 depending on the ciphersuite selected
882 by the SSL/TLS server library, indicating whether it can provide forward
884 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
886 *) New -verify_name option in command line utilities to set verification
890 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
891 Add CMAC pkey methods.
894 *) Experimental renegotiation in s_server -www mode. If the client
895 browses /reneg connection is renegotiated. If /renegcert it is
896 renegotiated requesting a certificate.
899 *) Add an "external" session cache for debugging purposes to s_server. This
900 should help trace issues which normally are only apparent in deployed
901 multi-process servers.
904 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
905 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
906 BIO_set_cipher() and some obscure PEM functions were changed so they
907 can now return an error. The RAND changes required a change to the
908 RAND_METHOD structure.
911 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
912 a gcc attribute to warn if the result of a function is ignored. This
913 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
914 whose return value is often ignored.
917 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
918 These allow SCTs (signed certificate timestamps) to be requested and
919 validated when establishing a connection.
920 [Rob Percival <robpercival@google.com>]
922 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
924 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
925 Builds that are not configured with "enable-weak-ssl-ciphers" will not
926 provide any "EXPORT" or "LOW" strength ciphers.
929 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
930 is by default disabled at build-time. Builds that are not configured with
931 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
932 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
933 will need to explicitly call either of:
935 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
937 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
939 as appropriate. Even if either of those is used, or the application
940 explicitly uses the version-specific SSLv2_method() or its client and
941 server variants, SSLv2 ciphers vulnerable to exhaustive search key
942 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
943 ciphers, and SSLv2 56-bit DES are no longer available.
947 *) Fix a double-free in DSA code
949 A double free bug was discovered when OpenSSL parses malformed DSA private
950 keys and could lead to a DoS attack or memory corruption for applications
951 that receive DSA private keys from untrusted sources. This scenario is
954 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
959 *) Disable SRP fake user seed to address a server memory leak.
961 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
963 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
964 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
965 was changed to ignore the "fake user" SRP seed, even if the seed
968 Users should use SRP_VBASE_get1_by_user instead. Note that in
969 SRP_VBASE_get1_by_user, caller must free the returned value. Note
970 also that even though configuring the SRP seed attempts to hide
971 invalid usernames by continuing the handshake with fake
972 credentials, this behaviour is not constant time and no strong
973 guarantees are made that the handshake is indistinguishable from
974 that of a valid user.
978 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
980 In the BN_hex2bn function the number of hex digits is calculated using an
981 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
982 large values of |i| this can result in |bn_expand| not allocating any
983 memory because |i * 4| is negative. This can leave the internal BIGNUM data
984 field as NULL leading to a subsequent NULL ptr deref. For very large values
985 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
986 In this case memory is allocated to the internal BIGNUM data field, but it
987 is insufficiently sized leading to heap corruption. A similar issue exists
988 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
989 is ever called by user applications with very large untrusted hex/dec data.
990 This is anticipated to be a rare occurrence.
992 All OpenSSL internal usage of these functions use data that is not expected
993 to be untrusted, e.g. config file data or application command line
994 arguments. If user developed applications generate config file data based
995 on untrusted data then it is possible that this could also lead to security
996 consequences. This is also anticipated to be rare.
998 This issue was reported to OpenSSL by Guido Vranken.
1002 *) Fix memory issues in BIO_*printf functions
1004 The internal |fmtstr| function used in processing a "%s" format string in
1005 the BIO_*printf functions could overflow while calculating the length of a
1006 string and cause an OOB read when printing very long strings.
1008 Additionally the internal |doapr_outch| function can attempt to write to an
1009 OOB memory location (at an offset from the NULL pointer) in the event of a
1010 memory allocation failure. In 1.0.2 and below this could be caused where
1011 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1012 could be in processing a very long "%s" format string. Memory leaks can
1015 The first issue may mask the second issue dependent on compiler behaviour.
1016 These problems could enable attacks where large amounts of untrusted data
1017 is passed to the BIO_*printf functions. If applications use these functions
1018 in this way then they could be vulnerable. OpenSSL itself uses these
1019 functions when printing out human-readable dumps of ASN.1 data. Therefore
1020 applications that print this data could be vulnerable if the data is from
1021 untrusted sources. OpenSSL command line applications could also be
1022 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1023 as command line arguments.
1025 Libssl is not considered directly vulnerable. Additionally certificates etc
1026 received via remote connections via libssl are also unlikely to be able to
1027 trigger these issues because of message size limits enforced within libssl.
1029 This issue was reported to OpenSSL Guido Vranken.
1033 *) Side channel attack on modular exponentiation
1035 A side-channel attack was found which makes use of cache-bank conflicts on
1036 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1037 of RSA keys. The ability to exploit this issue is limited as it relies on
1038 an attacker who has control of code in a thread running on the same
1039 hyper-threaded core as the victim thread which is performing decryptions.
1041 This issue was reported to OpenSSL by Yuval Yarom, The University of
1042 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1043 Nadia Heninger, University of Pennsylvania with more information at
1044 http://cachebleed.info.
1048 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1049 if no keysize is specified with default_bits. This fixes an
1050 omission in an earlier change that changed all RSA/DSA key generation
1051 apps to use 2048 bits by default.
1054 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1055 *) DH small subgroups
1057 Historically OpenSSL only ever generated DH parameters based on "safe"
1058 primes. More recently (in version 1.0.2) support was provided for
1059 generating X9.42 style parameter files such as those required for RFC 5114
1060 support. The primes used in such files may not be "safe". Where an
1061 application is using DH configured with parameters based on primes that are
1062 not "safe" then an attacker could use this fact to find a peer's private
1063 DH exponent. This attack requires that the attacker complete multiple
1064 handshakes in which the peer uses the same private DH exponent. For example
1065 this could be used to discover a TLS server's private DH exponent if it's
1066 reusing the private DH exponent or it's using a static DH ciphersuite.
1068 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1069 TLS. It is not on by default. If the option is not set then the server
1070 reuses the same private DH exponent for the life of the server process and
1071 would be vulnerable to this attack. It is believed that many popular
1072 applications do set this option and would therefore not be at risk.
1074 The fix for this issue adds an additional check where a "q" parameter is
1075 available (as is the case in X9.42 based parameters). This detects the
1076 only known attack, and is the only possible defense for static DH
1077 ciphersuites. This could have some performance impact.
1079 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1080 default and cannot be disabled. This could have some performance impact.
1082 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1086 *) SSLv2 doesn't block disabled ciphers
1088 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1089 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1090 been disabled, provided that the SSLv2 protocol was not also disabled via
1093 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1094 and Sebastian Schinzel.
1098 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1100 *) BN_mod_exp may produce incorrect results on x86_64
1102 There is a carry propagating bug in the x86_64 Montgomery squaring
1103 procedure. No EC algorithms are affected. Analysis suggests that attacks
1104 against RSA and DSA as a result of this defect would be very difficult to
1105 perform and are not believed likely. Attacks against DH are considered just
1106 feasible (although very difficult) because most of the work necessary to
1107 deduce information about a private key may be performed offline. The amount
1108 of resources required for such an attack would be very significant and
1109 likely only accessible to a limited number of attackers. An attacker would
1110 additionally need online access to an unpatched system using the target
1111 private key in a scenario with persistent DH parameters and a private
1112 key that is shared between multiple clients. For example this can occur by
1113 default in OpenSSL DHE based SSL/TLS ciphersuites.
1115 This issue was reported to OpenSSL by Hanno Böck.
1119 *) Certificate verify crash with missing PSS parameter
1121 The signature verification routines will crash with a NULL pointer
1122 dereference if presented with an ASN.1 signature using the RSA PSS
1123 algorithm and absent mask generation function parameter. Since these
1124 routines are used to verify certificate signature algorithms this can be
1125 used to crash any certificate verification operation and exploited in a
1126 DoS attack. Any application which performs certificate verification is
1127 vulnerable including OpenSSL clients and servers which enable client
1130 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1134 *) X509_ATTRIBUTE memory leak
1136 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1137 memory. This structure is used by the PKCS#7 and CMS routines so any
1138 application which reads PKCS#7 or CMS data from untrusted sources is
1139 affected. SSL/TLS is not affected.
1141 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1146 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1147 This changes the decoding behaviour for some invalid messages,
1148 though the change is mostly in the more lenient direction, and
1149 legacy behaviour is preserved as much as possible.
1152 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1154 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1156 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1158 *) Alternate chains certificate forgery
1160 During certificate verfification, OpenSSL will attempt to find an
1161 alternative certificate chain if the first attempt to build such a chain
1162 fails. An error in the implementation of this logic can mean that an
1163 attacker could cause certain checks on untrusted certificates to be
1164 bypassed, such as the CA flag, enabling them to use a valid leaf
1165 certificate to act as a CA and "issue" an invalid certificate.
1167 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1171 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1173 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1174 incompatibility in the handling of HMAC. The previous ABI has now been
1178 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1180 *) Malformed ECParameters causes infinite loop
1182 When processing an ECParameters structure OpenSSL enters an infinite loop
1183 if the curve specified is over a specially malformed binary polynomial
1186 This can be used to perform denial of service against any
1187 system which processes public keys, certificate requests or
1188 certificates. This includes TLS clients and TLS servers with
1189 client authentication enabled.
1191 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1195 *) Exploitable out-of-bounds read in X509_cmp_time
1197 X509_cmp_time does not properly check the length of the ASN1_TIME
1198 string and can read a few bytes out of bounds. In addition,
1199 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1202 An attacker can use this to craft malformed certificates and CRLs of
1203 various sizes and potentially cause a segmentation fault, resulting in
1204 a DoS on applications that verify certificates or CRLs. TLS clients
1205 that verify CRLs are affected. TLS clients and servers with client
1206 authentication enabled may be affected if they use custom verification
1209 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1210 independently by Hanno Böck.
1214 *) PKCS7 crash with missing EnvelopedContent
1216 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1217 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1218 with missing content and trigger a NULL pointer dereference on parsing.
1220 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1221 structures from untrusted sources are affected. OpenSSL clients and
1222 servers are not affected.
1224 This issue was reported to OpenSSL by Michal Zalewski (Google).
1228 *) CMS verify infinite loop with unknown hash function
1230 When verifying a signedData message the CMS code can enter an infinite loop
1231 if presented with an unknown hash function OID. This can be used to perform
1232 denial of service against any system which verifies signedData messages using
1234 This issue was reported to OpenSSL by Johannes Bauer.
1238 *) Race condition handling NewSessionTicket
1240 If a NewSessionTicket is received by a multi-threaded client when attempting to
1241 reuse a previous ticket then a race condition can occur potentially leading to
1242 a double free of the ticket data.
1246 *) Only support 256-bit or stronger elliptic curves with the
1247 'ecdh_auto' setting (server) or by default (client). Of supported
1248 curves, prefer P-256 (both).
1251 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1253 *) ClientHello sigalgs DoS fix
1255 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1256 invalid signature algorithms extension a NULL pointer dereference will
1257 occur. This can be exploited in a DoS attack against the server.
1259 This issue was was reported to OpenSSL by David Ramos of Stanford
1262 [Stephen Henson and Matt Caswell]
1264 *) Multiblock corrupted pointer fix
1266 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1267 feature only applies on 64 bit x86 architecture platforms that support AES
1268 NI instructions. A defect in the implementation of "multiblock" can cause
1269 OpenSSL's internal write buffer to become incorrectly set to NULL when
1270 using non-blocking IO. Typically, when the user application is using a
1271 socket BIO for writing, this will only result in a failed connection.
1272 However if some other BIO is used then it is likely that a segmentation
1273 fault will be triggered, thus enabling a potential DoS attack.
1275 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1279 *) Segmentation fault in DTLSv1_listen fix
1281 The DTLSv1_listen function is intended to be stateless and processes the
1282 initial ClientHello from many peers. It is common for user code to loop
1283 over the call to DTLSv1_listen until a valid ClientHello is received with
1284 an associated cookie. A defect in the implementation of DTLSv1_listen means
1285 that state is preserved in the SSL object from one invocation to the next
1286 that can lead to a segmentation fault. Errors processing the initial
1287 ClientHello can trigger this scenario. An example of such an error could be
1288 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1291 This issue was reported to OpenSSL by Per Allansson.
1295 *) Segmentation fault in ASN1_TYPE_cmp fix
1297 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1298 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1299 certificate signature algorithm consistency this can be used to crash any
1300 certificate verification operation and exploited in a DoS attack. Any
1301 application which performs certificate verification is vulnerable including
1302 OpenSSL clients and servers which enable client authentication.
1306 *) Segmentation fault for invalid PSS parameters fix
1308 The signature verification routines will crash with a NULL pointer
1309 dereference if presented with an ASN.1 signature using the RSA PSS
1310 algorithm and invalid parameters. Since these routines are used to verify
1311 certificate signature algorithms this can be used to crash any
1312 certificate verification operation and exploited in a DoS attack. Any
1313 application which performs certificate verification is vulnerable including
1314 OpenSSL clients and servers which enable client authentication.
1316 This issue was was reported to OpenSSL by Brian Carpenter.
1320 *) ASN.1 structure reuse memory corruption fix
1322 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1323 memory corruption via an invalid write. Such reuse is and has been
1324 strongly discouraged and is believed to be rare.
1326 Applications that parse structures containing CHOICE or ANY DEFINED BY
1327 components may be affected. Certificate parsing (d2i_X509 and related
1328 functions) are however not affected. OpenSSL clients and servers are
1333 *) PKCS7 NULL pointer dereferences fix
1335 The PKCS#7 parsing code does not handle missing outer ContentInfo
1336 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1337 missing content and trigger a NULL pointer dereference on parsing.
1339 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1340 otherwise parse PKCS#7 structures from untrusted sources are
1341 affected. OpenSSL clients and servers are not affected.
1343 This issue was reported to OpenSSL by Michal Zalewski (Google).
1347 *) DoS via reachable assert in SSLv2 servers fix
1349 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1350 servers that both support SSLv2 and enable export cipher suites by sending
1351 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1353 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1354 (OpenSSL development team).
1358 *) Empty CKE with client auth and DHE fix
1360 If client auth is used then a server can seg fault in the event of a DHE
1361 ciphersuite being selected and a zero length ClientKeyExchange message
1362 being sent by the client. This could be exploited in a DoS attack.
1366 *) Handshake with unseeded PRNG fix
1368 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1369 with an unseeded PRNG. The conditions are:
1370 - The client is on a platform where the PRNG has not been seeded
1371 automatically, and the user has not seeded manually
1372 - A protocol specific client method version has been used (i.e. not
1373 SSL_client_methodv23)
1374 - A ciphersuite is used that does not require additional random data from
1375 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1377 If the handshake succeeds then the client random that has been used will
1378 have been generated from a PRNG with insufficient entropy and therefore the
1379 output may be predictable.
1381 For example using the following command with an unseeded openssl will
1382 succeed on an unpatched platform:
1384 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1388 *) Use After Free following d2i_ECPrivatekey error fix
1390 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1391 could cause a use after free condition. This, in turn, could cause a double
1392 free in several private key parsing functions (such as d2i_PrivateKey
1393 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1394 for applications that receive EC private keys from untrusted
1395 sources. This scenario is considered rare.
1397 This issue was discovered by the BoringSSL project and fixed in their
1402 *) X509_to_X509_REQ NULL pointer deref fix
1404 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1405 the certificate key is invalid. This function is rarely used in practice.
1407 This issue was discovered by Brian Carpenter.
1411 *) Removed the export ciphers from the DEFAULT ciphers
1414 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1416 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1417 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1418 So far those who have to target multiple plaforms would compromise
1419 and argue that binary targeting say ARMv5 would still execute on
1420 ARMv8. "Universal" build resolves this compromise by providing
1421 near-optimal performance even on newer platforms.
1424 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1425 (other platforms pending).
1426 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1428 *) Add support for the SignedCertificateTimestampList certificate and
1429 OCSP response extensions from RFC6962.
1432 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1433 for corner cases. (Certain input points at infinity could lead to
1434 bogus results, with non-infinity inputs mapped to infinity too.)
1437 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1438 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1439 common cases are optimized and there still is room for further
1440 improvements. Vector Permutation AES for Altivec is also added.
1443 *) Add support for little-endian ppc64 Linux target.
1444 [Marcelo Cerri (IBM)]
1446 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1447 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1448 are optimized and there still is room for further improvements.
1449 Both 32- and 64-bit modes are supported.
1450 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1452 *) Improved ARMv7 NEON support.
1455 *) Support for SPARC Architecture 2011 crypto extensions, first
1456 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1457 SHA256/512, MD5, GHASH and modular exponentiation.
1458 [Andy Polyakov, David Miller]
1460 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1462 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1464 *) Support for new and upcoming Intel processors, including AVX2,
1465 BMI and SHA ISA extensions. This includes additional "stitched"
1466 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1469 This work was sponsored by Intel Corp.
1472 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1473 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1474 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1477 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1478 this fixes a limiation in previous versions of OpenSSL.
1481 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1482 MGF1 digest and OAEP label.
1485 *) Add EVP support for key wrapping algorithms, to avoid problems with
1486 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1487 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1488 algorithms and include tests cases.
1491 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1493 [Douglas E. Engert, Steve Henson]
1495 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1496 difference in days and seconds between two tm or ASN1_TIME structures.
1499 *) Add -rev test option to s_server to just reverse order of characters
1500 received by client and send back to server. Also prints an abbreviated
1501 summary of the connection parameters.
1504 *) New option -brief for s_client and s_server to print out a brief summary
1505 of connection parameters.
1508 *) Add callbacks for arbitrary TLS extensions.
1509 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1511 *) New option -crl_download in several openssl utilities to download CRLs
1512 from CRLDP extension in certificates.
1515 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1518 *) New function X509_CRL_diff to generate a delta CRL from the difference
1519 of two full CRLs. Add support to "crl" utility.
1522 *) New functions to set lookup_crls function and to retrieve
1523 X509_STORE from X509_STORE_CTX.
1526 *) Print out deprecated issuer and subject unique ID fields in
1530 *) Extend OCSP I/O functions so they can be used for simple general purpose
1531 HTTP as well as OCSP. New wrapper function which can be used to download
1532 CRLs using the OCSP API.
1535 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1538 *) SSL_CONF* functions. These provide a common framework for application
1539 configuration using configuration files or command lines.
1542 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1543 message callback and prints the results. Needs compile time option
1544 "enable-ssl-trace". New options to s_client and s_server to enable
1548 *) New ctrl and macro to retrieve supported points extensions.
1549 Print out extension in s_server and s_client.
1552 *) New functions to retrieve certificate signature and signature
1556 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1560 *) New Suite B modes for TLS code. These use and enforce the requirements
1561 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1562 only use Suite B curves. The Suite B modes can be set by using the
1563 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1566 *) New chain verification flags for Suite B levels of security. Check
1567 algorithms are acceptable when flags are set in X509_verify_cert.
1570 *) Make tls1_check_chain return a set of flags indicating checks passed
1571 by a certificate chain. Add additional tests to handle client
1572 certificates: checks for matching certificate type and issuer name
1576 *) If an attempt is made to use a signature algorithm not in the peer
1577 preference list abort the handshake. If client has no suitable
1578 signature algorithms in response to a certificate request do not
1579 use the certificate.
1582 *) If server EC tmp key is not in client preference list abort handshake.
1585 *) Add support for certificate stores in CERT structure. This makes it
1586 possible to have different stores per SSL structure or one store in
1587 the parent SSL_CTX. Include distint stores for certificate chain
1588 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1589 to build and store a certificate chain in CERT structure: returing
1590 an error if the chain cannot be built: this will allow applications
1591 to test if a chain is correctly configured.
1593 Note: if the CERT based stores are not set then the parent SSL_CTX
1594 store is used to retain compatibility with existing behaviour.
1598 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1599 mask based on the current session, check mask when sending client
1600 hello and checking the requested ciphersuite.
1603 *) New ctrls to retrieve and set certificate types in a certificate
1604 request message. Print out received values in s_client. If certificate
1605 types is not set with custom values set sensible values based on
1606 supported signature algorithms.
1609 *) Support for distinct client and server supported signature algorithms.
1612 *) Add certificate callback. If set this is called whenever a certificate
1613 is required by client or server. An application can decide which
1614 certificate chain to present based on arbitrary criteria: for example
1615 supported signature algorithms. Add very simple example to s_server.
1616 This fixes many of the problems and restrictions of the existing client
1617 certificate callback: for example you can now clear an existing
1618 certificate and specify the whole chain.
1621 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1622 the certificate can be used for (if anything). Set valid_flags field
1623 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1624 to have similar checks in it.
1626 Add new "cert_flags" field to CERT structure and include a "strict mode".
1627 This enforces some TLS certificate requirements (such as only permitting
1628 certificate signature algorithms contained in the supported algorithms
1629 extension) which some implementations ignore: this option should be used
1630 with caution as it could cause interoperability issues.
1633 *) Update and tidy signature algorithm extension processing. Work out
1634 shared signature algorithms based on preferences and peer algorithms
1635 and print them out in s_client and s_server. Abort handshake if no
1636 shared signature algorithms.
1639 *) Add new functions to allow customised supported signature algorithms
1640 for SSL and SSL_CTX structures. Add options to s_client and s_server
1644 *) New function SSL_certs_clear() to delete all references to certificates
1645 from an SSL structure. Before this once a certificate had been added
1646 it couldn't be removed.
1649 *) Integrate hostname, email address and IP address checking with certificate
1650 verification. New verify options supporting checking in opensl utility.
1653 *) Fixes and wildcard matching support to hostname and email checking
1654 functions. Add manual page.
1655 [Florian Weimer (Red Hat Product Security Team)]
1657 *) New functions to check a hostname email or IP address against a
1658 certificate. Add options x509 utility to print results of checks against
1662 *) Fix OCSP checking.
1663 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1665 *) Initial experimental support for explicitly trusted non-root CAs.
1666 OpenSSL still tries to build a complete chain to a root but if an
1667 intermediate CA has a trust setting included that is used. The first
1668 setting is used: whether to trust (e.g., -addtrust option to the x509
1672 *) Add -trusted_first option which attempts to find certificates in the
1673 trusted store even if an untrusted chain is also supplied.
1676 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1677 platform support for Linux and Android.
1680 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1683 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1684 When in FIPS mode the approved implementations are used as normal,
1685 when not in FIPS mode the internal unapproved versions are used instead.
1686 This means that the FIPS capable OpenSSL isn't forced to use the
1687 (often lower performance) FIPS implementations outside FIPS mode.
1690 *) Transparently support X9.42 DH parameters when calling
1691 PEM_read_bio_DHparameters. This means existing applications can handle
1692 the new parameter format automatically.
1695 *) Initial experimental support for X9.42 DH parameter format: mainly
1696 to support use of 'q' parameter for RFC5114 parameters.
1699 *) Add DH parameters from RFC5114 including test data to dhtest.
1702 *) Support for automatic EC temporary key parameter selection. If enabled
1703 the most preferred EC parameters are automatically used instead of
1704 hardcoded fixed parameters. Now a server just has to call:
1705 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1706 support ECDH and use the most appropriate parameters.
1709 *) Enhance and tidy EC curve and point format TLS extension code. Use
1710 static structures instead of allocation if default values are used.
1711 New ctrls to set curves we wish to support and to retrieve shared curves.
1712 Print out shared curves in s_server. New options to s_server and s_client
1713 to set list of supported curves.
1716 *) New ctrls to retrieve supported signature algorithms and
1717 supported curve values as an array of NIDs. Extend openssl utility
1718 to print out received values.
1721 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1722 between NIDs and the more common NIST names such as "P-256". Enhance
1723 ecparam utility and ECC method to recognise the NIST names for curves.
1726 *) Enhance SSL/TLS certificate chain handling to support different
1727 chains for each certificate instead of one chain in the parent SSL_CTX.
1730 *) Support for fixed DH ciphersuite client authentication: where both
1731 server and client use DH certificates with common parameters.
1734 *) Support for fixed DH ciphersuites: those requiring DH server
1738 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1740 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1741 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1742 X509_CINF_get_signature were reverted post internal team review.
1744 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1746 *) Build fixes for the Windows and OpenVMS platforms
1747 [Matt Caswell and Richard Levitte]
1749 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1751 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1752 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1753 dereference. This could lead to a Denial Of Service attack. Thanks to
1754 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1758 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1759 dtls1_buffer_record function under certain conditions. In particular this
1760 could occur if an attacker sent repeated DTLS records with the same
1761 sequence number but for the next epoch. The memory leak could be exploited
1762 by an attacker in a Denial of Service attack through memory exhaustion.
1763 Thanks to Chris Mueller for reporting this issue.
1767 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1768 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1769 method would be set to NULL which could later result in a NULL pointer
1770 dereference. Thanks to Frank Schmirler for reporting this issue.
1774 *) Abort handshake if server key exchange message is omitted for ephemeral
1777 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1778 reporting this issue.
1782 *) Remove non-export ephemeral RSA code on client and server. This code
1783 violated the TLS standard by allowing the use of temporary RSA keys in
1784 non-export ciphersuites and could be used by a server to effectively
1785 downgrade the RSA key length used to a value smaller than the server
1786 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1787 INRIA or reporting this issue.
1791 *) Fixed issue where DH client certificates are accepted without verification.
1792 An OpenSSL server will accept a DH certificate for client authentication
1793 without the certificate verify message. This effectively allows a client to
1794 authenticate without the use of a private key. This only affects servers
1795 which trust a client certificate authority which issues certificates
1796 containing DH keys: these are extremely rare and hardly ever encountered.
1797 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1802 *) Ensure that the session ID context of an SSL is updated when its
1803 SSL_CTX is updated via SSL_set_SSL_CTX.
1805 The session ID context is typically set from the parent SSL_CTX,
1806 and can vary with the CTX.
1809 *) Fix various certificate fingerprint issues.
1811 By using non-DER or invalid encodings outside the signed portion of a
1812 certificate the fingerprint can be changed without breaking the signature.
1813 Although no details of the signed portion of the certificate can be changed
1814 this can cause problems with some applications: e.g. those using the
1815 certificate fingerprint for blacklists.
1817 1. Reject signatures with non zero unused bits.
1819 If the BIT STRING containing the signature has non zero unused bits reject
1820 the signature. All current signature algorithms require zero unused bits.
1822 2. Check certificate algorithm consistency.
1824 Check the AlgorithmIdentifier inside TBS matches the one in the
1825 certificate signature. NB: this will result in signature failure
1826 errors for some broken certificates.
1828 Thanks to Konrad Kraszewski from Google for reporting this issue.
1830 3. Check DSA/ECDSA signatures use DER.
1832 Reencode DSA/ECDSA signatures and compare with the original received
1833 signature. Return an error if there is a mismatch.
1835 This will reject various cases including garbage after signature
1836 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1837 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1838 (negative or with leading zeroes).
1840 Further analysis was conducted and fixes were developed by Stephen Henson
1841 of the OpenSSL core team.
1846 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1847 results on some platforms, including x86_64. This bug occurs at random
1848 with a very low probability, and is not known to be exploitable in any
1849 way, though its exact impact is difficult to determine. Thanks to Pieter
1850 Wuille (Blockstream) who reported this issue and also suggested an initial
1851 fix. Further analysis was conducted by the OpenSSL development team and
1852 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1853 the OpenSSL core team.
1857 *) Do not resume sessions on the server if the negotiated protocol
1858 version does not match the session's version. Resuming with a different
1859 version, while not strictly forbidden by the RFC, is of questionable
1860 sanity and breaks all known clients.
1861 [David Benjamin, Emilia Käsper]
1863 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1864 early CCS messages during renegotiation. (Note that because
1865 renegotiation is encrypted, this early CCS was not exploitable.)
1868 *) Tighten client-side session ticket handling during renegotiation:
1869 ensure that the client only accepts a session ticket if the server sends
1870 the extension anew in the ServerHello. Previously, a TLS client would
1871 reuse the old extension state and thus accept a session ticket if one was
1872 announced in the initial ServerHello.
1874 Similarly, ensure that the client requires a session ticket if one
1875 was advertised in the ServerHello. Previously, a TLS client would
1876 ignore a missing NewSessionTicket message.
1879 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1881 *) SRTP Memory Leak.
1883 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1884 sends a carefully crafted handshake message, to cause OpenSSL to fail
1885 to free up to 64k of memory causing a memory leak. This could be
1886 exploited in a Denial Of Service attack. This issue affects OpenSSL
1887 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1888 whether SRTP is used or configured. Implementations of OpenSSL that
1889 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1891 The fix was developed by the OpenSSL team.
1895 *) Session Ticket Memory Leak.
1897 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1898 integrity of that ticket is first verified. In the event of a session
1899 ticket integrity check failing, OpenSSL will fail to free memory
1900 causing a memory leak. By sending a large number of invalid session
1901 tickets an attacker could exploit this issue in a Denial Of Service
1906 *) Build option no-ssl3 is incomplete.
1908 When OpenSSL is configured with "no-ssl3" as a build option, servers
1909 could accept and complete a SSL 3.0 handshake, and clients could be
1910 configured to send them.
1912 [Akamai and the OpenSSL team]
1914 *) Add support for TLS_FALLBACK_SCSV.
1915 Client applications doing fallback retries should call
1916 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1918 [Adam Langley, Bodo Moeller]
1920 *) Add additional DigestInfo checks.
1922 Reencode DigestInto in DER and check against the original when
1923 verifying RSA signature: this will reject any improperly encoded
1924 DigestInfo structures.
1926 Note: this is a precautionary measure and no attacks are currently known.
1930 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1932 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1933 SRP code can be overrun an internal buffer. Add sanity check that
1934 g, A, B < N to SRP code.
1936 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1937 Group for discovering this issue.
1941 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1942 TLS 1.0 instead of higher protocol versions when the ClientHello message
1943 is badly fragmented. This allows a man-in-the-middle attacker to force a
1944 downgrade to TLS 1.0 even if both the server and the client support a
1945 higher protocol version, by modifying the client's TLS records.
1947 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1948 researching this issue.
1952 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1953 to a denial of service attack. A malicious server can crash the client
1954 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1955 ciphersuite and sending carefully crafted handshake messages.
1957 Thanks to Felix Gröbert (Google) for discovering and researching this
1962 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1963 to leak memory. This can be exploited through a Denial of Service attack.
1964 Thanks to Adam Langley for discovering and researching this issue.
1968 *) An attacker can force openssl to consume large amounts of memory whilst
1969 processing DTLS handshake messages. This can be exploited through a
1970 Denial of Service attack.
1971 Thanks to Adam Langley for discovering and researching this issue.
1975 *) An attacker can force an error condition which causes openssl to crash
1976 whilst processing DTLS packets due to memory being freed twice. This
1977 can be exploited through a Denial of Service attack.
1978 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1983 *) If a multithreaded client connects to a malicious server using a resumed
1984 session and the server sends an ec point format extension it could write
1985 up to 255 bytes to freed memory.
1987 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1992 *) A malicious server can crash an OpenSSL client with a null pointer
1993 dereference (read) by specifying an SRP ciphersuite even though it was not
1994 properly negotiated with the client. This can be exploited through a
1995 Denial of Service attack.
1997 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1998 discovering and researching this issue.
2002 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2003 X509_name_oneline, X509_name_print_ex et al. to leak some information
2004 from the stack. Applications may be affected if they echo pretty printing
2005 output to the attacker.
2007 Thanks to Ivan Fratric (Google) for discovering this issue.
2009 [Emilia Käsper, and Steve Henson]
2011 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2012 for corner cases. (Certain input points at infinity could lead to
2013 bogus results, with non-infinity inputs mapped to infinity too.)
2016 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2018 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2019 handshake can force the use of weak keying material in OpenSSL
2020 SSL/TLS clients and servers.
2022 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2023 researching this issue. (CVE-2014-0224)
2024 [KIKUCHI Masashi, Steve Henson]
2026 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2027 OpenSSL DTLS client the code can be made to recurse eventually crashing
2030 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2032 [Imre Rad, Steve Henson]
2034 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2035 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2036 client or server. This is potentially exploitable to run arbitrary
2037 code on a vulnerable client or server.
2039 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2040 [Jüri Aedla, Steve Henson]
2042 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2043 are subject to a denial of service attack.
2045 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2046 this issue. (CVE-2014-3470)
2047 [Felix Gröbert, Ivan Fratric, Steve Henson]
2049 *) Harmonize version and its documentation. -f flag is used to display
2051 [mancha <mancha1@zoho.com>]
2053 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2054 in i2d_ECPrivateKey.
2055 [mancha <mancha1@zoho.com>]
2057 *) Fix some double frees. These are not thought to be exploitable.
2058 [mancha <mancha1@zoho.com>]
2060 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2062 *) A missing bounds check in the handling of the TLS heartbeat extension
2063 can be used to reveal up to 64k of memory to a connected client or
2066 Thanks for Neel Mehta of Google Security for discovering this bug and to
2067 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2068 preparing the fix (CVE-2014-0160)
2069 [Adam Langley, Bodo Moeller]
2071 *) Fix for the attack described in the paper "Recovering OpenSSL
2072 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2073 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2074 http://eprint.iacr.org/2014/140
2076 Thanks to Yuval Yarom and Naomi Benger for discovering this
2077 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2078 [Yuval Yarom and Naomi Benger]
2080 *) TLS pad extension: draft-agl-tls-padding-03
2082 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2083 TLS client Hello record length value would otherwise be > 255 and
2084 less that 512 pad with a dummy extension containing zeroes so it
2085 is at least 512 bytes long.
2087 [Adam Langley, Steve Henson]
2089 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2091 *) Fix for TLS record tampering bug. A carefully crafted invalid
2092 handshake could crash OpenSSL with a NULL pointer exception.
2093 Thanks to Anton Johansson for reporting this issues.
2096 *) Keep original DTLS digest and encryption contexts in retransmission
2097 structures so we can use the previous session parameters if they need
2098 to be resent. (CVE-2013-6450)
2101 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2102 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2103 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2104 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2105 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2106 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2107 [Rob Stradling, Adam Langley]
2109 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2111 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2112 supporting platforms or when small records were transferred.
2113 [Andy Polyakov, Steve Henson]
2115 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2117 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2119 This addresses the flaw in CBC record processing discovered by
2120 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2121 at: http://www.isg.rhul.ac.uk/tls/
2123 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2124 Security Group at Royal Holloway, University of London
2125 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2126 Emilia Käsper for the initial patch.
2128 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2130 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2131 ciphersuites which can be exploited in a denial of service attack.
2132 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2133 and detecting this bug and to Wolfgang Ettlinger
2134 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2138 *) Return an error when checking OCSP signatures when key is NULL.
2139 This fixes a DoS attack. (CVE-2013-0166)
2142 *) Make openssl verify return errors.
2143 [Chris Palmer <palmer@google.com> and Ben Laurie]
2145 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2146 the right response is stapled. Also change SSL_get_certificate()
2147 so it returns the certificate actually sent.
2148 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2149 [Rob Stradling <rob.stradling@comodo.com>]
2151 *) Fix possible deadlock when decoding public keys.
2154 *) Don't use TLS 1.0 record version number in initial client hello
2158 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2160 *) Sanity check record length before skipping explicit IV in TLS
2161 1.2, 1.1 and DTLS to fix DoS attack.
2163 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2164 fuzzing as a service testing platform.
2168 *) Initialise tkeylen properly when encrypting CMS messages.
2169 Thanks to Solar Designer of Openwall for reporting this issue.
2172 *) In FIPS mode don't try to use composite ciphers as they are not
2176 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2178 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2179 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2180 mean any application compiled against OpenSSL 1.0.0 headers setting
2181 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2182 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2183 0x10000000L Any application which was previously compiled against
2184 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2185 will need to be recompiled as a result. Letting be results in
2186 inability to disable specifically TLS 1.1 and in client context,
2187 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2190 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2191 disable just protocol X, but all protocols above X *if* there are
2192 protocols *below* X still enabled. In more practical terms it means
2193 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2194 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2195 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2199 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2201 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2202 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2203 in CRYPTO_realloc_clean.
2205 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2206 issue and to Adam Langley <agl@chromium.org> for fixing it.
2208 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2210 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2213 *) Workarounds for some broken servers that "hang" if a client hello
2214 record length exceeds 255 bytes.
2216 1. Do not use record version number > TLS 1.0 in initial client
2217 hello: some (but not all) hanging servers will now work.
2218 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2219 the number of ciphers sent in the client hello. This should be
2220 set to an even number, such as 50, for example by passing:
2221 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2222 Most broken servers should now work.
2223 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2224 TLS 1.2 client support entirely.
2227 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2230 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2232 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2233 STRING form instead of a DigestInfo.
2236 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2237 and the RSA_sign/RSA_verify functions. This was made more apparent when
2238 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2239 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2240 the correct format in RSA_verify so both forms transparently work.
2243 *) Some servers which support TLS 1.0 can choke if we initially indicate
2244 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2245 encrypted premaster secret. As a workaround use the maximum permitted
2246 client version in client hello, this should keep such servers happy
2247 and still work with previous versions of OpenSSL.
2250 *) Add support for TLS/DTLS heartbeats.
2251 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2253 *) Add support for SCTP.
2254 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2256 *) Improved PRNG seeding for VOS.
2257 [Paul Green <Paul.Green@stratus.com>]
2259 *) Extensive assembler packs updates, most notably:
2261 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2262 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2263 - x86_64: bit-sliced AES implementation;
2264 - ARM: NEON support, contemporary platforms optimizations;
2265 - s390x: z196 support;
2266 - *: GHASH and GF(2^m) multiplication implementations;
2270 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2271 (removal of unnecessary code)
2272 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2274 *) Add TLS key material exporter from RFC 5705.
2277 *) Add DTLS-SRTP negotiation from RFC 5764.
2280 *) Add Next Protocol Negotiation,
2281 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2282 disabled with a no-npn flag to config or Configure. Code donated
2284 [Adam Langley <agl@google.com> and Ben Laurie]
2286 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2287 NIST-P256, NIST-P521, with constant-time single point multiplication on
2288 typical inputs. Compiler support for the nonstandard type __uint128_t is
2289 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2290 Code made available under Apache License version 2.0.
2292 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2293 line to include this in your build of OpenSSL, and run "make depend" (or
2294 "make update"). This enables the following EC_METHODs:
2296 EC_GFp_nistp224_method()
2297 EC_GFp_nistp256_method()
2298 EC_GFp_nistp521_method()
2300 EC_GROUP_new_by_curve_name() will automatically use these (while
2301 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2303 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2305 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2306 all platforms. Move ssize_t definition from e_os.h to the public
2307 header file e_os2.h as it now appears in public header file cms.h
2310 *) New -sigopt option to the ca, req and x509 utilities. Additional
2311 signature parameters can be passed using this option and in
2315 *) Add RSA PSS signing function. This will generate and set the
2316 appropriate AlgorithmIdentifiers for PSS based on those in the
2317 corresponding EVP_MD_CTX structure. No application support yet.
2320 *) Support for companion algorithm specific ASN1 signing routines.
2321 New function ASN1_item_sign_ctx() signs a pre-initialised
2322 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2323 the appropriate parameters.
2326 *) Add new algorithm specific ASN1 verification initialisation function
2327 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2328 handling will be the same no matter what EVP_PKEY_METHOD is used.
2329 Add a PSS handler to support verification of PSS signatures: checked
2330 against a number of sample certificates.
2333 *) Add signature printing for PSS. Add PSS OIDs.
2334 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2336 *) Add algorithm specific signature printing. An individual ASN1 method
2337 can now print out signatures instead of the standard hex dump.
2339 More complex signatures (e.g. PSS) can print out more meaningful
2340 information. Include DSA version that prints out the signature
2344 *) Password based recipient info support for CMS library: implementing
2348 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2349 neatly separates the code into cipher and PBE sections and is required
2350 for some algorithms that split PBES2 into separate pieces (such as
2351 password based CMS).
2354 *) Session-handling fixes:
2355 - Fix handling of connections that are resuming with a session ID,
2356 but also support Session Tickets.
2357 - Fix a bug that suppressed issuing of a new ticket if the client
2358 presented a ticket with an expired session.
2359 - Try to set the ticket lifetime hint to something reasonable.
2360 - Make tickets shorter by excluding irrelevant information.
2361 - On the client side, don't ignore renewed tickets.
2362 [Adam Langley, Bodo Moeller (Google)]
2364 *) Fix PSK session representation.
2367 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2369 This work was sponsored by Intel.
2372 *) Add GCM support to TLS library. Some custom code is needed to split
2373 the IV between the fixed (from PRF) and explicit (from TLS record)
2374 portions. This adds all GCM ciphersuites supported by RFC5288 and
2375 RFC5289. Generalise some AES* cipherstrings to include GCM and
2376 add a special AESGCM string for GCM only.
2379 *) Expand range of ctrls for AES GCM. Permit setting invocation
2380 field on decrypt and retrieval of invocation field only on encrypt.
2383 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2384 As required by RFC5289 these ciphersuites cannot be used if for
2385 versions of TLS earlier than 1.2.
2388 *) For FIPS capable OpenSSL interpret a NULL default public key method
2389 as unset and return the appropriate default but do *not* set the default.
2390 This means we can return the appropriate method in applications that
2391 switch between FIPS and non-FIPS modes.
2394 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2395 ENGINE is used then we cannot handle that in the FIPS module so we
2396 keep original code iff non-FIPS operations are allowed.
2399 *) Add -attime option to openssl utilities.
2400 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2402 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2405 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2406 FIPS EC methods unconditionally for now.
2409 *) New build option no-ec2m to disable characteristic 2 code.
2412 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2413 all cases can be covered as some introduce binary incompatibilities.
2416 *) Redirect RSA operations to FIPS module including keygen,
2417 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2420 *) Add similar low level API blocking to ciphers.
2423 *) Low level digest APIs are not approved in FIPS mode: any attempt
2424 to use these will cause a fatal error. Applications that *really* want
2425 to use them can use the private_* version instead.
2428 *) Redirect cipher operations to FIPS module for FIPS builds.
2431 *) Redirect digest operations to FIPS module for FIPS builds.
2434 *) Update build system to add "fips" flag which will link in fipscanister.o
2435 for static and shared library builds embedding a signature if needed.
2438 *) Output TLS supported curves in preference order instead of numerical
2439 order. This is currently hardcoded for the highest order curves first.
2440 This should be configurable so applications can judge speed vs strength.
2443 *) Add TLS v1.2 server support for client authentication.
2446 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2450 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2451 FIPS modules versions.
2454 *) Add TLS v1.2 client side support for client authentication. Keep cache
2455 of handshake records longer as we don't know the hash algorithm to use
2456 until after the certificate request message is received.
2459 *) Initial TLS v1.2 client support. Add a default signature algorithms
2460 extension including all the algorithms we support. Parse new signature
2461 format in client key exchange. Relax some ECC signing restrictions for
2462 TLS v1.2 as indicated in RFC5246.
2465 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2466 to new signature format when needed using client digest preference.
2467 All server ciphersuites should now work correctly in TLS v1.2. No client
2468 support yet and no support for client certificates.
2471 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2472 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2473 ciphersuites. At present only RSA key exchange ciphersuites work with
2474 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2475 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2476 and version checking.
2479 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2480 with this defined it will not be affected by any changes to ssl internal
2481 structures. Add several utility functions to allow openssl application
2482 to work with OPENSSL_NO_SSL_INTERN defined.
2486 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2488 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2491 *) Permit abbreviated handshakes when renegotiating using the function
2492 SSL_renegotiate_abbreviated().
2493 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2495 *) Add call to ENGINE_register_all_complete() to
2496 ENGINE_load_builtin_engines(), so some implementations get used
2497 automatically instead of needing explicit application support.
2500 *) Add support for TLS key exporter as described in RFC5705.
2501 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2503 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2504 a few changes are required:
2506 Add SSL_OP_NO_TLSv1_1 flag.
2507 Add TLSv1_1 methods.
2508 Update version checking logic to handle version 1.1.
2509 Add explicit IV handling (ported from DTLS code).
2510 Add command line options to s_client/s_server.
2513 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2515 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2516 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2517 content decryption and always return the same error. Note: this attack
2518 needs on average 2^20 messages so it only affects automated senders. The
2519 old behaviour can be reenabled in the CMS code by setting the
2520 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2521 an MMA defence is not necessary.
2522 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2523 this issue. (CVE-2012-0884)
2526 *) Fix CVE-2011-4619: make sure we really are receiving a
2527 client hello before rejecting multiple SGC restarts. Thanks to
2528 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2531 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2533 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2534 Thanks to Antonio Martin, Enterprise Secure Access Research and
2535 Development, Cisco Systems, Inc. for discovering this bug and
2536 preparing a fix. (CVE-2012-0050)
2539 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2541 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2542 of the Vaudenay padding oracle attack on CBC mode encryption
2543 which enables an efficient plaintext recovery attack against
2544 the OpenSSL implementation of DTLS. Their attack exploits timing
2545 differences arising during decryption processing. A research
2546 paper describing this attack can be found at:
2547 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2548 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2549 Security Group at Royal Holloway, University of London
2550 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2551 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2552 for preparing the fix. (CVE-2011-4108)
2553 [Robin Seggelmann, Michael Tuexen]
2555 *) Clear bytes used for block padding of SSL 3.0 records.
2557 [Adam Langley (Google)]
2559 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2560 Kadianakis <desnacked@gmail.com> for discovering this issue and
2561 Adam Langley for preparing the fix. (CVE-2011-4619)
2562 [Adam Langley (Google)]
2564 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2565 [Andrey Kulikov <amdeich@gmail.com>]
2567 *) Prevent malformed RFC3779 data triggering an assertion failure.
2568 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2569 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2570 [Rob Austein <sra@hactrn.net>]
2572 *) Improved PRNG seeding for VOS.
2573 [Paul Green <Paul.Green@stratus.com>]
2575 *) Fix ssl_ciph.c set-up race.
2576 [Adam Langley (Google)]
2578 *) Fix spurious failures in ecdsatest.c.
2579 [Emilia Käsper (Google)]
2581 *) Fix the BIO_f_buffer() implementation (which was mixing different
2582 interpretations of the '..._len' fields).
2583 [Adam Langley (Google)]
2585 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2586 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2587 threads won't reuse the same blinding coefficients.
2589 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2590 lock to call BN_BLINDING_invert_ex, and avoids one use of
2591 BN_BLINDING_update for each BN_BLINDING structure (previously,
2592 the last update always remained unused).
2593 [Emilia Käsper (Google)]
2595 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2596 [Bob Buckholz (Google)]
2598 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2600 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2601 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2602 [Kaspar Brand <ossl@velox.ch>]
2604 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2605 for multi-threaded use of ECDH. (CVE-2011-3210)
2606 [Adam Langley (Google)]
2608 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2611 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2612 signature public key algorithm by using OID xref utilities instead.
2613 Before this you could only use some ECC ciphersuites with SHA1 only.
2616 *) Add protection against ECDSA timing attacks as mentioned in the paper
2617 by Billy Bob Brumley and Nicola Tuveri, see:
2619 http://eprint.iacr.org/2011/232.pdf
2621 [Billy Bob Brumley and Nicola Tuveri]
2623 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2625 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2626 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2628 *) Fix bug in string printing code: if *any* escaping is enabled we must
2629 escape the escape character (backslash) or the resulting string is
2633 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2635 *) Disable code workaround for ancient and obsolete Netscape browsers
2636 and servers: an attacker can use it in a ciphersuite downgrade attack.
2637 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2640 *) Fixed J-PAKE implementation error, originally discovered by
2641 Sebastien Martini, further info and confirmation from Stefan
2642 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2645 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2647 *) Fix extension code to avoid race conditions which can result in a buffer
2648 overrun vulnerability: resumed sessions must not be modified as they can
2649 be shared by multiple threads. CVE-2010-3864
2652 *) Fix WIN32 build system to correctly link an ENGINE directory into
2656 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2658 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2660 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2662 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2664 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2665 context. The operation can be customised via the ctrl mechanism in
2666 case ENGINEs want to include additional functionality.
2669 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2672 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2673 output hashes compatible with older versions of OpenSSL.
2674 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2676 *) Fix compression algorithm handling: if resuming a session use the
2677 compression algorithm of the resumed session instead of determining
2678 it from client hello again. Don't allow server to change algorithm.
2681 *) Add load_crls() function to apps tidying load_certs() too. Add option
2682 to verify utility to allow additional CRLs to be included.
2685 *) Update OCSP request code to permit adding custom headers to the request:
2686 some responders need this.
2689 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2691 [Julia Lawall <julia@diku.dk>]
2693 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2694 needlessly dereferenced structures, used obsolete functions and
2695 didn't handle all updated verify codes correctly.
2698 *) Disable MD2 in the default configuration.
2701 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2702 indicate the initial BIO being pushed or popped. This makes it possible
2703 to determine whether the BIO is the one explicitly called or as a result
2704 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2705 it handles reference counts correctly and doesn't zero out the I/O bio
2706 when it is not being explicitly popped. WARNING: applications which
2707 included workarounds for the old buggy behaviour will need to be modified
2708 or they could free up already freed BIOs.
2711 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2712 renaming to all platforms (within the 0.9.8 branch, this was
2713 done conditionally on Netware platforms to avoid a name clash).
2714 [Guenter <lists@gknw.net>]
2716 *) Add ECDHE and PSK support to DTLS.
2717 [Michael Tuexen <tuexen@fh-muenster.de>]
2719 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2723 *) Add "missing" function EVP_MD_flags() (without this the only way to
2724 retrieve a digest flags is by accessing the structure directly. Update
2725 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2726 or cipher is registered as in the "from" argument. Print out all
2727 registered digests in the dgst usage message instead of manually
2728 attempting to work them out.
2731 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2732 this allows the use of compression and extensions. Change default cipher
2733 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2734 by default unless an application cipher string requests it.
2737 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2738 key ids to find matching certificates and keys but some PKCS#12 files
2739 don't follow the (somewhat unwritten) rules and this strategy fails.
2740 Now just gather all certificates together and the first private key
2741 then look for the first certificate that matches the key.
2744 *) Support use of registered digest and cipher names for dgst and cipher
2745 commands instead of having to add each one as a special case. So now
2752 openssl dgst -sha256 foo
2754 and this works for ENGINE based algorithms too.
2758 *) Update Gost ENGINE to support parameter files.
2759 [Victor B. Wagner <vitus@cryptocom.ru>]
2761 *) Support GeneralizedTime in ca utility.
2762 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2764 *) Enhance the hash format used for certificate directory links. The new
2765 form uses the canonical encoding (meaning equivalent names will work
2766 even if they aren't identical) and uses SHA1 instead of MD5. This form
2767 is incompatible with the older format and as a result c_rehash should
2768 be used to rebuild symbolic links.
2771 *) Make PKCS#8 the default write format for private keys, replacing the
2772 traditional format. This form is standardised, more secure and doesn't
2773 include an implicit MD5 dependency.
2776 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2777 committed to OpenSSL should pass this lot as a minimum.
2780 *) Add session ticket override functionality for use by EAP-FAST.
2781 [Jouni Malinen <j@w1.fi>]
2783 *) Modify HMAC functions to return a value. Since these can be implemented
2784 in an ENGINE errors can occur.
2787 *) Type-checked OBJ_bsearch_ex.
2790 *) Type-checked OBJ_bsearch. Also some constification necessitated
2791 by type-checking. Still to come: TXT_DB, bsearch(?),
2792 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2796 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2797 seconds to a tm structure directly, instead of going through OS
2798 specific date routines. This avoids any issues with OS routines such
2799 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2800 and X509_time_adj_ex() to cover the extended range. The existing
2801 X509_time_adj() is still usable and will no longer have any date issues.
2804 *) Delta CRL support. New use deltas option which will attempt to locate
2805 and search any appropriate delta CRLs available.
2807 This work was sponsored by Google.
2810 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2811 code and add additional score elements. Validate alternate CRL paths
2812 as part of the CRL checking and indicate a new error "CRL path validation
2813 error" in this case. Applications wanting additional details can use
2814 the verify callback and check the new "parent" field. If this is not
2815 NULL CRL path validation is taking place. Existing applications wont
2816 see this because it requires extended CRL support which is off by
2819 This work was sponsored by Google.
2822 *) Support for freshest CRL extension.
2824 This work was sponsored by Google.
2827 *) Initial indirect CRL support. Currently only supported in the CRLs
2828 passed directly and not via lookup. Process certificate issuer
2829 CRL entry extension and lookup CRL entries by bother issuer name
2830 and serial number. Check and process CRL issuer entry in IDP extension.
2832 This work was sponsored by Google.
2835 *) Add support for distinct certificate and CRL paths. The CRL issuer
2836 certificate is validated separately in this case. Only enabled if
2837 an extended CRL support flag is set: this flag will enable additional
2838 CRL functionality in future.
2840 This work was sponsored by Google.
2843 *) Add support for policy mappings extension.
2845 This work was sponsored by Google.
2848 *) Fixes to pathlength constraint, self issued certificate handling,
2849 policy processing to align with RFC3280 and PKITS tests.
2851 This work was sponsored by Google.
2854 *) Support for name constraints certificate extension. DN, email, DNS
2855 and URI types are currently supported.
2857 This work was sponsored by Google.
2860 *) To cater for systems that provide a pointer-based thread ID rather
2861 than numeric, deprecate the current numeric thread ID mechanism and
2862 replace it with a structure and associated callback type. This
2863 mechanism allows a numeric "hash" to be extracted from a thread ID in
2864 either case, and on platforms where pointers are larger than 'long',
2865 mixing is done to help ensure the numeric 'hash' is usable even if it
2866 can't be guaranteed unique. The default mechanism is to use "&errno"
2867 as a pointer-based thread ID to distinguish between threads.
2869 Applications that want to provide their own thread IDs should now use
2870 CRYPTO_THREADID_set_callback() to register a callback that will call
2871 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2873 Note that ERR_remove_state() is now deprecated, because it is tied
2874 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2875 to free the current thread's error state should be replaced by
2876 ERR_remove_thread_state(NULL).
2878 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2879 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2880 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2881 application was previously providing a numeric thread callback that
2882 was inappropriate for distinguishing threads, then uniqueness might
2883 have been obtained with &errno that happened immediately in the
2884 intermediate development versions of OpenSSL; this is no longer the
2885 case, the numeric thread callback will now override the automatic use
2887 [Geoff Thorpe, with help from Bodo Moeller]
2889 *) Initial support for different CRL issuing certificates. This covers a
2890 simple case where the self issued certificates in the chain exist and
2891 the real CRL issuer is higher in the existing chain.
2893 This work was sponsored by Google.
2896 *) Removed effectively defunct crypto/store from the build.
2899 *) Revamp of STACK to provide stronger type-checking. Still to come:
2900 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2901 ASN1_STRING, CONF_VALUE.
2904 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2905 RAM on SSL connections. This option can save about 34k per idle SSL.
2908 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2909 STACK, TXT_DB, bsearch, qsort.
2912 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2913 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2914 support for data, signedData, compressedData, digestedData and
2915 encryptedData, envelopedData types included. Scripts to check against
2916 RFC4134 examples draft and interop and consistency checks of many
2917 content types and variants.
2920 *) Add options to enc utility to support use of zlib compression BIO.
2923 *) Extend mk1mf to support importing of options and assembly language
2924 files from Configure script, currently only included in VC-WIN32.
2925 The assembly language rules can now optionally generate the source
2926 files from the associated perl scripts.
2929 *) Implement remaining functionality needed to support GOST ciphersuites.
2930 Interop testing has been performed using CryptoPro implementations.
2931 [Victor B. Wagner <vitus@cryptocom.ru>]
2933 *) s390x assembler pack.
2936 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2940 *) Implement Opaque PRF Input TLS extension as specified in
2941 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2942 official specification yet and no extension type assignment by
2943 IANA exists, this extension (for now) will have to be explicitly
2944 enabled when building OpenSSL by providing the extension number
2945 to use. For example, specify an option
2947 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2949 to the "config" or "Configure" script to enable the extension,
2950 assuming extension number 0x9527 (which is a completely arbitrary
2951 and unofficial assignment based on the MD5 hash of the Internet
2952 Draft). Note that by doing so, you potentially lose
2953 interoperability with other TLS implementations since these might
2954 be using the same extension number for other purposes.
2956 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2957 opaque PRF input value to use in the handshake. This will create
2958 an interal copy of the length-'len' string at 'src', and will
2959 return non-zero for success.
2961 To get more control and flexibility, provide a callback function
2964 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2965 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2969 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2972 Callback function 'cb' will be called in handshakes, and is
2973 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2974 Argument 'arg' is for application purposes (the value as given to
2975 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2976 be provided to the callback function). The callback function
2977 has to return non-zero to report success: usually 1 to use opaque
2978 PRF input just if possible, or 2 to enforce use of the opaque PRF
2979 input. In the latter case, the library will abort the handshake
2980 if opaque PRF input is not successfully negotiated.
2982 Arguments 'peerinput' and 'len' given to the callback function
2983 will always be NULL and 0 in the case of a client. A server will
2984 see the client's opaque PRF input through these variables if
2985 available (NULL and 0 otherwise). Note that if the server
2986 provides an opaque PRF input, the length must be the same as the
2987 length of the client's opaque PRF input.
2989 Note that the callback function will only be called when creating
2990 a new session (session resumption can resume whatever was
2991 previously negotiated), and will not be called in SSL 2.0
2992 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2993 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2994 for applications that need to enforce opaque PRF input.
2998 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3001 [Victor B. Wagner <vitus@cryptocom.ru>]
3003 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3004 RFC4507bis. The encrypted ticket format is an encrypted encoded
3005 SSL_SESSION structure, that way new session features are automatically
3008 If a client application caches session in an SSL_SESSION structure
3009 support is transparent because tickets are now stored in the encoded
3012 The SSL_CTX structure automatically generates keys for ticket
3013 protection in servers so again support should be possible
3014 with no application modification.
3016 If a client or server wishes to disable RFC4507 support then the option
3017 SSL_OP_NO_TICKET can be set.
3019 Add a TLS extension debugging callback to allow the contents of any client
3020 or server extensions to be examined.
3022 This work was sponsored by Google.
3025 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3026 OpenSSL should now compile cleanly on gcc 4.2
3027 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3029 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3030 support including streaming MAC support: this is required for GOST
3031 ciphersuite support.
3032 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3034 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3035 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3036 to output in BER and PEM format.
3039 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3040 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3041 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3042 ENGINE support for HMAC keys which are unextractable. New -mac and
3043 -macopt options to dgst utility.
3046 *) New option -sigopt to dgst utility. Update dgst to use
3047 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3048 alternative signing parameters such as X9.31 or PSS in the dgst
3052 *) Change ssl_cipher_apply_rule(), the internal function that does
3053 the work each time a ciphersuite string requests enabling
3054 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3055 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3056 the order of disabled ciphersuites such that those ciphersuites
3057 that most recently went from enabled to disabled not only stay
3058 in order with respect to each other, but also have higher priority
3059 than other disabled ciphersuites the next time ciphersuites are
3062 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3063 the same ciphersuites as with "HIGH" alone, but in a specific
3064 order where the PSK ciphersuites come first (since they are the
3065 most recently disabled ciphersuites when "HIGH" is parsed).
3067 Also, change ssl_create_cipher_list() (using this new
3068 funcionality) such that between otherwise identical
3069 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3073 *) Change ssl_create_cipher_list() so that it automatically
3074 arranges the ciphersuites in reasonable order before starting
3075 to process the rule string. Thus, the definition for "DEFAULT"
3076 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3077 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3078 This makes it much easier to arrive at a reasonable default order
3079 in applications for which anonymous ciphers are OK (meaning
3080 that you can't actually use DEFAULT).
3081 [Bodo Moeller; suggested by Victor Duchovni]
3083 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3084 processing) into multiple integers instead of setting
3085 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3086 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3087 (These masks as well as the individual bit definitions are hidden
3088 away into the non-exported interface ssl/ssl_locl.h, so this
3089 change to the definition of the SSL_CIPHER structure shouldn't
3090 affect applications.) This give us more bits for each of these
3091 categories, so there is no longer a need to coagulate AES128 and
3092 AES256 into a single algorithm bit, and to coagulate Camellia128
3093 and Camellia256 into a single algorithm bit, which has led to all
3096 Thus, among other things, the kludge introduced in 0.9.7m and
3097 0.9.8e for masking out AES256 independently of AES128 or masking
3098 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3100 With the change, we also introduce new ciphersuite aliases that
3101 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3105 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3106 Use the leftmost N bytes of the signature input if the input is
3107 larger than the prime q (with N being the size in bytes of q).
3110 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3111 it yet and it is largely untested.
3114 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3117 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3118 some compilers (gcc 4.2 and later) reject their use. Safestack is
3119 reimplemented. Update ASN1 to avoid use of legacy functions.
3122 *) Win32/64 targets are linked with Winsock2.
3125 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3126 to external functions. This can be used to increase CRL handling
3127 efficiency especially when CRLs are very large by (for example) storing
3128 the CRL revoked certificates in a database.
3131 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3132 new CRLs added to a directory can be used. New command line option
3133 -verify_return_error to s_client and s_server. This causes real errors
3134 to be returned by the verify callback instead of carrying on no matter
3135 what. This reflects the way a "real world" verify callback would behave.
3138 *) GOST engine, supporting several GOST algorithms and public key formats.
3139 Kindly donated by Cryptocom.
3142 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3143 partitioned by DP are handled but no indirect CRL or reason partitioning
3144 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3145 selected via a scoring technique which handles IDP and AKID in CRLs.
3148 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3149 will ultimately be used for all verify operations: this will remove the
3150 X509_STORE dependency on certificate verification and allow alternative
3151 lookup methods. X509_STORE based implementations of these two callbacks.
3154 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3155 Modify get_crl() to find a valid (unexpired) CRL if possible.
3158 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3159 this would be called X509_CRL_cmp() but that name is already used by
3160 a function that just compares CRL issuer names. Cache several CRL
3161 extensions in X509_CRL structure and cache CRLDP in X509.
3164 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3165 this maps equivalent X509_NAME structures into a consistent structure.
3166 Name comparison can then be performed rapidly using memcmp().
3169 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3173 *) Allow digests to supply their own micalg string for S/MIME type using
3174 the ctrl EVP_MD_CTRL_MICALG.
3177 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3178 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3179 ctrl. It can then customise the structure before and/or after signing
3183 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3184 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3185 to free up any added signature OIDs.
3188 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3189 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3190 digest and cipher tables. New options added to openssl utility:
3191 list-message-digest-algorithms and list-cipher-algorithms.
3194 *) Change the array representation of binary polynomials: the list
3195 of degrees of non-zero coefficients is now terminated with -1.
3196 Previously it was terminated with 0, which was also part of the
3197 value; thus, the array representation was not applicable to
3198 polynomials where t^0 has coefficient zero. This change makes
3199 the array representation useful in a more general context.
3202 *) Various modifications and fixes to SSL/TLS cipher string
3203 handling. For ECC, the code now distinguishes between fixed ECDH
3204 with RSA certificates on the one hand and with ECDSA certificates
3205 on the other hand, since these are separate ciphersuites. The
3206 unused code for Fortezza ciphersuites has been removed.
3208 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3209 (not "ECDHE"). For consistency with the code for DH
3210 certificates, use of ECDH certificates is now considered ECDH
3211 authentication, not RSA or ECDSA authentication (the latter is
3212 merely the CA's signing algorithm and not actively used in the
3215 The temporary ciphersuite alias "ECCdraft" is no longer
3216 available, and ECC ciphersuites are no longer excluded from "ALL"
3217 and "DEFAULT". The following aliases now exist for RFC 4492
3218 ciphersuites, most of these by analogy with the DH case:
3220 kECDHr - ECDH cert, signed with RSA
3221 kECDHe - ECDH cert, signed with ECDSA
3222 kECDH - ECDH cert (signed with either RSA or ECDSA)
3223 kEECDH - ephemeral ECDH
3224 ECDH - ECDH cert or ephemeral ECDH
3230 AECDH - anonymous ECDH
3231 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3235 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3236 Use correct micalg parameters depending on digest(s) in signed message.
3239 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3240 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3243 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3244 an engine to register a method. Add ENGINE lookups for methods and
3245 functional reference processing.
3248 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3249 EVP_{Sign,Verify}* which allow an application to customise the signature
3253 *) New -resign option to smime utility. This adds one or more signers
3254 to an existing PKCS#7 signedData structure. Also -md option to use an
3255 alternative message digest algorithm for signing.
3258 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3259 create PKCS7 structures containing multiple signers. Update smime
3260 application to support multiple signers.
3263 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3267 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3268 Reorganize PBE internals to lookup from a static table using NIDs,
3269 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3270 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3271 PRF which will be automatically used with PBES2.
3274 *) Replace the algorithm specific calls to generate keys in "req" with the
3278 *) Update PKCS#7 enveloped data routines to use new API. This is now
3279 supported by any public key method supporting the encrypt operation. A
3280 ctrl is added to allow the public key algorithm to examine or modify
3281 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3285 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3286 a default digest type to use. In most cases this will be SHA1 but some
3287 algorithms (such as GOST) need to specify an alternative digest. The
3288 return value indicates how strong the preference is 1 means optional and
3289 2 is mandatory (that is it is the only supported type). Modify
3290 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3291 use the default md. Update openssl utilities to use the default digest
3292 type for signing if it is not explicitly indicated.
3295 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3296 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3297 signing method from the key type. This effectively removes the link
3298 between digests and public key types.
3301 *) Add an OID cross reference table and utility functions. Its purpose is to
3302 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3303 rsaEncryption. This will allow some of the algorithm specific hackery
3304 needed to use the correct OID to be removed.
3307 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3308 structures for PKCS7_sign(). They are now set up by the relevant public
3312 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3315 *) Add support for key derivation (agreement) in the API, DH method and
3319 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3320 public and private key formats. As a side effect these add additional
3321 command line functionality not previously available: DSA signatures can be
3322 generated and verified using pkeyutl and DH key support and generation in
3327 [Oliver Tappe <zooey@hirschkaefer.de>]
3329 *) New make target "install_html_docs" installs HTML renditions of the
3331 [Oliver Tappe <zooey@hirschkaefer.de>]
3333 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3334 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3335 support key and parameter generation and add initial key generation
3336 functionality for RSA.
3339 *) Add functions for main EVP_PKEY_method operations. The undocumented
3340 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3341 EVP_PKEY_{encrypt,decrypt}_old.
3344 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3345 key API, doesn't do much yet.
3348 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3349 public key algorithms. New option to openssl utility:
3350 "list-public-key-algorithms" to print out info.
3353 *) Implement the Supported Elliptic Curves Extension for
3354 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3357 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3358 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3361 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3362 utilities such as rsa, dsa, dsaparam etc except they process any key
3366 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3367 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3368 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3372 *) Initial support for pluggable public key ASN1.
3373 De-spaghettify the public key ASN1 handling. Move public and private
3374 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3375 algorithm specific handling to a single module within the relevant
3376 algorithm directory. Add functions to allow (near) opaque processing
3377 of public and private key structures.
3380 *) Implement the Supported Point Formats Extension for
3381 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3384 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3385 for the psk identity [hint] and the psk callback functions to the
3386 SSL_SESSION, SSL and SSL_CTX structure.
3389 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3393 SSL_CTX_use_psk_identity_hint
3394 SSL_get_psk_identity_hint
3395 SSL_get_psk_identity
3396 SSL_use_psk_identity_hint
3398 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3400 *) Add RFC 3161 compliant time stamp request creation, response generation
3401 and response verification functionality.
3402 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3404 *) Add initial support for TLS extensions, specifically for the server_name
3405 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3406 have new members for a host name. The SSL data structure has an
3407 additional member SSL_CTX *initial_ctx so that new sessions can be
3408 stored in that context to allow for session resumption, even after the
3409 SSL has been switched to a new SSL_CTX in reaction to a client's
3410 server_name extension.
3412 New functions (subject to change):
3414 SSL_get_servername()
3415 SSL_get_servername_type()
3418 New CTRL codes and macros (subject to change):
3420 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3421 - SSL_CTX_set_tlsext_servername_callback()
3422 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3423 - SSL_CTX_set_tlsext_servername_arg()
3424 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3426 openssl s_client has a new '-servername ...' option.
3428 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3429 '-key2 ...', '-servername_fatal' (subject to change). This allows
3430 testing the HostName extension for a specific single host name ('-cert'
3431 and '-key' remain fallbacks for handshakes without HostName
3432 negotiation). If the unrecognized_name alert has to be sent, this by
3433 default is a warning; it becomes fatal with the '-servername_fatal'
3436 [Peter Sylvester, Remy Allais, Christophe Renou]
3438 *) Whirlpool hash implementation is added.
3441 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3442 bn(64,32). Because of instruction set limitations it doesn't have
3443 any negative impact on performance. This was done mostly in order
3444 to make it possible to share assembler modules, such as bn_mul_mont
3445 implementations, between 32- and 64-bit builds without hassle.
3448 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3449 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3453 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3454 dedicated Montgomery multiplication procedure, is introduced.
3455 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3456 "64-bit" performance on certain 32-bit targets.
3459 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3460 in SSL structures. New SSL ctrl to set maximum send fragment size.
3461 Save memory by seeting the I/O buffer sizes dynamically instead of
3462 using the maximum available value.
3465 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3466 in addition to the text details.
3469 *) Very, very preliminary EXPERIMENTAL support for printing of general
3470 ASN1 structures. This currently produces rather ugly output and doesn't
3471 handle several customised structures at all.
3474 *) Integrated support for PVK file format and some related formats such
3475 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3476 these in the 'rsa' and 'dsa' utilities.
3479 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3482 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3483 place for the (very old) "NETSCAPE" format certificates which are now
3484 handled using new ASN1 code equivalents.
3487 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3488 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3489 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3492 *) Modify CRL distribution points extension code to print out previously
3493 unsupported fields. Enhance extension setting code to allow setting of
3497 *) Add print and set support for Issuing Distribution Point CRL extension.
3500 *) Change 'Configure' script to enable Camellia by default.
3503 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3505 *) When rejecting SSL/TLS records due to an incorrect version number, never
3506 update s->server with a new major version number. As of
3507 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3508 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3509 the previous behavior could result in a read attempt at NULL when
3510 receiving specific incorrect SSL/TLS records once record payload
3511 protection is active. (CVE-2010-0740)
3512 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3514 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3515 could be crashed if the relevant tables were not present (e.g. chrooted).
3516 [Tomas Hoger <thoger@redhat.com>]
3518 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3520 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3521 [Martin Olsson, Neel Mehta]
3523 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3524 accommodate for stack sorting, always a write lock!).
3527 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3528 excessive delays in the RAND_poll(): over a minute. As a workaround
3529 include a time check in the inner Heap32Next loop too.
3532 *) The code that handled flushing of data in SSL/TLS originally used the
3533 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3534 the problem outlined in PR#1949. The fix suggested there however can
3535 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3536 of Apache). So instead simplify the code to flush unconditionally.
3537 This should be fine since flushing with no data to flush is a no op.
3540 *) Handle TLS versions 2.0 and later properly and correctly use the
3541 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3542 off ancient servers have a habit of sticking around for a while...
3545 *) Modify compression code so it frees up structures without using the
3546 ex_data callbacks. This works around a problem where some applications
3547 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3548 restarting) then use compression (e.g. SSL with compression) later.
3549 This results in significant per-connection memory leaks and
3550 has caused some security issues including CVE-2008-1678 and
3554 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3555 change when encrypting or decrypting.
3558 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3559 connect and renegotiate with servers which do not support RI.
3560 Until RI is more widely deployed this option is enabled by default.
3563 *) Add "missing" ssl ctrls to clear options and mode.
3566 *) If client attempts to renegotiate and doesn't support RI respond with
3567 a no_renegotiation alert as required by RFC5746. Some renegotiating
3568 TLS clients will continue a connection gracefully when they receive
3569 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3570 waiting for a server hello which it will never receive. Now we treat a
3571 received no_renegotiation alert as a fatal error. This is because
3572 applications requesting a renegotiation might well expect it to succeed
3573 and would have no code in place to handle the server denying it so the
3574 only safe thing to do is to terminate the connection.
3577 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3578 peer supports secure renegotiation and 0 otherwise. Print out peer
3579 renegotiation support in s_client/s_server.
3582 *) Replace the highly broken and deprecated SPKAC certification method with
3583 the updated NID creation version. This should correctly handle UTF8.
3586 *) Implement RFC5746. Re-enable renegotiation but require the extension
3587 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3588 turns out to be a bad idea. It has been replaced by
3589 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3590 SSL_CTX_set_options(). This is really not recommended unless you
3591 know what you are doing.
3592 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3594 *) Fixes to stateless session resumption handling. Use initial_ctx when
3595 issuing and attempting to decrypt tickets in case it has changed during
3596 servername handling. Use a non-zero length session ID when attempting
3597 stateless session resumption: this makes it possible to determine if
3598 a resumption has occurred immediately after receiving server hello
3599 (several places in OpenSSL subtly assume this) instead of later in
3603 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3604 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3605 fixes for a few places where the return code is not checked
3607 [Julia Lawall <julia@diku.dk>]
3609 *) Add --strict-warnings option to Configure script to include devteam
3610 warnings in other configurations.
3613 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3614 makes it possible to install openssl libraries in locations which
3615 have names other than "lib", for example "/usr/lib64" which some
3617 [Steve Henson, based on patch from Jeremy Utley]
3619 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3620 X690 8.9.12 and can produce some misleading textual output of OIDs.
3621 [Steve Henson, reported by Dan Kaminsky]
3623 *) Delete MD2 from algorithm tables. This follows the recommendation in
3624 several standards that it is not used in new applications due to
3625 several cryptographic weaknesses. For binary compatibility reasons
3626 the MD2 API is still compiled in by default.
3629 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3633 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3634 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3636 [Guenter <lists@gknw.net>]
3638 *) Fix the server certificate chain building code to use X509_verify_cert(),
3639 it used to have an ad-hoc builder which was unable to cope with anything
3640 other than a simple chain.
3641 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3643 *) Don't check self signed certificate signatures in X509_verify_cert()
3644 by default (a flag can override this): it just wastes time without
3645 adding any security. As a useful side effect self signed root CAs
3646 with non-FIPS digests are now usable in FIPS mode.
3649 *) In dtls1_process_out_of_seq_message() the check if the current message
3650 is already buffered was missing. For every new message was memory
3651 allocated, allowing an attacker to perform an denial of service attack
3652 with sending out of seq handshake messages until there is no memory
3653 left. Additionally every future messege was buffered, even if the
3654 sequence number made no sense and would be part of another handshake.
3655 So only messages with sequence numbers less than 10 in advance will be
3656 buffered. (CVE-2009-1378)
3657 [Robin Seggelmann, discovered by Daniel Mentz]
3659 *) Records are buffered if they arrive with a future epoch to be
3660 processed after finishing the corresponding handshake. There is
3661 currently no limitation to this buffer allowing an attacker to perform
3662 a DOS attack with sending records with future epochs until there is no
3663 memory left. This patch adds the pqueue_size() function to determine
3664 the size of a buffer and limits the record buffer to 100 entries.
3666 [Robin Seggelmann, discovered by Daniel Mentz]
3668 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3669 parent structure is freed. (CVE-2009-1379)
3672 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3673 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3676 [Ilya O. <vrghost@gmail.com>]
3678 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3680 *) Disable renegotiation completely - this fixes a severe security
3681 problem (CVE-2009-3555) at the cost of breaking all
3682 renegotiation. Renegotiation can be re-enabled by setting
3683 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3684 run-time. This is really not recommended unless you know what
3688 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3690 *) Don't set val to NULL when freeing up structures, it is freed up by
3691 underlying code. If sizeof(void *) > sizeof(long) this can result in
3692 zeroing past the valid field. (CVE-2009-0789)
3693 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3695 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3696 checked correctly. This would allow some invalid signed attributes to
3697 appear to verify correctly. (CVE-2009-0591)
3698 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3700 *) Reject UniversalString and BMPString types with invalid lengths. This
3701 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3702 a legal length. (CVE-2009-0590)
3705 *) Set S/MIME signing as the default purpose rather than setting it
3706 unconditionally. This allows applications to override it at the store
3710 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3711 to handle some structures.
3714 *) Improve efficiency of mem_gets: don't search whole buffer each time
3716 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3718 *) New -hex option for openssl rand.
3721 *) Print out UTF8String and NumericString when parsing ASN1.
3724 *) Support NumericString type for name components.
3727 *) Allow CC in the environment to override the automatically chosen
3728 compiler. Note that nothing is done to ensure flags work with the
3732 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3734 *) Properly check EVP_VerifyFinal() and similar return values
3736 [Ben Laurie, Bodo Moeller, Google Security Team]
3738 *) Enable TLS extensions by default.
3741 *) Allow the CHIL engine to be loaded, whether the application is
3742 multithreaded or not. (This does not release the developer from the
3743 obligation to set up the dynamic locking callbacks.)
3744 [Sander Temme <sander@temme.net>]
3746 *) Use correct exit code if there is an error in dgst command.
3747 [Steve Henson; problem pointed out by Roland Dirlewanger]
3749 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3750 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3753 *) Add experimental JPAKE support, including demo authentication in
3754 s_client and s_server.
3757 *) Set the comparison function in v3_addr_canonize().
3758 [Rob Austein <sra@hactrn.net>]
3760 *) Add support for XMPP STARTTLS in s_client.
3761 [Philip Paeps <philip@freebsd.org>]
3763 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3764 to ensure that even with this option, only ciphersuites in the
3765 server's preference list will be accepted. (Note that the option
3766 applies only when resuming a session, so the earlier behavior was
3767 just about the algorithm choice for symmetric cryptography.)
3770 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3772 *) Fix NULL pointer dereference if a DTLS server received
3773 ChangeCipherSpec as first record (CVE-2009-1386).
3776 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3777 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3780 *) The fix in 0.9.8c that supposedly got rid of unsafe
3781 double-checked locking was incomplete for RSA blinding,
3782 addressing just one layer of what turns out to have been
3783 doubly unsafe triple-checked locking.
3785 So now fix this for real by retiring the MONT_HELPER macro
3786 in crypto/rsa/rsa_eay.c.
3788 [Bodo Moeller; problem pointed out by Marius Schilder]
3790 *) Various precautionary measures:
3792 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3794 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3795 (NB: This would require knowledge of the secret session ticket key
3796 to exploit, in which case you'd be SOL either way.)
3798 - Change bn_nist.c so that it will properly handle input BIGNUMs
3799 outside the expected range.
3801 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3804 [Neel Mehta, Bodo Moeller]
3806 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3807 the load fails. Useful for distros.
3808 [Ben Laurie and the FreeBSD team]
3810 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3813 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3816 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3818 This work was sponsored by Logica.
3821 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3822 keystores. Support for SSL/TLS client authentication too.
3823 Not compiled unless enable-capieng specified to Configure.
3825 This work was sponsored by Logica.
3828 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3829 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3830 attribute creation routines such as certifcate requests and PKCS#12
3834 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3836 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3837 handshake which could lead to a cilent crash as found using the
3838 Codenomicon TLS test suite (CVE-2008-1672)
3839 [Steve Henson, Mark Cox]
3841 *) Fix double free in TLS server name extensions which could lead to
3842 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3845 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3847 Clear the error queue to ensure that error entries left from
3848 older function calls do not interfere with the correct operation.
3849 [Lutz Jaenicke, Erik de Castro Lopo]
3851 *) Remove root CA certificates of commercial CAs:
3853 The OpenSSL project does not recommend any specific CA and does not
3854 have any policy with respect to including or excluding any CA.
3855 Therefore it does not make any sense to ship an arbitrary selection
3856 of root CA certificates with the OpenSSL software.
3859 *) RSA OAEP patches to fix two separate invalid memory reads.
3860 The first one involves inputs when 'lzero' is greater than
3861 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3862 before the beginning of from). The second one involves inputs where
3863 the 'db' section contains nothing but zeroes (there is a one-byte
3864 invalid read after the end of 'db').
3865 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3867 *) Partial backport from 0.9.9-dev:
3869 Introduce bn_mul_mont (dedicated Montgomery multiplication
3870 procedure) as a candidate for BIGNUM assembler implementation.
3871 While 0.9.9-dev uses assembler for various architectures, only
3872 x86_64 is available by default here in the 0.9.8 branch, and
3873 32-bit x86 is available through a compile-time setting.
3875 To try the 32-bit x86 assembler implementation, use Configure
3876 option "enable-montasm" (which exists only for this backport).
3878 As "enable-montasm" for 32-bit x86 disclaims code stability
3879 anyway, in this constellation we activate additional code
3880 backported from 0.9.9-dev for further performance improvements,
3881 namely BN_from_montgomery_word. (To enable this otherwise,
3882 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3884 [Andy Polyakov (backport partially by Bodo Moeller)]
3886 *) Add TLS session ticket callback. This allows an application to set
3887 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3888 values. This is useful for key rollover for example where several key
3889 sets may exist with different names.
3892 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3893 This was broken until now in 0.9.8 releases, such that the only way
3894 a registered ENGINE could be used (assuming it initialises
3895 successfully on the host) was to explicitly set it as the default
3896 for the relevant algorithms. This is in contradiction with 0.9.7
3897 behaviour and the documentation. With this fix, when an ENGINE is
3898 registered into a given algorithm's table of implementations, the
3899 'uptodate' flag is reset so that auto-discovery will be used next
3900 time a new context for that algorithm attempts to select an
3902 [Ian Lister (tweaked by Geoff Thorpe)]
3904 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3905 implemention in the following ways:
3907 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3910 Lack of BER streaming support means one pass streaming processing is
3911 only supported if data is detached: setting the streaming flag is
3912 ignored for embedded content.
3914 CMS support is disabled by default and must be explicitly enabled
3915 with the enable-cms configuration option.
3918 *) Update the GMP engine glue to do direct copies between BIGNUM and
3919 mpz_t when openssl and GMP use the same limb size. Otherwise the
3920 existing "conversion via a text string export" trick is still used.
3921 [Paul Sheer <paulsheer@gmail.com>]
3923 *) Zlib compression BIO. This is a filter BIO which compressed and
3924 uncompresses any data passed through it.
3927 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3928 RFC3394 compatible AES key wrapping.
3931 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3932 sets string data without copying. X509_ALGOR_set0() and
3933 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3934 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3935 from an X509_ATTRIBUTE structure optionally checking it occurs only
3936 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3940 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3941 to get the expected BN_FLG_CONSTTIME behavior.
3942 [Bodo Moeller (Google)]
3946 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3947 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3948 - added some more tests to do_tests.pl
3949 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3950 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3951 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3952 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3953 - various changes to netware.pl to enable gcc-cross builds on Win32
3955 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3956 - various changes to fix missing prototype warnings
3957 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3958 - added AES, WHIRLPOOL and CPUID assembler code to build files
3959 - added missing AES assembler make rules to mk1mf.pl
3960 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3961 [Guenter Knauf <eflash@gmx.net>]
3963 *) Implement certificate status request TLS extension defined in RFC3546.
3964 A client can set the appropriate parameters and receive the encoded
3965 OCSP response via a callback. A server can query the supplied parameters
3966 and set the encoded OCSP response in the callback. Add simplified examples
3967 to s_client and s_server.
3970 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3972 *) Fix various bugs:
3973 + Binary incompatibility of ssl_ctx_st structure
3974 + DTLS interoperation with non-compliant servers
3975 + Don't call get_session_cb() without proposed session
3976 + Fix ia64 assembler code
3977 [Andy Polyakov, Steve Henson]
3979 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3981 *) DTLS Handshake overhaul. There were longstanding issues with
3982 OpenSSL DTLS implementation, which were making it impossible for
3983 RFC 4347 compliant client to communicate with OpenSSL server.
3984 Unfortunately just fixing these incompatibilities would "cut off"
3985 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3986 server keeps tolerating non RFC compliant syntax. The opposite is
3987 not true, 0.9.8f client can not communicate with earlier server.
3988 This update even addresses CVE-2007-4995.
3991 *) Changes to avoid need for function casts in OpenSSL: some compilers
3992 (gcc 4.2 and later) reject their use.
3993 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3996 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3997 RFC4507bis. The encrypted ticket format is an encrypted encoded
3998 SSL_SESSION structure, that way new session features are automatically
4001 If a client application caches session in an SSL_SESSION structure
4002 support is transparent because tickets are now stored in the encoded
4005 The SSL_CTX structure automatically generates keys for ticket
4006 protection in servers so again support should be possible
4007 with no application modification.
4009 If a client or server wishes to disable RFC4507 support then the option
4010 SSL_OP_NO_TICKET can be set.
4012 Add a TLS extension debugging callback to allow the contents of any client
4013 or server extensions to be examined.
4015 This work was sponsored by Google.
4018 *) Add initial support for TLS extensions, specifically for the server_name
4019 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4020 have new members for a host name. The SSL data structure has an
4021 additional member SSL_CTX *initial_ctx so that new sessions can be
4022 stored in that context to allow for session resumption, even after the
4023 SSL has been switched to a new SSL_CTX in reaction to a client's
4024 server_name extension.
4026 New functions (subject to change):
4028 SSL_get_servername()
4029 SSL_get_servername_type()
4032 New CTRL codes and macros (subject to change):
4034 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4035 - SSL_CTX_set_tlsext_servername_callback()
4036 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4037 - SSL_CTX_set_tlsext_servername_arg()
4038 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4040 openssl s_client has a new '-servername ...' option.
4042 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4043 '-key2 ...', '-servername_fatal' (subject to change). This allows
4044 testing the HostName extension for a specific single host name ('-cert'
4045 and '-key' remain fallbacks for handshakes without HostName
4046 negotiation). If the unrecognized_name alert has to be sent, this by
4047 default is a warning; it becomes fatal with the '-servername_fatal'
4050 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4052 *) Add AES and SSE2 assembly language support to VC++ build.
4055 *) Mitigate attack on final subtraction in Montgomery reduction.
4058 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4059 (which previously caused an internal error).
4062 *) Squeeze another 10% out of IGE mode when in != out.
4065 *) AES IGE mode speedup.
4066 [Dean Gaudet (Google)]
4068 *) Add the Korean symmetric 128-bit cipher SEED (see
4069 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4070 add SEED ciphersuites from RFC 4162:
4072 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4073 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4074 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4075 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4077 To minimize changes between patchlevels in the OpenSSL 0.9.8
4078 series, SEED remains excluded from compilation unless OpenSSL
4079 is configured with 'enable-seed'.
4080 [KISA, Bodo Moeller]
4082 *) Mitigate branch prediction attacks, which can be practical if a
4083 single processor is shared, allowing a spy process to extract
4084 information. For detailed background information, see
4085 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4086 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4087 and Necessary Software Countermeasures"). The core of the change
4088 are new versions BN_div_no_branch() and
4089 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4090 respectively, which are slower, but avoid the security-relevant
4091 conditional branches. These are automatically called by BN_div()
4092 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4093 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4094 remove a conditional branch.
4096 BN_FLG_CONSTTIME is the new name for the previous
4097 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4098 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4099 in the exponent causes BN_mod_exp_mont() to use the alternative
4100 implementation in BN_mod_exp_mont_consttime().) The old name
4101 remains as a deprecated alias.
4103 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4104 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4105 constant-time implementations for more than just exponentiation.
4106 Here too the old name is kept as a deprecated alias.
4108 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4109 the BN_BLINDING structure gets an independent copy of the
4110 modulus. This means that the previous "BIGNUM *m" argument to
4111 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4112 essentially becomes "const BIGNUM *m", although we can't actually
4113 change this in the header file before 0.9.9. It allows
4114 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4115 enable BN_FLG_CONSTTIME.
4117 [Matthew D Wood (Intel Corp)]
4119 *) In the SSL/TLS server implementation, be strict about session ID
4120 context matching (which matters if an application uses a single
4121 external cache for different purposes). Previously,
4122 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4123 set. This did ensure strict client verification, but meant that,
4124 with applications using a single external cache for quite
4125 different requirements, clients could circumvent ciphersuite
4126 restrictions for a given session ID context by starting a session
4127 in a different context.
4130 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4131 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4132 authentication-only ciphersuites.
4135 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4136 not complete and could lead to a possible single byte overflow
4137 (CVE-2007-5135) [Ben Laurie]
4139 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4141 *) Since AES128 and AES256 (and similarly Camellia128 and
4142 Camellia256) share a single mask bit in the logic of
4143 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4144 kludge to work properly if AES128 is available and AES256 isn't
4145 (or if Camellia128 is available and Camellia256 isn't).
4148 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4149 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4150 When a point or a seed is encoded in a BIT STRING, we need to
4151 prevent the removal of trailing zero bits to get the proper DER
4152 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4153 of a NamedBitList, for which trailing 0 bits need to be removed.)
4156 *) Have SSL/TLS server implementation tolerate "mismatched" record
4157 protocol version while receiving ClientHello even if the
4158 ClientHello is fragmented. (The server can't insist on the
4159 particular protocol version it has chosen before the ServerHello
4160 message has informed the client about his choice.)
4163 *) Add RFC 3779 support.
4164 [Rob Austein for ARIN, Ben Laurie]
4166 *) Load error codes if they are not already present instead of using a
4167 static variable. This allows them to be cleanly unloaded and reloaded.
4168 Improve header file function name parsing.
4171 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4172 or CAPABILITY handshake as required by RFCs.
4175 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4177 *) Introduce limits to prevent malicious keys being able to
4178 cause a denial of service. (CVE-2006-2940)
4179 [Steve Henson, Bodo Moeller]
4181 *) Fix ASN.1 parsing of certain invalid structures that can result
4182 in a denial of service. (CVE-2006-2937) [Steve Henson]
4184 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4185 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4187 *) Fix SSL client code which could crash if connecting to a
4188 malicious SSLv2 server. (CVE-2006-4343)
4189 [Tavis Ormandy and Will Drewry, Google Security Team]
4191 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4192 match only those. Before that, "AES256-SHA" would be interpreted
4193 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4194 the same strength classification in 0.9.7h) as we currently only
4195 have a single AES bit in the ciphersuite description bitmap.
4196 That change, however, also applied to ciphersuite strings such as
4197 "RC4-MD5" that intentionally matched multiple ciphersuites --
4198 namely, SSL 2.0 ciphersuites in addition to the more common ones
4199 from SSL 3.0/TLS 1.0.
4201 So we change the selection algorithm again: Naming an explicit
4202 ciphersuite selects this one ciphersuite, and any other similar
4203 ciphersuite (same bitmap) from *other* protocol versions.
4204 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4205 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4207 Since SSL 2.0 does not have any ciphersuites for which the
4208 128/256 bit distinction would be relevant, this works for now.
4209 The proper fix will be to use different bits for AES128 and
4210 AES256, which would have avoided the problems from the beginning;
4211 however, bits are scarce, so we can only do this in a new release
4212 (not just a patchlevel) when we can change the SSL_CIPHER
4213 definition to split the single 'unsigned long mask' bitmap into
4214 multiple values to extend the available space.
4218 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4220 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4221 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4223 *) Add AES IGE and biIGE modes.
4226 *) Change the Unix randomness entropy gathering to use poll() when
4227 possible instead of select(), since the latter has some
4228 undesirable limitations.
4229 [Darryl Miles via Richard Levitte and Bodo Moeller]
4231 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4232 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4233 cannot be implicitly activated as part of, e.g., the "AES" alias.
4234 However, please upgrade to OpenSSL 0.9.9[-dev] for
4235 non-experimental use of the ECC ciphersuites to get TLS extension
4236 support, which is required for curve and point format negotiation
4237 to avoid potential handshake problems.
4240 *) Disable rogue ciphersuites:
4242 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4243 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4244 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4246 The latter two were purportedly from
4247 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4250 Also deactivate the remaining ciphersuites from
4251 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4252 unofficial, and the ID has long expired.
4255 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4256 dual-core machines) and other potential thread-safety issues.
4259 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4260 versions), which is now available for royalty-free use
4261 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4262 Also, add Camellia TLS ciphersuites from RFC 4132.
4264 To minimize changes between patchlevels in the OpenSSL 0.9.8
4265 series, Camellia remains excluded from compilation unless OpenSSL
4266 is configured with 'enable-camellia'.
4269 *) Disable the padding bug check when compression is in use. The padding
4270 bug check assumes the first packet is of even length, this is not
4271 necessarily true if compresssion is enabled and can result in false
4272 positives causing handshake failure. The actual bug test is ancient
4273 code so it is hoped that implementations will either have fixed it by
4274 now or any which still have the bug do not support compression.
4277 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4279 *) When applying a cipher rule check to see if string match is an explicit
4280 cipher suite and only match that one cipher suite if it is.
4283 *) Link in manifests for VC++ if needed.
4284 [Austin Ziegler <halostatue@gmail.com>]
4286 *) Update support for ECC-based TLS ciphersuites according to
4287 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4288 TLS extensions, which are supported starting with the 0.9.9
4289 branch, not in the OpenSSL 0.9.8 branch).
4292 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4293 opaque EVP_CIPHER_CTX handling.
4296 *) Fixes and enhancements to zlib compression code. We now only use
4297 "zlib1.dll" and use the default __cdecl calling convention on Win32
4298 to conform with the standards mentioned here:
4299 http://www.zlib.net/DLL_FAQ.txt
4300 Static zlib linking now works on Windows and the new --with-zlib-include
4301 --with-zlib-lib options to Configure can be used to supply the location
4302 of the headers and library. Gracefully handle case where zlib library
4306 *) Several fixes and enhancements to the OID generation code. The old code
4307 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4308 handle numbers larger than ULONG_MAX, truncated printing and had a
4309 non standard OBJ_obj2txt() behaviour.
4312 *) Add support for building of engines under engine/ as shared libraries
4313 under VC++ build system.
4316 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4317 Hopefully, we will not see any false combination of paths any more.
4320 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4322 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4323 (part of SSL_OP_ALL). This option used to disable the
4324 countermeasure against man-in-the-middle protocol-version
4325 rollback in the SSL 2.0 server implementation, which is a bad
4326 idea. (CVE-2005-2969)
4328 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4329 for Information Security, National Institute of Advanced Industrial
4330 Science and Technology [AIST], Japan)]
4332 *) Add two function to clear and return the verify parameter flags.
4335 *) Keep cipherlists sorted in the source instead of sorting them at
4336 runtime, thus removing the need for a lock.
4339 *) Avoid some small subgroup attacks in Diffie-Hellman.
4340 [Nick Mathewson and Ben Laurie]
4342 *) Add functions for well-known primes.
4345 *) Extended Windows CE support.
4346 [Satoshi Nakamura and Andy Polyakov]
4348 *) Initialize SSL_METHOD structures at compile time instead of during
4349 runtime, thus removing the need for a lock.
4352 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4353 attempting to decrypt each encrypted key in turn. Add support to
4357 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4359 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4362 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4365 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4366 key into the same file any more.
4369 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4372 *) Add -utf8 command line and config file option to 'ca'.
4373 [Stefan <stf@udoma.org]
4375 *) Removed the macro des_crypt(), as it seems to conflict with some
4376 libraries. Use DES_crypt().
4379 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4380 involves renaming the source and generated shared-libs for
4381 both. The engines will accept the corrected or legacy ids
4382 ('ncipher' and '4758_cca' respectively) when binding. NB,
4383 this only applies when building 'shared'.
4384 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4386 *) Add attribute functions to EVP_PKEY structure. Modify
4387 PKCS12_create() to recognize a CSP name attribute and
4388 use it. Make -CSP option work again in pkcs12 utility.
4391 *) Add new functionality to the bn blinding code:
4392 - automatic re-creation of the BN_BLINDING parameters after
4393 a fixed number of uses (currently 32)
4394 - add new function for parameter creation
4395 - introduce flags to control the update behaviour of the
4396 BN_BLINDING parameters
4397 - hide BN_BLINDING structure
4398 Add a second BN_BLINDING slot to the RSA structure to improve
4399 performance when a single RSA object is shared among several
4403 *) Add support for DTLS.
4404 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4406 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4407 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4410 *) Remove buggy and incomplete DH cert support from
4411 ssl/ssl_rsa.c and ssl/s3_both.c
4414 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4415 the apps/openssl applications.
4418 *) Compile clean with "-Wall -Wmissing-prototypes
4419 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4420 DEBUG_SAFESTACK must also be set.
4423 *) Change ./Configure so that certain algorithms can be disabled by default.
4424 The new counterpiece to "no-xxx" is "enable-xxx".
4426 The patented RC5 and MDC2 algorithms will now be disabled unless
4427 "enable-rc5" and "enable-mdc2", respectively, are specified.
4429 (IDEA remains enabled despite being patented. This is because IDEA
4430 is frequently required for interoperability, and there is no license
4431 fee for non-commercial use. As before, "no-idea" can be used to
4432 avoid this algorithm.)
4436 *) Add processing of proxy certificates (see RFC 3820). This work was
4437 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4438 EGEE (Enabling Grids for E-science in Europe).
4441 *) RC4 performance overhaul on modern architectures/implementations, such
4442 as Intel P4, IA-64 and AMD64.
4445 *) New utility extract-section.pl. This can be used specify an alternative
4446 section number in a pod file instead of having to treat each file as
4447 a separate case in Makefile. This can be done by adding two lines to the
4450 =for comment openssl_section:XXX
4452 The blank line is mandatory.
4456 *) New arguments -certform, -keyform and -pass for s_client and s_server
4457 to allow alternative format key and certificate files and passphrase
4461 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4462 update associated structures and add various utility functions.
4464 Add new policy related verify parameters, include policy checking in
4465 standard verify code. Enhance 'smime' application with extra parameters
4466 to support policy checking and print out.
4469 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4470 Nehemiah processors. These extensions support AES encryption in hardware
4471 as well as RNG (though RNG support is currently disabled).
4472 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4474 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4477 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4478 [Andy Polyakov and a number of other people]
4480 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4481 implementation contributed by IBM.
4482 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4484 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4485 exponent rather than 'unsigned long'. There is a corresponding change to
4486 the new 'rsa_keygen' element of the RSA_METHOD structure.
4487 [Jelte Jansen, Geoff Thorpe]
4489 *) Functionality for creating the initial serial number file is now
4490 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4492 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4493 number file to 1, which is bound to cause problems. To avoid
4494 the problems while respecting compatibility between different 0.9.7
4495 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4496 CA.pl for serial number initialization. With the new release 0.9.8,
4497 we can fix the problem directly in the 'ca' utility.)
4500 *) Reduced header interdepencies by declaring more opaque objects in
4501 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4502 give fewer recursive includes, which could break lazy source code - so
4503 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4504 developers should define this symbol when building and using openssl to
4505 ensure they track the recommended behaviour, interfaces, [etc], but
4506 backwards-compatible behaviour prevails when this isn't defined.
4509 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4512 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4513 This will generate a random key of the appropriate length based on the
4514 cipher context. The EVP_CIPHER can provide its own random key generation
4515 routine to support keys of a specific form. This is used in the des and
4516 3des routines to generate a key of the correct parity. Update S/MIME
4517 code to use new functions and hence generate correct parity DES keys.
4518 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4519 valid (weak or incorrect parity).
4522 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4523 as looking them up. This is useful when the verified structure may contain
4524 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4525 present unless the new PKCS7_NO_CRL flag is asserted.
4528 *) Extend ASN1 oid configuration module. It now additionally accepts the
4531 shortName = some long name, 1.2.3.4
4534 *) Reimplemented the BN_CTX implementation. There is now no more static
4535 limitation on the number of variables it can handle nor the depth of the
4536 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4537 information can now expand as required, and rather than having a single
4538 static array of bignums, BN_CTX now uses a linked-list of such arrays
4539 allowing it to expand on demand whilst maintaining the usefulness of
4540 BN_CTX's "bundling".
4543 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4544 to allow all RSA operations to function using a single BN_CTX.
4547 *) Preliminary support for certificate policy evaluation and checking. This
4548 is initially intended to pass the tests outlined in "Conformance Testing
4549 of Relying Party Client Certificate Path Processing Logic" v1.07.
4552 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4553 remained unused and not that useful. A variety of other little bignum
4554 tweaks and fixes have also been made continuing on from the audit (see
4558 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4559 associated ASN1, EVP and SSL functions and old ASN1 macros.
4562 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4563 and this should never fail. So the return value from the use of
4564 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4565 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4568 *) BN_CTX_get() should return zero-valued bignums, providing the same
4569 initialised value as BN_new().
4570 [Geoff Thorpe, suggested by Ulf Möller]
4572 *) Support for inhibitAnyPolicy certificate extension.
4575 *) An audit of the BIGNUM code is underway, for which debugging code is
4576 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4577 is considered valid when processing BIGNUMs, and causes execution to
4578 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4579 further steps are taken to deliberately pollute unused data in BIGNUM
4580 structures to try and expose faulty code further on. For now, openssl will
4581 (in its default mode of operation) continue to tolerate the inconsistent
4582 forms that it has tolerated in the past, but authors and packagers should
4583 consider trying openssl and their own applications when compiled with
4584 these debugging symbols defined. It will help highlight potential bugs in
4585 their own code, and will improve the test coverage for OpenSSL itself. At
4586 some point, these tighter rules will become openssl's default to improve
4587 maintainability, though the assert()s and other overheads will remain only
4588 in debugging configurations. See bn.h for more details.
4589 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4591 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4592 that can only be obtained through BN_CTX_new() (which implicitly
4593 initialises it). The presence of this function only made it possible
4594 to overwrite an existing structure (and cause memory leaks).
4597 *) Because of the callback-based approach for implementing LHASH as a
4598 template type, lh_insert() adds opaque objects to hash-tables and
4599 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4600 to clean up those corresponding objects before destroying the hash table
4601 (and losing the object pointers). So some over-zealous constifications in
4602 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4603 objects as "const" and the lh_doall[_arg] callback wrappers are not
4604 prototyped to have "const" restrictions on the object pointers they are
4605 given (and so aren't required to cast them away any more).
4608 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4609 (speed) prefers to use its own implementation. The two implementations
4610 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4611 its object type properly exposed (MS_TM) instead of casting to/from "char
4612 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4613 aren't necessarily the greatest nomenclatures - but this is what was used
4614 internally to the implementation so I've used that for now.
4617 *) Ensure that deprecated functions do not get compiled when
4618 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4619 the self-tests were still using deprecated key-generation functions so
4620 these have been updated also.
4623 *) Reorganise PKCS#7 code to separate the digest location functionality
4624 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4625 New function PKCS7_set_digest() to set the digest type for PKCS#7
4626 digestedData type. Add additional code to correctly generate the
4627 digestedData type and add support for this type in PKCS7 initialization
4631 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4632 structure of type "other".
4635 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4636 sure the loop does correctly stop and breaking ("division by zero")
4637 modulus operations are not performed. The (pre-generated) prime
4638 table crypto/bn/bn_prime.h was already correct, but it could not be
4639 re-generated on some platforms because of the "division by zero"
4640 situation in the script.
4641 [Ralf S. Engelschall]
4643 *) Update support for ECC-based TLS ciphersuites according to
4644 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4645 SHA-1 now is only used for "small" curves (where the
4646 representation of a field element takes up to 24 bytes); for
4647 larger curves, the field element resulting from ECDH is directly
4648 used as premaster secret.
4649 [Douglas Stebila (Sun Microsystems Laboratories)]
4651 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4652 curve secp160r1 to the tests.
4653 [Douglas Stebila (Sun Microsystems Laboratories)]
4655 *) Add the possibility to load symbols globally with DSO.
4656 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4658 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4659 control of the error stack.
4662 *) Add support for STORE in ENGINE.
4665 *) Add the STORE type. The intention is to provide a common interface
4666 to certificate and key stores, be they simple file-based stores, or
4667 HSM-type store, or LDAP stores, or...
4668 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4671 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4672 pass a list of arguments to any function as well as provide a way
4673 for a function to pass data back to the caller.
4676 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
4677 works like BUF_strdup() but can be used to duplicate a portion of
4678 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4682 *) Add the function sk_find_ex() which works like sk_find(), but will
4683 return an index to an element even if an exact match couldn't be
4684 found. The index is guaranteed to point at the element where the
4685 searched-for key would be inserted to preserve sorting order.
4688 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4689 takes an extra flags argument for optional functionality. Currently,
4690 the following flags are defined:
4692 OBJ_BSEARCH_VALUE_ON_NOMATCH
4693 This one gets OBJ_bsearch_ex() to return a pointer to the first
4694 element where the comparing function returns a negative or zero
4697 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4698 This one gets OBJ_bsearch_ex() to return a pointer to the first
4699 element where the comparing function returns zero. This is useful
4700 if there are more than one element where the comparing function
4704 *) Make it possible to create self-signed certificates with 'openssl ca'
4705 in such a way that the self-signed certificate becomes part of the
4706 CA database and uses the same mechanisms for serial number generation
4707 as all other certificate signing. The new flag '-selfsign' enables
4708 this functionality. Adapt CA.sh and CA.pl.in.
4711 *) Add functionality to check the public key of a certificate request
4712 against a given private. This is useful to check that a certificate
4713 request can be signed by that key (self-signing).
4716 *) Make it possible to have multiple active certificates with the same
4717 subject in the CA index file. This is done only if the keyword
4718 'unique_subject' is set to 'no' in the main CA section (default
4719 if 'CA_default') of the configuration file. The value is saved
4720 with the database itself in a separate index attribute file,
4721 named like the index file with '.attr' appended to the name.
4724 *) Generate muti valued AVAs using '+' notation in config files for
4728 *) Support for nameConstraints certificate extension.
4731 *) Support for policyConstraints certificate extension.
4734 *) Support for policyMappings certificate extension.
4737 *) Make sure the default DSA_METHOD implementation only uses its
4738 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4739 and change its own handlers to be NULL so as to remove unnecessary
4740 indirection. This lets alternative implementations fallback to the
4741 default implementation more easily.
4744 *) Support for directoryName in GeneralName related extensions
4748 *) Make it possible to link applications using Makefile.shared.
4749 Make that possible even when linking against static libraries!
4752 *) Support for single pass processing for S/MIME signing. This now
4753 means that S/MIME signing can be done from a pipe, in addition
4754 cleartext signing (multipart/signed type) is effectively streaming
4755 and the signed data does not need to be all held in memory.
4757 This is done with a new flag PKCS7_STREAM. When this flag is set
4758 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4759 is done after the data is output (and digests calculated) in
4760 SMIME_write_PKCS7().
4763 *) Add full support for -rpath/-R, both in shared libraries and
4764 applications, at least on the platforms where it's known how
4768 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4769 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4770 will now compute a table of multiples of the generator that
4771 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4772 faster (notably in the case of a single point multiplication,
4773 scalar * generator).
4774 [Nils Larsch, Bodo Moeller]
4776 *) IPv6 support for certificate extensions. The various extensions
4777 which use the IP:a.b.c.d can now take IPv6 addresses using the
4778 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4782 *) Added an ENGINE that implements RSA by performing private key
4783 exponentiations with the GMP library. The conversions to and from
4784 GMP's mpz_t format aren't optimised nor are any montgomery forms
4785 cached, and on x86 it appears OpenSSL's own performance has caught up.
4786 However there are likely to be other architectures where GMP could
4787 provide a boost. This ENGINE is not built in by default, but it can be
4788 specified at Configure time and should be accompanied by the necessary
4789 linker additions, eg;
4790 ./config -DOPENSSL_USE_GMP -lgmp
4793 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4794 testing availability of engines with "-t" - the old behaviour is
4795 produced by increasing the feature's verbosity with "-tt".
4798 *) ECDSA routines: under certain error conditions uninitialized BN objects
4799 could be freed. Solution: make sure initialization is performed early
4800 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4804 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4805 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4806 software implementations. For DSA and DH, parameter generation can
4807 also be overriden by providing the appropriate method callbacks.
4810 *) Change the "progress" mechanism used in key-generation and
4811 primality testing to functions that take a new BN_GENCB pointer in
4812 place of callback/argument pairs. The new API functions have "_ex"
4813 postfixes and the older functions are reimplemented as wrappers for
4814 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4815 declarations of the old functions to help (graceful) attempts to
4816 migrate to the new functions. Also, the new key-generation API
4817 functions operate on a caller-supplied key-structure and return
4818 success/failure rather than returning a key or NULL - this is to
4819 help make "keygen" another member function of RSA_METHOD etc.
4821 Example for using the new callback interface:
4823 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4827 BN_GENCB_set(&my_cb, my_callback, my_arg);
4829 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4830 /* For the meaning of a, b in calls to my_callback(), see the
4831 * documentation of the function that calls the callback.
4832 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4833 * my_callback should return 1 if it wants BN_is_prime_ex()
4834 * to continue, or 0 to stop.
4839 *) Change the ZLIB compression method to be stateful, and make it
4840 available to TLS with the number defined in
4841 draft-ietf-tls-compression-04.txt.
4844 *) Add the ASN.1 structures and functions for CertificatePair, which
4845 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4847 CertificatePair ::= SEQUENCE {
4848 forward [0] Certificate OPTIONAL,
4849 reverse [1] Certificate OPTIONAL,
4850 -- at least one of the pair shall be present -- }
4852 Also implement the PEM functions to read and write certificate
4853 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4855 This needed to be defined, mostly for the sake of the LDAP
4856 attribute crossCertificatePair, but may prove useful elsewhere as
4860 *) Make it possible to inhibit symlinking of shared libraries in
4861 Makefile.shared, for Cygwin's sake.
4864 *) Extend the BIGNUM API by creating a function
4865 void BN_set_negative(BIGNUM *a, int neg);
4866 and a macro that behave like
4867 int BN_is_negative(const BIGNUM *a);
4869 to avoid the need to access 'a->neg' directly in applications.
4872 *) Implement fast modular reduction for pseudo-Mersenne primes
4873 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4874 EC_GROUP_new_curve_GFp() will now automatically use this
4876 [Nils Larsch <nla@trustcenter.de>]
4878 *) Add new lock type (CRYPTO_LOCK_BN).
4881 *) Change the ENGINE framework to automatically load engines
4882 dynamically from specific directories unless they could be
4883 found to already be built in or loaded. Move all the
4884 current engines except for the cryptodev one to a new
4886 The engines in engines/ are built as shared libraries if
4887 the "shared" options was given to ./Configure or ./config.
4888 Otherwise, they are inserted in libcrypto.a.
4889 /usr/local/ssl/engines is the default directory for dynamic
4890 engines, but that can be overriden at configure time through
4891 the usual use of --prefix and/or --openssldir, and at run
4892 time with the environment variable OPENSSL_ENGINES.
4893 [Geoff Thorpe and Richard Levitte]
4895 *) Add Makefile.shared, a helper makefile to build shared
4896 libraries. Adapt Makefile.org.
4899 *) Add version info to Win32 DLLs.
4900 [Peter 'Luna' Runestig" <peter@runestig.com>]
4902 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4903 can be added using this API to created arbitrary PKCS#12
4904 files while avoiding the low level API.
4906 New options to PKCS12_create(), key or cert can be NULL and
4907 will then be omitted from the output file. The encryption
4908 algorithm NIDs can be set to -1 for no encryption, the mac
4909 iteration count can be set to 0 to omit the mac.
4911 Enhance pkcs12 utility by making the -nokeys and -nocerts
4912 options work when creating a PKCS#12 file. New option -nomac
4913 to omit the mac, NONE can be set for an encryption algorithm.
4914 New code is modified to use the enhanced PKCS12_create()
4915 instead of the low level API.
4918 *) Extend ASN1 encoder to support indefinite length constructed
4919 encoding. This can output sequences tags and octet strings in
4920 this form. Modify pk7_asn1.c to support indefinite length
4921 encoding. This is experimental and needs additional code to
4922 be useful, such as an ASN1 bio and some enhanced streaming
4925 Extend template encode functionality so that tagging is passed
4926 down to the template encoder.
4929 *) Let 'openssl req' fail if an argument to '-newkey' is not
4930 recognized instead of using RSA as a default.
4933 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4934 As these are not official, they are not included in "ALL";
4935 the "ECCdraft" ciphersuite group alias can be used to select them.
4936 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4938 *) Add ECDH engine support.
4939 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4941 *) Add ECDH in new directory crypto/ecdh/.
4942 [Douglas Stebila (Sun Microsystems Laboratories)]
4944 *) Let BN_rand_range() abort with an error after 100 iterations
4945 without success (which indicates a broken PRNG).
4948 *) Change BN_mod_sqrt() so that it verifies that the input value
4949 is really the square of the return value. (Previously,
4950 BN_mod_sqrt would show GIGO behaviour.)
4953 *) Add named elliptic curves over binary fields from X9.62, SECG,
4954 and WAP/WTLS; add OIDs that were still missing.
4956 [Sheueling Chang Shantz and Douglas Stebila
4957 (Sun Microsystems Laboratories)]
4959 *) Extend the EC library for elliptic curves over binary fields
4960 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4963 EC_GF2m_simple_method
4967 EC_GROUP_new_curve_GF2m
4968 EC_GROUP_set_curve_GF2m
4969 EC_GROUP_get_curve_GF2m
4970 EC_POINT_set_affine_coordinates_GF2m
4971 EC_POINT_get_affine_coordinates_GF2m
4972 EC_POINT_set_compressed_coordinates_GF2m
4974 Point compression for binary fields is disabled by default for
4975 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4978 As binary polynomials are represented as BIGNUMs, various members
4979 of the EC_GROUP and EC_POINT data structures can be shared
4980 between the implementations for prime fields and binary fields;
4981 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4982 are essentially identical to their ..._GFp counterparts.
4983 (For simplicity, the '..._GFp' prefix has been dropped from
4984 various internal method names.)
4986 An internal 'field_div' method (similar to 'field_mul' and
4987 'field_sqr') has been added; this is used only for binary fields.
4989 [Sheueling Chang Shantz and Douglas Stebila
4990 (Sun Microsystems Laboratories)]
4992 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4993 through methods ('mul', 'precompute_mult').
4995 The generic implementations (now internally called 'ec_wNAF_mul'
4996 and 'ec_wNAF_precomputed_mult') remain the default if these
4997 methods are undefined.
4999 [Sheueling Chang Shantz and Douglas Stebila
5000 (Sun Microsystems Laboratories)]
5002 *) New function EC_GROUP_get_degree, which is defined through
5003 EC_METHOD. For curves over prime fields, this returns the bit
5004 length of the modulus.
5006 [Sheueling Chang Shantz and Douglas Stebila
5007 (Sun Microsystems Laboratories)]
5009 *) New functions EC_GROUP_dup, EC_POINT_dup.
5010 (These simply call ..._new and ..._copy).
5012 [Sheueling Chang Shantz and Douglas Stebila
5013 (Sun Microsystems Laboratories)]
5015 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5016 Polynomials are represented as BIGNUMs (where the sign bit is not
5017 used) in the following functions [macros]:
5020 BN_GF2m_sub [= BN_GF2m_add]
5021 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5022 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5023 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5025 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5026 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5027 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5028 BN_GF2m_cmp [= BN_ucmp]
5030 (Note that only the 'mod' functions are actually for fields GF(2^m).
5031 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5033 For some functions, an the irreducible polynomial defining a
5034 field can be given as an 'unsigned int[]' with strictly
5035 decreasing elements giving the indices of those bits that are set;
5036 i.e., p[] represents the polynomial
5037 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5039 p[0] > p[1] > ... > p[k] = 0.
5040 This applies to the following functions:
5045 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5046 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5048 BN_GF2m_mod_sqrt_arr
5049 BN_GF2m_mod_solve_quad_arr
5053 Conversion can be performed by the following functions:
5058 bntest.c has additional tests for binary polynomial arithmetic.
5060 Two implementations for BN_GF2m_mod_div() are available.
5061 The default algorithm simply uses BN_GF2m_mod_inv() and
5062 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5063 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5064 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5066 [Sheueling Chang Shantz and Douglas Stebila
5067 (Sun Microsystems Laboratories)]
5069 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5070 functionality is disabled at compile-time.
5071 [Douglas Stebila <douglas.stebila@sun.com>]
5073 *) Change default behaviour of 'openssl asn1parse' so that more
5074 information is visible when viewing, e.g., a certificate:
5076 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5077 mode the content of non-printable OCTET STRINGs is output in a
5078 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5079 avoid the appearance of a printable string.
5080 [Nils Larsch <nla@trustcenter.de>]
5082 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5084 EC_GROUP_set_asn1_flag()
5085 EC_GROUP_get_asn1_flag()
5086 EC_GROUP_set_point_conversion_form()
5087 EC_GROUP_get_point_conversion_form()
5088 These control ASN1 encoding details:
5089 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5090 has been set to OPENSSL_EC_NAMED_CURVE.
5091 - Points are encoded in uncompressed form by default; options for
5092 asn1_for are as for point2oct, namely
5093 POINT_CONVERSION_COMPRESSED
5094 POINT_CONVERSION_UNCOMPRESSED
5095 POINT_CONVERSION_HYBRID
5097 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5100 EC_GROUP_get0_seed()
5101 EC_GROUP_get_seed_len()
5102 This is used only for ASN1 purposes (so far).
5103 [Nils Larsch <nla@trustcenter.de>]
5105 *) Add 'field_type' member to EC_METHOD, which holds the NID
5106 of the appropriate field type OID. The new function
5107 EC_METHOD_get_field_type() returns this value.
5108 [Nils Larsch <nla@trustcenter.de>]
5113 EC_POINT_point2hex()
5114 EC_POINT_hex2point()
5115 providing useful interfaces to EC_POINT_point2oct() and
5116 EC_POINT_oct2point().
5117 [Nils Larsch <nla@trustcenter.de>]
5119 *) Change internals of the EC library so that the functions
5120 EC_GROUP_set_generator()
5121 EC_GROUP_get_generator()
5122 EC_GROUP_get_order()
5123 EC_GROUP_get_cofactor()
5124 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5125 to methods, which would lead to unnecessary code duplication when
5126 adding different types of curves.
5127 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5129 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5130 arithmetic, and such that modified wNAFs are generated
5131 (which avoid length expansion in many cases).
5134 *) Add a function EC_GROUP_check_discriminant() (defined via
5135 EC_METHOD) that verifies that the curve discriminant is non-zero.
5137 Add a function EC_GROUP_check() that makes some sanity tests
5138 on a EC_GROUP, its generator and order. This includes
5139 EC_GROUP_check_discriminant().
5140 [Nils Larsch <nla@trustcenter.de>]
5142 *) Add ECDSA in new directory crypto/ecdsa/.
5144 Add applications 'openssl ecparam' and 'openssl ecdsa'
5145 (these are based on 'openssl dsaparam' and 'openssl dsa').
5147 ECDSA support is also included in various other files across the
5148 library. Most notably,
5149 - 'openssl req' now has a '-newkey ecdsa:file' option;
5150 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5151 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5152 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5153 them suitable for ECDSA where domain parameters must be
5154 extracted before the specific public key;
5155 - ECDSA engine support has been added.
5156 [Nils Larsch <nla@trustcenter.de>]
5158 *) Include some named elliptic curves, and add OIDs from X9.62,
5159 SECG, and WAP/WTLS. Each curve can be obtained from the new
5161 EC_GROUP_new_by_curve_name(),
5162 and the list of available named curves can be obtained with
5163 EC_get_builtin_curves().
5164 Also add a 'curve_name' member to EC_GROUP objects, which can be
5166 EC_GROUP_set_curve_name()
5167 EC_GROUP_get_curve_name()
5168 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5170 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5171 was actually never needed) and in BN_mul(). The removal in BN_mul()
5172 required a small change in bn_mul_part_recursive() and the addition
5173 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5174 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5175 bn_sub_words() and bn_add_words() except they take arrays with
5179 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5181 *) Cleanse PEM buffers before freeing them since they may contain
5183 [Benjamin Bennett <ben@psc.edu>]
5185 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5186 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5187 authentication-only ciphersuites.
5190 *) Since AES128 and AES256 share a single mask bit in the logic of
5191 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5192 kludge to work properly if AES128 is available and AES256 isn't.
5195 *) Expand security boundary to match 1.1.1 module.
5198 *) Remove redundant features: hash file source, editing of test vectors
5199 modify fipsld to use external fips_premain.c signature.
5202 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5203 run algorithm test programs.
5206 *) Make algorithm test programs more tolerant of whitespace.
5209 *) Have SSL/TLS server implementation tolerate "mismatched" record
5210 protocol version while receiving ClientHello even if the
5211 ClientHello is fragmented. (The server can't insist on the
5212 particular protocol version it has chosen before the ServerHello
5213 message has informed the client about his choice.)
5216 *) Load error codes if they are not already present instead of using a
5217 static variable. This allows them to be cleanly unloaded and reloaded.
5220 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5222 *) Introduce limits to prevent malicious keys being able to
5223 cause a denial of service. (CVE-2006-2940)
5224 [Steve Henson, Bodo Moeller]
5226 *) Fix ASN.1 parsing of certain invalid structures that can result
5227 in a denial of service. (CVE-2006-2937) [Steve Henson]
5229 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5230 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5232 *) Fix SSL client code which could crash if connecting to a
5233 malicious SSLv2 server. (CVE-2006-4343)
5234 [Tavis Ormandy and Will Drewry, Google Security Team]
5236 *) Change ciphersuite string processing so that an explicit
5237 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5238 will no longer include "AES128-SHA"), and any other similar
5239 ciphersuite (same bitmap) from *other* protocol versions (so that
5240 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5241 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5242 changes from 0.9.8b and 0.9.8d.
5245 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5247 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5248 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5250 *) Change the Unix randomness entropy gathering to use poll() when
5251 possible instead of select(), since the latter has some
5252 undesirable limitations.
5253 [Darryl Miles via Richard Levitte and Bodo Moeller]
5255 *) Disable rogue ciphersuites:
5257 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5258 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5259 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5261 The latter two were purportedly from
5262 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5265 Also deactive the remaining ciphersuites from
5266 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5267 unofficial, and the ID has long expired.
5270 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5271 dual-core machines) and other potential thread-safety issues.
5274 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5276 *) Adapt fipsld and the build system to link against the validated FIPS
5277 module in FIPS mode.
5280 *) Fixes for VC++ 2005 build under Windows.
5283 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5284 from a Windows bash shell such as MSYS. It is autodetected from the
5285 "config" script when run from a VC++ environment. Modify standard VC++
5286 build to use fipscanister.o from the GNU make build.
5289 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5291 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5292 The value now differs depending on if you build for FIPS or not.
5293 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5294 safely run with a non-FIPSed libcrypto, as it may crash because of
5295 the difference induced by this change.
5298 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5300 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5301 (part of SSL_OP_ALL). This option used to disable the
5302 countermeasure against man-in-the-middle protocol-version
5303 rollback in the SSL 2.0 server implementation, which is a bad
5304 idea. (CVE-2005-2969)
5306 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5307 for Information Security, National Institute of Advanced Industrial
5308 Science and Technology [AIST], Japan)]
5310 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5311 mainly for FIPS compliance and not fully integrated at this stage.
5314 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5315 the exponentiation using a fixed-length exponent. (Otherwise,
5316 the information leaked through timing could expose the secret key
5317 after many signatures; cf. Bleichenbacher's attack on DSA with
5321 *) Make a new fixed-window mod_exp implementation the default for
5322 RSA, DSA, and DH private-key operations so that the sequence of
5323 squares and multiplies and the memory access pattern are
5324 independent of the particular secret key. This will mitigate
5325 cache-timing and potential related attacks.
5327 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5328 and this is automatically used by BN_mod_exp_mont() if the new flag
5329 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5330 will use this BN flag for private exponents unless the flag
5331 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5332 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5334 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5336 *) Change the client implementation for SSLv23_method() and
5337 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5338 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5339 (Previously, the SSL 2.0 backwards compatible Client Hello
5340 message format would be used even with SSL_OP_NO_SSLv2.)
5343 *) Add support for smime-type MIME parameter in S/MIME messages which some
5347 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5348 a threadsafe manner. Modify rsa code to use new function and add calls
5349 to dsa and dh code (which had race conditions before).
5352 *) Include the fixed error library code in the C error file definitions
5353 instead of fixing them up at runtime. This keeps the error code
5354 structures constant.
5357 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5359 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5362 *) Fixes for newer kerberos headers. NB: the casts are needed because
5363 the 'length' field is signed on one version and unsigned on another
5364 with no (?) obvious way to tell the difference, without these VC++
5365 complains. Also the "definition" of FAR (blank) is no longer included
5366 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5367 some needed definitions.
5370 *) Undo Cygwin change.
5373 *) Added support for proxy certificates according to RFC 3820.
5374 Because they may be a security thread to unaware applications,
5375 they must be explicitly allowed in run-time. See
5376 docs/HOWTO/proxy_certificates.txt for further information.
5379 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5381 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5382 server and client random values. Previously
5383 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5384 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5386 This change has negligible security impact because:
5388 1. Server and client random values still have 24 bytes of pseudo random
5391 2. Server and client random values are sent in the clear in the initial
5394 3. The master secret is derived using the premaster secret (48 bytes in
5395 size for static RSA ciphersuites) as well as client server and random
5398 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5401 [Stephen Henson, reported by UK NISCC]
5403 *) Use Windows randomness collection on Cygwin.
5406 *) Fix hang in EGD/PRNGD query when communication socket is closed
5407 prematurely by EGD/PRNGD.
5408 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5410 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5413 *) Back-port of selected performance improvements from development
5414 branch, as well as improved support for PowerPC platforms.
5417 *) Add lots of checks for memory allocation failure, error codes to indicate
5418 failure and freeing up memory if a failure occurs.
5419 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5421 *) Add new -passin argument to dgst.
5424 *) Perform some character comparisons of different types in X509_NAME_cmp:
5425 this is needed for some certificates that reencode DNs into UTF8Strings
5426 (in violation of RFC3280) and can't or wont issue name rollover
5430 *) Make an explicit check during certificate validation to see that
5431 the CA setting in each certificate on the chain is correct. As a
5432 side effect always do the following basic checks on extensions,
5433 not just when there's an associated purpose to the check:
5435 - if there is an unhandled critical extension (unless the user
5436 has chosen to ignore this fault)
5437 - if the path length has been exceeded (if one is set at all)
5438 - that certain extensions fit the associated purpose (if one has
5442 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5444 *) Avoid a race condition when CRLs are checked in a multi threaded
5445 environment. This would happen due to the reordering of the revoked
5446 entries during signature checking and serial number lookup. Now the
5447 encoding is cached and the serial number sort performed under a lock.
5448 Add new STACK function sk_is_sorted().
5451 *) Add Delta CRL to the extension code.
5454 *) Various fixes to s3_pkt.c so alerts are sent properly.
5455 [David Holmes <d.holmes@f5.com>]
5457 *) Reduce the chances of duplicate issuer name and serial numbers (in
5458 violation of RFC3280) using the OpenSSL certificate creation utilities.
5459 This is done by creating a random 64 bit value for the initial serial
5460 number when a serial number file is created or when a self signed
5461 certificate is created using 'openssl req -x509'. The initial serial
5462 number file is created using 'openssl x509 -next_serial' in CA.pl
5463 rather than being initialized to 1.
5466 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5468 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5469 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5470 [Joe Orton, Steve Henson]
5472 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5474 [Joe Orton, Steve Henson]
5476 *) Make it possible to have multiple active certificates with the same
5477 subject in the CA index file. This is done only if the keyword
5478 'unique_subject' is set to 'no' in the main CA section (default
5479 if 'CA_default') of the configuration file. The value is saved
5480 with the database itself in a separate index attribute file,
5481 named like the index file with '.attr' appended to the name.
5484 *) X509 verify fixes. Disable broken certificate workarounds when
5485 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5486 keyUsage extension present. Don't accept CRLs with unhandled critical
5487 extensions: since verify currently doesn't process CRL extensions this
5488 rejects a CRL with *any* critical extensions. Add new verify error codes
5492 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5493 A clarification of RFC2560 will require the use of OCTET STRINGs and
5494 some implementations cannot handle the current raw format. Since OpenSSL
5495 copies and compares OCSP nonces as opaque blobs without any attempt at
5496 parsing them this should not create any compatibility issues.
5499 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5500 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5501 this HMAC (and other) operations are several times slower than OpenSSL
5505 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5506 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5508 *) Use the correct content when signing type "other".
5511 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5513 *) Fix various bugs revealed by running the NISCC test suite:
5515 Stop out of bounds reads in the ASN1 code when presented with
5516 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5518 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5520 If verify callback ignores invalid public key errors don't try to check
5521 certificate signature with the NULL public key.
5525 *) New -ignore_err option in ocsp application to stop the server
5526 exiting on the first error in a request.
5529 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5530 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5534 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5535 extra data after the compression methods not only for TLS 1.0
5536 but also for SSL 3.0 (as required by the specification).
5537 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5539 *) Change X509_certificate_type() to mark the key as exported/exportable
5540 when it's 512 *bits* long, not 512 bytes.
5543 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5544 blocks during encryption.
5547 *) Various fixes to base64 BIO and non blocking I/O. On write
5548 flushes were not handled properly if the BIO retried. On read
5549 data was not being buffered properly and had various logic bugs.
5550 This also affects blocking I/O when the data being decoded is a
5554 *) Various S/MIME bugfixes and compatibility changes:
5555 output correct application/pkcs7 MIME type if
5556 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5557 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5558 of files as .eml work). Correctly handle very long lines in MIME
5562 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5564 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5565 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5566 a protocol version number mismatch like a decryption error
5567 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5570 *) Turn on RSA blinding by default in the default implementation
5571 to avoid a timing attack. Applications that don't want it can call
5572 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5573 They would be ill-advised to do so in most cases.
5574 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5576 *) Change RSA blinding code so that it works when the PRNG is not
5577 seeded (in this case, the secret RSA exponent is abused as
5578 an unpredictable seed -- if it is not unpredictable, there
5579 is no point in blinding anyway). Make RSA blinding thread-safe
5580 by remembering the creator's thread ID in rsa->blinding and
5581 having all other threads use local one-time blinding factors
5582 (this requires more computation than sharing rsa->blinding, but
5583 avoids excessive locking; and if an RSA object is not shared
5584 between threads, blinding will still be very fast).
5587 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5588 ENGINE as defaults for all supported algorithms irrespective of
5589 the 'flags' parameter. 'flags' is now honoured, so applications
5590 should make sure they are passing it correctly.
5593 *) Target "mingw" now allows native Windows code to be generated in
5594 the Cygwin environment as well as with the MinGW compiler.
5597 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5599 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5600 via timing by performing a MAC computation even if incorrect
5601 block cipher padding has been found. This is a countermeasure
5602 against active attacks where the attacker has to distinguish
5603 between bad padding and a MAC verification error. (CVE-2003-0078)
5605 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5606 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5607 Martin Vuagnoux (EPFL, Ilion)]
5609 *) Make the no-err option work as intended. The intention with no-err
5610 is not to have the whole error stack handling routines removed from
5611 libcrypto, it's only intended to remove all the function name and
5612 reason texts, thereby removing some of the footprint that may not
5613 be interesting if those errors aren't displayed anyway.
5615 NOTE: it's still possible for any application or module to have it's
5616 own set of error texts inserted. The routines are there, just not
5617 used by default when no-err is given.
5620 *) Add support for FreeBSD on IA64.
5621 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5623 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5624 Kerberos function mit_des_cbc_cksum(). Before this change,
5625 the value returned by DES_cbc_cksum() was like the one from
5626 mit_des_cbc_cksum(), except the bytes were swapped.
5627 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5629 *) Allow an application to disable the automatic SSL chain building.
5630 Before this a rather primitive chain build was always performed in
5631 ssl3_output_cert_chain(): an application had no way to send the
5632 correct chain if the automatic operation produced an incorrect result.
5634 Now the chain builder is disabled if either:
5636 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5638 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5640 The reasoning behind this is that an application would not want the
5641 auto chain building to take place if extra chain certificates are
5642 present and it might also want a means of sending no additional
5643 certificates (for example the chain has two certificates and the
5647 *) Add the possibility to build without the ENGINE framework.
5648 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5650 *) Under Win32 gmtime() can return NULL: check return value in
5651 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5654 *) DSA routines: under certain error conditions uninitialized BN objects
5655 could be freed. Solution: make sure initialization is performed early
5656 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5657 Nils Larsch <nla@trustcenter.de> via PR#459)
5660 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5661 checked on reconnect on the client side, therefore session resumption
5662 could still fail with a "ssl session id is different" error. This
5663 behaviour is masked when SSL_OP_ALL is used due to
5664 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5665 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5666 followup to PR #377.
5669 *) IA-32 assembler support enhancements: unified ELF targets, support
5670 for SCO/Caldera platforms, fix for Cygwin shared build.
5673 *) Add support for FreeBSD on sparc64. As a consequence, support for
5674 FreeBSD on non-x86 processors is separate from x86 processors on
5675 the config script, much like the NetBSD support.
5676 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
5678 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5680 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5683 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5684 code (06) was taken as the first octet of the session ID and the last
5685 octet was ignored consequently. As a result SSLv2 client side session
5686 caching could not have worked due to the session ID mismatch between
5688 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5692 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5693 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5697 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5698 seems that in spite of existing for more than a year, many application
5699 author have done nothing to provide the necessary callbacks, which
5700 means that this particular engine will not work properly anywhere.
5701 This is a very unfortunate situation which forces us, in the name
5702 of usability, to give the hw_ncipher.c a static lock, which is part
5704 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5705 appear in 0.9.8 or later. We EXPECT application authors to have
5706 dealt properly with this when 0.9.8 is released (unless we actually
5707 make such changes in the libcrypto locking code that changes will
5708 have to be made anyway).
5711 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5712 octets have been read, EOF or an error occurs. Without this change
5713 some truncated ASN1 structures will not produce an error.
5716 *) Disable Heimdal support, since it hasn't been fully implemented.
5717 Still give the possibility to force the use of Heimdal, but with
5718 warnings and a request that patches get sent to openssl-dev.
5721 *) Add the VC-CE target, introduce the WINCE sysname, and add
5722 INSTALL.WCE and appropriate conditionals to make it build.
5723 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5725 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5726 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5727 edit numbers of the version.
5728 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5730 *) Introduce safe string copy and catenation functions
5731 (BUF_strlcpy() and BUF_strlcat()).
5732 [Ben Laurie (CHATS) and Richard Levitte]
5734 *) Avoid using fixed-size buffers for one-line DNs.
5735 [Ben Laurie (CHATS)]
5737 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5738 resizing buffers containing secrets, and use where appropriate.
5739 [Ben Laurie (CHATS)]
5741 *) Avoid using fixed size buffers for configuration file location.
5742 [Ben Laurie (CHATS)]
5744 *) Avoid filename truncation for various CA files.
5745 [Ben Laurie (CHATS)]
5747 *) Use sizeof in preference to magic numbers.
5748 [Ben Laurie (CHATS)]
5750 *) Avoid filename truncation in cert requests.
5751 [Ben Laurie (CHATS)]
5753 *) Add assertions to check for (supposedly impossible) buffer
5755 [Ben Laurie (CHATS)]
5757 *) Don't cache truncated DNS entries in the local cache (this could
5758 potentially lead to a spoofing attack).
5759 [Ben Laurie (CHATS)]
5761 *) Fix various buffers to be large enough for hex/decimal
5762 representations in a platform independent manner.
5763 [Ben Laurie (CHATS)]
5765 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5766 resizing buffers containing secrets, and use where appropriate.
5767 [Ben Laurie (CHATS)]
5769 *) Add BIO_indent() to avoid much slightly worrying code to do
5771 [Ben Laurie (CHATS)]
5773 *) Convert sprintf()/BIO_puts() to BIO_printf().
5774 [Ben Laurie (CHATS)]
5776 *) buffer_gets() could terminate with the buffer only half
5778 [Ben Laurie (CHATS)]
5780 *) Add assertions to prevent user-supplied crypto functions from
5781 overflowing internal buffers by having large block sizes, etc.
5782 [Ben Laurie (CHATS)]
5784 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5786 [Ben Laurie (CHATS)]
5788 *) Eliminate unused copy of key in RC4.
5789 [Ben Laurie (CHATS)]
5791 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5792 [Ben Laurie (CHATS)]
5794 *) Fix off-by-one error in EGD path.
5795 [Ben Laurie (CHATS)]
5797 *) If RANDFILE path is too long, ignore instead of truncating.
5798 [Ben Laurie (CHATS)]
5800 *) Eliminate unused and incorrectly sized X.509 structure
5802 [Ben Laurie (CHATS)]
5804 *) Eliminate unused and dangerous function knumber().
5805 [Ben Laurie (CHATS)]
5807 *) Eliminate unused and dangerous structure, KSSL_ERR.
5808 [Ben Laurie (CHATS)]
5810 *) Protect against overlong session ID context length in an encoded
5811 session object. Since these are local, this does not appear to be
5813 [Ben Laurie (CHATS)]
5815 *) Change from security patch (see 0.9.6e below) that did not affect
5816 the 0.9.6 release series:
5818 Remote buffer overflow in SSL3 protocol - an attacker could
5819 supply an oversized master key in Kerberos-enabled versions.
5821 [Ben Laurie (CHATS)]
5823 *) Change the SSL kerb5 codes to match RFC 2712.
5826 *) Make -nameopt work fully for req and add -reqopt switch.
5827 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5829 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5830 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5832 *) Make sure tests can be performed even if the corresponding algorithms
5833 have been removed entirely. This was also the last step to make
5834 OpenSSL compilable with DJGPP under all reasonable conditions.
5835 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5837 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5838 to allow version independent disabling of normally unselected ciphers,
5839 which may be activated as a side-effect of selecting a single cipher.
5841 (E.g., cipher list string "RSA" enables ciphersuites that are left
5842 out of "ALL" because they do not provide symmetric encryption.
5843 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5844 [Lutz Jaenicke, Bodo Moeller]
5846 *) Add appropriate support for separate platform-dependent build
5847 directories. The recommended way to make a platform-dependent
5848 build directory is the following (tested on Linux), maybe with
5851 # Place yourself outside of the OpenSSL source tree. In
5852 # this example, the environment variable OPENSSL_SOURCE
5853 # is assumed to contain the absolute OpenSSL source directory.
5854 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5855 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5856 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5857 mkdir -p `dirname $F`
5858 ln -s $OPENSSL_SOURCE/$F $F
5861 To be absolutely sure not to disturb the source tree, a "make clean"
5862 is a good thing. If it isn't successful, don't worry about it,
5863 it probably means the source directory is very clean.
5866 *) Make sure any ENGINE control commands make local copies of string
5867 pointers passed to them whenever necessary. Otherwise it is possible
5868 the caller may have overwritten (or deallocated) the original string
5869 data when a later ENGINE operation tries to use the stored values.
5870 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5872 *) Improve diagnostics in file reading and command-line digests.
5873 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5875 *) Add AES modes CFB and OFB to the object database. Correct an
5876 error in AES-CFB decryption.
5879 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5880 allows existing EVP_CIPHER_CTX structures to be reused after
5881 calling EVP_*Final(). This behaviour is used by encryption
5882 BIOs and some applications. This has the side effect that
5883 applications must explicitly clean up cipher contexts with
5884 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5887 *) Check the values of dna and dnb in bn_mul_recursive before calling
5888 bn_mul_comba (a non zero value means the a or b arrays do not contain
5889 n2 elements) and fallback to bn_mul_normal if either is not zero.
5892 *) Fix escaping of non-ASCII characters when using the -subj option
5893 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5896 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5897 form for "surname", serialNumber has no short form.
5898 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5899 therefore remove "mail" short name for "internet 7".
5900 The OID for unique identifiers in X509 certificates is
5901 x500UniqueIdentifier, not uniqueIdentifier.
5902 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5905 *) Add an "init" command to the ENGINE config module and auto initialize
5906 ENGINEs. Without any "init" command the ENGINE will be initialized
5907 after all ctrl commands have been executed on it. If init=1 the
5908 ENGINE is initailized at that point (ctrls before that point are run
5909 on the uninitialized ENGINE and after on the initialized one). If
5910 init=0 then the ENGINE will not be iniatialized at all.
5913 *) Fix the 'app_verify_callback' interface so that the user-defined
5914 argument is actually passed to the callback: In the
5915 SSL_CTX_set_cert_verify_callback() prototype, the callback
5916 declaration has been changed from
5919 int (*cb)(X509_STORE_CTX *,void *);
5920 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5921 i=s->ctx->app_verify_callback(&ctx)
5922 has been changed into
5923 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5925 To update applications using SSL_CTX_set_cert_verify_callback(),
5926 a dummy argument can be added to their callback functions.
5927 [D. K. Smetters <smetters@parc.xerox.com>]
5929 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5930 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5932 *) Add and OPENSSL_LOAD_CONF define which will cause
5933 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5934 This allows older applications to transparently support certain
5935 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5936 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5937 load the config file and OPENSSL_add_all_algorithms_conf() which will
5938 always load it have also been added.
5941 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5942 Adjust NIDs and EVP layer.
5943 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5945 *) Config modules support in openssl utility.
5947 Most commands now load modules from the config file,
5948 though in a few (such as version) this isn't done
5949 because it couldn't be used for anything.
5951 In the case of ca and req the config file used is
5952 the same as the utility itself: that is the -config
5953 command line option can be used to specify an
5957 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5958 use "openssl_conf" if filename is NULL use default openssl config file.
5961 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5962 config section name. Add a new flag to tolerate a missing config file
5963 and move code to CONF_modules_load_file().
5966 *) Support for crypto accelerator cards from Accelerated Encryption
5967 Processing, www.aep.ie. (Use engine 'aep')
5968 The support was copied from 0.9.6c [engine] and adapted/corrected
5969 to work with the new engine framework.
5970 [AEP Inc. and Richard Levitte]
5972 *) Support for SureWare crypto accelerator cards from Baltimore
5973 Technologies. (Use engine 'sureware')
5974 The support was copied from 0.9.6c [engine] and adapted
5975 to work with the new engine framework.
5978 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5979 make the newer ENGINE framework commands for the CHIL engine work.
5980 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5982 *) Make it possible to produce shared libraries on ReliantUNIX.
5983 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5985 *) Add the configuration target debug-linux-ppro.
5986 Make 'openssl rsa' use the general key loading routines
5987 implemented in apps.c, and make those routines able to
5988 handle the key format FORMAT_NETSCAPE and the variant
5990 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5992 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5993 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5995 *) Add -keyform to rsautl, and document -engine.
5996 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5998 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5999 BIO_R_NO_SUCH_FILE error code rather than the generic
6000 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6003 *) Add new functions
6005 ERR_peek_last_error_line
6006 ERR_peek_last_error_line_data.
6007 These are similar to
6010 ERR_peek_error_line_data,
6011 but report on the latest error recorded rather than the first one
6012 still in the error queue.
6013 [Ben Laurie, Bodo Moeller]
6015 *) default_algorithms option in ENGINE config module. This allows things
6017 default_algorithms = ALL
6018 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6021 *) Preliminary ENGINE config module.
6024 *) New experimental application configuration code.
6027 *) Change the AES code to follow the same name structure as all other
6028 symmetric ciphers, and behave the same way. Move everything to
6029 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6030 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6032 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6033 [Ben Laurie and Theo de Raadt]
6035 *) Add option to output public keys in req command.
6036 [Massimiliano Pala madwolf@openca.org]
6038 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6039 (up to about 10% better than before for P-192 and P-224).
6042 *) New functions/macros
6044 SSL_CTX_set_msg_callback(ctx, cb)
6045 SSL_CTX_set_msg_callback_arg(ctx, arg)
6046 SSL_set_msg_callback(ssl, cb)
6047 SSL_set_msg_callback_arg(ssl, arg)
6049 to request calling a callback function
6051 void cb(int write_p, int version, int content_type,
6052 const void *buf, size_t len, SSL *ssl, void *arg)
6054 whenever a protocol message has been completely received
6055 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6056 protocol version according to which the SSL library interprets
6057 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6058 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6059 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6060 specification (change_cipher_spec(20), alert(21), handshake(22)).
6061 'buf' and 'len' point to the actual message, 'ssl' to the
6062 SSL object, and 'arg' is the application-defined value set by
6063 SSL[_CTX]_set_msg_callback_arg().
6065 'openssl s_client' and 'openssl s_server' have new '-msg' options
6066 to enable a callback that displays all protocol messages.
6069 *) Change the shared library support so shared libraries are built as
6070 soon as the corresponding static library is finished, and thereby get
6071 openssl and the test programs linked against the shared library.
6072 This still only happens when the keyword "shard" has been given to
6073 the configuration scripts.
6075 NOTE: shared library support is still an experimental thing, and
6076 backward binary compatibility is still not guaranteed.
6077 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6079 *) Add support for Subject Information Access extension.
6080 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6082 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6083 additional bytes when new memory had to be allocated, not just
6084 when reusing an existing buffer.
6087 *) New command line and configuration option 'utf8' for the req command.
6088 This allows field values to be specified as UTF8 strings.
6091 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6092 runs for the former and machine-readable output for the latter.
6095 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6096 of the e-mail address in the DN (i.e., it will go into a certificate
6097 extension only). The new configuration file option 'email_in_dn = no'
6098 has the same effect.
6099 [Massimiliano Pala madwolf@openca.org]
6101 *) Change all functions with names starting with des_ to be starting
6102 with DES_ instead. Add wrappers that are compatible with libdes,
6103 but are named _ossl_old_des_*. Finally, add macros that map the
6104 des_* symbols to the corresponding _ossl_old_des_* if libdes
6105 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6106 desired, the des_* symbols will be mapped to DES_*, with one
6109 Since we provide two compatibility mappings, the user needs to
6110 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6111 compatibility is desired. The default (i.e., when that macro
6112 isn't defined) is OpenSSL 0.9.6c compatibility.
6114 There are also macros that enable and disable the support of old
6115 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6116 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6117 are defined, the default will apply: to support the old des routines.
6119 In either case, one must include openssl/des.h to get the correct
6120 definitions. Do not try to just include openssl/des_old.h, that
6123 NOTE: This is a major break of an old API into a new one. Software
6124 authors are encouraged to switch to the DES_ style functions. Some
6125 time in the future, des_old.h and the libdes compatibility functions
6126 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6127 default), and then completely removed.
6130 *) Test for certificates which contain unsupported critical extensions.
6131 If such a certificate is found during a verify operation it is
6132 rejected by default: this behaviour can be overridden by either
6133 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6134 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6135 X509_supported_extension() has also been added which returns 1 if a
6136 particular extension is supported.
6139 *) Modify the behaviour of EVP cipher functions in similar way to digests
6140 to retain compatibility with existing code.
6143 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6144 compatibility with existing code. In particular the 'ctx' parameter does
6145 not have to be to be initialized before the call to EVP_DigestInit() and
6146 it is tidied up after a call to EVP_DigestFinal(). New function
6147 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6148 EVP_MD_CTX_copy() changed to not require the destination to be
6149 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6150 requires the destination to be valid.
6152 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6153 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6156 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6157 so that complete 'Handshake' protocol structures are kept in memory
6158 instead of overwriting 'msg_type' and 'length' with 'body' data.
6161 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6162 [Massimo Santin via Richard Levitte]
6164 *) Major restructuring to the underlying ENGINE code. This includes
6165 reduction of linker bloat, separation of pure "ENGINE" manipulation
6166 (initialisation, etc) from functionality dealing with implementations
6167 of specific crypto iterfaces. This change also introduces integrated
6168 support for symmetric ciphers and digest implementations - so ENGINEs
6169 can now accelerate these by providing EVP_CIPHER and EVP_MD
6170 implementations of their own. This is detailed in crypto/engine/README
6171 as it couldn't be adequately described here. However, there are a few
6172 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6173 were changed in the original introduction of ENGINE code have now
6174 reverted back - the hooking from this code to ENGINE is now a good
6175 deal more passive and at run-time, operations deal directly with
6176 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6177 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6178 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6179 they were not being used by the framework as there is no concept of a
6180 BIGNUM_METHOD and they could not be generalised to the new
6181 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6182 ENGINE_cpy() has been removed as it cannot be consistently defined in
6186 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6189 *) Change mkdef.pl to sort symbols that get the same entry number,
6190 and make sure the automatically generated functions ERR_load_*
6191 become part of libeay.num as well.
6194 *) New function SSL_renegotiate_pending(). This returns true once
6195 renegotiation has been requested (either SSL_renegotiate() call
6196 or HelloRequest/ClientHello received from the peer) and becomes
6197 false once a handshake has been completed.
6198 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6199 sends a HelloRequest, but does not ensure that a handshake takes
6200 place. SSL_renegotiate_pending() is useful for checking if the
6201 client has followed the request.)
6204 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6205 By default, clients may request session resumption even during
6206 renegotiation (if session ID contexts permit); with this option,
6207 session resumption is possible only in the first handshake.
6209 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6210 more bits available for options that should not be part of
6211 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6214 *) Add some demos for certificate and certificate request creation.
6217 *) Make maximum certificate chain size accepted from the peer application
6218 settable (SSL*_get/set_max_cert_list()), as proposed by
6219 "Douglas E. Engert" <deengert@anl.gov>.
6222 *) Add support for shared libraries for Unixware-7
6223 (Boyd Lynn Gerber <gerberb@zenez.com>).
6226 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6227 be done prior to destruction. Use this to unload error strings from
6228 ENGINEs that load their own error strings. NB: This adds two new API
6229 functions to "get" and "set" this destroy handler in an ENGINE.
6232 *) Alter all existing ENGINE implementations (except "openssl" and
6233 "openbsd") to dynamically instantiate their own error strings. This
6234 makes them more flexible to be built both as statically-linked ENGINEs
6235 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6236 Also, add stub code to each that makes building them as self-contained
6237 shared-libraries easier (see README.ENGINE).
6240 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6241 implementations into applications that are completely implemented in
6242 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6243 commands that can be used to configure what shared-library to load and
6244 to control aspects of the way it is handled. Also, made an update to
6245 the README.ENGINE file that brings its information up-to-date and
6246 provides some information and instructions on the "dynamic" ENGINE
6247 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6250 *) Make it possible to unload ranges of ERR strings with a new
6251 "ERR_unload_strings" function.
6254 *) Add a copy() function to EVP_MD.
6257 *) Make EVP_MD routines take a context pointer instead of just the
6258 md_data void pointer.
6261 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6262 that the digest can only process a single chunk of data
6263 (typically because it is provided by a piece of
6264 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6265 is only going to provide a single chunk of data, and hence the
6266 framework needn't accumulate the data for oneshot drivers.
6269 *) As with "ERR", make it possible to replace the underlying "ex_data"
6270 functions. This change also alters the storage and management of global
6271 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6272 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6273 index counters. The API functions that use this state have been changed
6274 to take a "class_index" rather than pointers to the class's local STACK
6275 and counter, and there is now an API function to dynamically create new
6276 classes. This centralisation allows us to (a) plug a lot of the
6277 thread-safety problems that existed, and (b) makes it possible to clean
6278 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6279 such data would previously have always leaked in application code and
6280 workarounds were in place to make the memory debugging turn a blind eye
6281 to it. Application code that doesn't use this new function will still
6282 leak as before, but their memory debugging output will announce it now
6283 rather than letting it slide.
6285 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6286 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6287 has a return value to indicate success or failure.
6290 *) Make it possible to replace the underlying "ERR" functions such that the
6291 global state (2 LHASH tables and 2 locks) is only used by the "default"
6292 implementation. This change also adds two functions to "get" and "set"
6293 the implementation prior to it being automatically set the first time
6294 any other ERR function takes place. Ie. an application can call "get",
6295 pass the return value to a module it has just loaded, and that module
6296 can call its own "set" function using that value. This means the
6297 module's "ERR" operations will use (and modify) the error state in the
6298 application and not in its own statically linked copy of OpenSSL code.
6301 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6302 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6303 the operation, and provides a more encapsulated way for external code
6304 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6305 to use these functions rather than manually incrementing the counts.
6307 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6310 *) Add EVP test program.
6313 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6316 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6317 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6318 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6319 These allow a CRL to be built without having to access X509_CRL fields
6320 directly. Modify 'ca' application to use new functions.
6323 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6324 bug workarounds. Rollback attack detection is a security feature.
6325 The problem will only arise on OpenSSL servers when TLSv1 is not
6326 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6327 Software authors not wanting to support TLSv1 will have special reasons
6328 for their choice and can explicitly enable this option.
6329 [Bodo Moeller, Lutz Jaenicke]
6331 *) Rationalise EVP so it can be extended: don't include a union of
6332 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6333 (similar to those existing for EVP_CIPHER_CTX).
6338 EVP_MD_CTX_init(&md); /* new function call */
6339 EVP_DigestInit(&md, EVP_sha1());
6340 EVP_DigestUpdate(&md, in, len);
6341 EVP_DigestFinal(&md, out, NULL);
6342 EVP_MD_CTX_cleanup(&md); /* new function call */
6346 *) Make DES key schedule conform to the usual scheme, as well as
6347 correcting its structure. This means that calls to DES functions
6348 now have to pass a pointer to a des_key_schedule instead of a
6349 plain des_key_schedule (which was actually always a pointer
6352 des_key_schedule ks;
6354 des_set_key_checked(..., &ks);
6355 des_ncbc_encrypt(..., &ks, ...);
6357 (Note that a later change renames 'des_...' into 'DES_...'.)
6360 *) Initial reduction of linker bloat: the use of some functions, such as
6361 PEM causes large amounts of unused functions to be linked in due to
6362 poor organisation. For example pem_all.c contains every PEM function
6363 which has a knock on effect of linking in large amounts of (unused)
6364 ASN1 code. Grouping together similar functions and splitting unrelated
6365 functions prevents this.
6368 *) Cleanup of EVP macros.
6371 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6372 correct _ecb suffix.
6375 *) Add initial OCSP responder support to ocsp application. The
6376 revocation information is handled using the text based index
6377 use by the ca application. The responder can either handle
6378 requests generated internally, supplied in files (for example
6379 via a CGI script) or using an internal minimal server.
6382 *) Add configuration choices to get zlib compression for TLS.
6385 *) Changes to Kerberos SSL for RFC 2712 compliance:
6386 1. Implemented real KerberosWrapper, instead of just using
6387 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6388 2. Implemented optional authenticator field of KerberosWrapper.
6390 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6391 and authenticator structs; see crypto/krb5/.
6393 Generalized Kerberos calls to support multiple Kerberos libraries.
6394 [Vern Staats <staatsvr@asc.hpc.mil>,
6395 Jeffrey Altman <jaltman@columbia.edu>
6396 via Richard Levitte]
6398 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6399 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6400 values for each of the key sizes rather than having just
6401 parameters (and 'speed' generating keys each time).
6404 *) Speed up EVP routines.
6407 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6408 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6409 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6410 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6412 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6413 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6414 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6417 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6419 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6422 *) Added the OS2-EMX target.
6423 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6425 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6426 to support NCONF routines in extension code. New function CONF_set_nconf()
6427 to allow functions which take an NCONF to also handle the old LHASH
6428 structure: this means that the old CONF compatible routines can be
6429 retained (in particular wrt extensions) without having to duplicate the
6430 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6433 *) Enhance the general user interface with mechanisms for inner control
6434 and with possibilities to have yes/no kind of prompts.
6437 *) Change all calls to low level digest routines in the library and
6438 applications to use EVP. Add missing calls to HMAC_cleanup() and
6439 don't assume HMAC_CTX can be copied using memcpy().
6440 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6442 *) Add the possibility to control engines through control names but with
6443 arbitrary arguments instead of just a string.
6444 Change the key loaders to take a UI_METHOD instead of a callback
6445 function pointer. NOTE: this breaks binary compatibility with earlier
6446 versions of OpenSSL [engine].
6447 Adapt the nCipher code for these new conditions and add a card insertion
6451 *) Enhance the general user interface with mechanisms to better support
6452 dialog box interfaces, application-defined prompts, the possibility
6453 to use defaults (for example default passwords from somewhere else)
6454 and interrupts/cancellations.
6457 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6458 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6461 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6462 tidy up some unnecessarily weird code in 'sk_new()').
6463 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6465 *) Change the key loading routines for ENGINEs to use the same kind
6466 callback (pem_password_cb) as all other routines that need this
6470 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6471 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6472 than this minimum value is recommended.
6475 *) New random seeder for OpenVMS, using the system process statistics
6476 that are easily reachable.
6479 *) Windows apparently can't transparently handle global
6480 variables defined in DLLs. Initialisations such as:
6482 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6484 wont compile. This is used by the any applications that need to
6485 declare their own ASN1 modules. This was fixed by adding the option
6486 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6487 needed for static libraries under Win32.
6490 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6491 setting of purpose and trust fields. New X509_STORE trust and
6492 purpose functions and tidy up setting in other SSL functions.
6495 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6496 structure. These are inherited by X509_STORE_CTX when it is
6497 initialised. This allows various defaults to be set in the
6498 X509_STORE structure (such as flags for CRL checking and custom
6499 purpose or trust settings) for functions which only use X509_STORE_CTX
6500 internally such as S/MIME.
6502 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6503 trust settings if they are not set in X509_STORE. This allows X509_STORE
6504 purposes and trust (in S/MIME for example) to override any set by default.
6506 Add command line options for CRL checking to smime, s_client and s_server
6510 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6511 are set then the CRL is looked up in the X509_STORE structure and
6512 its validity and signature checked, then if the certificate is found
6513 in the CRL the verify fails with a revoked error.
6515 Various new CRL related callbacks added to X509_STORE_CTX structure.
6517 Command line options added to 'verify' application to support this.
6519 This needs some additional work, such as being able to handle multiple
6520 CRLs with different times, extension based lookup (rather than just
6521 by subject name) and ultimately more complete V2 CRL extension
6525 *) Add a general user interface API (crypto/ui/). This is designed
6526 to replace things like des_read_password and friends (backward
6527 compatibility functions using this new API are provided).
6528 The purpose is to remove prompting functions from the DES code
6529 section as well as provide for prompting through dialog boxes in
6530 a window system and the like.
6533 *) Add "ex_data" support to ENGINE so implementations can add state at a
6534 per-structure level rather than having to store it globally.
6537 *) Make it possible for ENGINE structures to be copied when retrieved by
6538 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6539 This causes the "original" ENGINE structure to act like a template,
6540 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6541 operational state can be localised to each ENGINE structure, despite the
6542 fact they all share the same "methods". New ENGINE structures returned in
6543 this case have no functional references and the return value is the single
6544 structural reference. This matches the single structural reference returned
6545 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6549 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6550 needs to match any other type at all we need to manually clear the
6554 *) Changes to the "openssl engine" utility to include;
6555 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6556 about an ENGINE's available control commands.
6557 - executing control commands from command line arguments using the
6558 '-pre' and '-post' switches. '-post' is only used if '-t' is
6559 specified and the ENGINE is successfully initialised. The syntax for
6560 the individual commands are colon-separated, for example;
6561 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6564 *) New dynamic control command support for ENGINEs. ENGINEs can now
6565 declare their own commands (numbers), names (strings), descriptions,
6566 and input types for run-time discovery by calling applications. A
6567 subset of these commands are implicitly classed as "executable"
6568 depending on their input type, and only these can be invoked through
6569 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6570 can be based on user input, config files, etc). The distinction is
6571 that "executable" commands cannot return anything other than a boolean
6572 result and can only support numeric or string input, whereas some
6573 discoverable commands may only be for direct use through
6574 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6575 pointers, or other custom uses. The "executable" commands are to
6576 support parameterisations of ENGINE behaviour that can be
6577 unambiguously defined by ENGINEs and used consistently across any
6578 OpenSSL-based application. Commands have been added to all the
6579 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6580 control over shared-library paths without source code alterations.
6583 *) Changed all ENGINE implementations to dynamically allocate their
6584 ENGINEs rather than declaring them statically. Apart from this being
6585 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6586 this also allows the implementations to compile without using the
6587 internal engine_int.h header.
6590 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6591 'const' value. Any code that should be able to modify a RAND_METHOD
6592 should already have non-const pointers to it (ie. they should only
6593 modify their own ones).
6596 *) Made a variety of little tweaks to the ENGINE code.
6597 - "atalla" and "ubsec" string definitions were moved from header files
6598 to C code. "nuron" string definitions were placed in variables
6599 rather than hard-coded - allowing parameterisation of these values
6600 later on via ctrl() commands.
6601 - Removed unused "#if 0"'d code.
6602 - Fixed engine list iteration code so it uses ENGINE_free() to release
6603 structural references.
6604 - Constified the RAND_METHOD element of ENGINE structures.
6605 - Constified various get/set functions as appropriate and added
6606 missing functions (including a catch-all ENGINE_cpy that duplicates
6607 all ENGINE values onto a new ENGINE except reference counts/state).
6608 - Removed NULL parameter checks in get/set functions. Setting a method
6609 or function to NULL is a way of cancelling out a previously set
6610 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6611 and doesn't justify the extra error symbols and code.
6612 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6613 flags from engine_int.h to engine.h.
6614 - Changed prototypes for ENGINE handler functions (init(), finish(),
6615 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6618 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6619 to the algorithm using long division. The binary algorithm can be
6620 used only if the modulus is odd. On 32-bit systems, it is faster
6621 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6622 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6623 up to 450 bits. In 64-bit environments, the binary algorithm
6624 appears to be advantageous for much longer moduli; here we use it
6625 for moduli up to 2048 bits.
6628 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6629 could not support the combine flag in choice fields.
6632 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6633 extensions from a certificate request to the certificate.
6636 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6637 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6638 file: this allows the display of the certificate about to be
6639 signed to be customised, to allow certain fields to be included
6640 or excluded and extension details. The old system didn't display
6641 multicharacter strings properly, omitted fields not in the policy
6642 and couldn't display additional details such as extensions.
6645 *) Function EC_POINTs_mul for multiple scalar multiplication
6646 of an arbitrary number of elliptic curve points
6647 \sum scalars[i]*points[i],
6648 optionally including the generator defined for the EC_GROUP:
6649 scalar*generator + \sum scalars[i]*points[i].
6651 EC_POINT_mul is a simple wrapper function for the typical case
6652 that the point list has just one item (besides the optional
6656 *) First EC_METHODs for curves over GF(p):
6658 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6659 operations and provides various method functions that can also
6660 operate with faster implementations of modular arithmetic.
6662 EC_GFp_mont_method() reuses most functions that are part of
6663 EC_GFp_simple_method, but uses Montgomery arithmetic.
6665 [Bodo Moeller; point addition and point doubling
6666 implementation directly derived from source code provided by
6667 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6669 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6670 crypto/ec/ec_lib.c):
6672 Curves are EC_GROUP objects (with an optional group generator)
6673 based on EC_METHODs that are built into the library.
6675 Points are EC_POINT objects based on EC_GROUP objects.
6677 Most of the framework would be able to handle curves over arbitrary
6678 finite fields, but as there are no obvious types for fields other
6679 than GF(p), some functions are limited to that for now.
6682 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6683 that the file contains a complete HTTP response.
6686 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6687 change the def and num file printf format specifier from "%-40sXXX"
6688 to "%-39s XXX". The latter will always guarantee a space after the
6689 field while the former will cause them to run together if the field
6690 is 40 of more characters long.
6693 *) Constify the cipher and digest 'method' functions and structures
6694 and modify related functions to take constant EVP_MD and EVP_CIPHER
6698 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6699 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6702 *) Modify EVP_Digest*() routines so they now return values. Although the
6703 internal software routines can never fail additional hardware versions
6707 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6709 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6710 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6713 ERR_R_NESTED_ASN1_ERROR
6715 ERR_R_MISSING_ASN1_EOS
6716 were 4 .. 9, conflicting with
6717 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6719 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6720 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6722 Add new error code 'ERR_R_INTERNAL_ERROR'.
6725 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6729 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6730 sets the subject name for a new request or supersedes the
6731 subject name in a given request. Formats that can be parsed are
6732 'CN=Some Name, OU=myOU, C=IT'
6734 'CN=Some Name/OU=myOU/C=IT'.
6736 Add options '-batch' and '-verbose' to 'openssl req'.
6737 [Massimiliano Pala <madwolf@hackmasters.net>]
6739 *) Introduce the possibility to access global variables through
6740 functions on platform were that's the best way to handle exporting
6741 global variables in shared libraries. To enable this functionality,
6742 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6743 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6744 is normally done by Configure or something similar).
6746 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6747 in the source file (foo.c) like this:
6749 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6750 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6752 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6753 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6755 OPENSSL_DECLARE_GLOBAL(int,foo);
6756 #define foo OPENSSL_GLOBAL_REF(foo)
6757 OPENSSL_DECLARE_GLOBAL(double,bar);
6758 #define bar OPENSSL_GLOBAL_REF(bar)
6760 The #defines are very important, and therefore so is including the
6761 header file everywhere where the defined globals are used.
6763 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6764 of ASN.1 items, but that structure is a bit different.
6766 The largest change is in util/mkdef.pl which has been enhanced with
6767 better and easier to understand logic to choose which symbols should
6768 go into the Windows .def files as well as a number of fixes and code
6769 cleanup (among others, algorithm keywords are now sorted
6770 lexicographically to avoid constant rewrites).
6773 *) In BN_div() keep a copy of the sign of 'num' before writing the
6774 result to 'rm' because if rm==num the value will be overwritten
6775 and produce the wrong result if 'num' is negative: this caused
6776 problems with BN_mod() and BN_nnmod().
6779 *) Function OCSP_request_verify(). This checks the signature on an
6780 OCSP request and verifies the signer certificate. The signer
6781 certificate is just checked for a generic purpose and OCSP request
6785 *) Add OCSP_check_validity() function to check the validity of OCSP
6786 responses. OCSP responses are prepared in real time and may only
6787 be a few seconds old. Simply checking that the current time lies
6788 between thisUpdate and nextUpdate max reject otherwise valid responses
6789 caused by either OCSP responder or client clock inaccuracy. Instead
6790 we allow thisUpdate and nextUpdate to fall within a certain period of
6791 the current time. The age of the response can also optionally be
6792 checked. Two new options -validity_period and -status_age added to
6796 *) If signature or public key algorithm is unrecognized print out its
6797 OID rather that just UNKNOWN.
6800 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6801 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6802 ID to be generated from the issuer certificate alone which can then be
6803 passed to OCSP_id_issuer_cmp().
6806 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6807 ASN1 modules to export functions returning ASN1_ITEM pointers
6808 instead of the ASN1_ITEM structures themselves. This adds several
6809 new macros which allow the underlying ASN1 function/structure to
6810 be accessed transparently. As a result code should not use ASN1_ITEM
6811 references directly (such as &X509_it) but instead use the relevant
6812 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6813 use of the new ASN1 code on platforms where exporting structures
6814 is problematical (for example in shared libraries) but exporting
6815 functions returning pointers to structures is not.
6818 *) Add support for overriding the generation of SSL/TLS session IDs.
6819 These callbacks can be registered either in an SSL_CTX or per SSL.
6820 The purpose of this is to allow applications to control, if they wish,
6821 the arbitrary values chosen for use as session IDs, particularly as it
6822 can be useful for session caching in multiple-server environments. A
6823 command-line switch for testing this (and any client code that wishes
6824 to use such a feature) has been added to "s_server".
6825 [Geoff Thorpe, Lutz Jaenicke]
6827 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6828 of the form '#if defined(...) || defined(...) || ...' and
6829 '#if !defined(...) && !defined(...) && ...'. This also avoids
6830 the growing number of special cases it was previously handling.
6833 *) Make all configuration macros available for application by making
6834 sure they are available in opensslconf.h, by giving them names starting
6835 with "OPENSSL_" to avoid conflicts with other packages and by making
6836 sure e_os2.h will cover all platform-specific cases together with
6838 Additionally, it is now possible to define configuration/platform-
6839 specific names (called "system identities"). In the C code, these
6840 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6841 macro with the name beginning with "OPENSSL_SYS_", which is determined
6842 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6846 *) New option -set_serial to 'req' and 'x509' this allows the serial
6847 number to use to be specified on the command line. Previously self
6848 signed certificates were hard coded with serial number 0 and the
6849 CA options of 'x509' had to use a serial number in a file which was
6853 *) New options to 'ca' utility to support V2 CRL entry extensions.
6854 Currently CRL reason, invalidity date and hold instruction are
6855 supported. Add new CRL extensions to V3 code and some new objects.
6858 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6859 disable standard block padding (aka PKCS#5 padding) in the EVP
6860 API, which was previously mandatory. This means that the data is
6861 not padded in any way and so the total length much be a multiple
6862 of the block size, otherwise an error occurs.
6865 *) Initial (incomplete) OCSP SSL support.
6868 *) New function OCSP_parse_url(). This splits up a URL into its host,
6869 port and path components: primarily to parse OCSP URLs. New -url
6870 option to ocsp utility.
6873 *) New nonce behavior. The return value of OCSP_check_nonce() now
6874 reflects the various checks performed. Applications can decide
6875 whether to tolerate certain situations such as an absent nonce
6876 in a response when one was present in a request: the ocsp application
6877 just prints out a warning. New function OCSP_add1_basic_nonce()
6878 this is to allow responders to include a nonce in a response even if
6879 the request is nonce-less.
6882 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6883 skipped when using openssl x509 multiple times on a single input file,
6884 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6887 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6888 set string type: to handle setting ASN1_TIME structures. Fix ca
6889 utility to correctly initialize revocation date of CRLs.
6892 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6893 the clients preferred ciphersuites and rather use its own preferences.
6894 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6895 Internet Explorer by ensuring unchanged hash method during stepup.
6896 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6899 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6900 to aes and add a new 'exist' option to print out symbols that don't
6904 *) Additional options to ocsp utility to allow flags to be set and
6905 additional certificates supplied.
6908 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6909 OCSP client a number of certificate to only verify the response
6913 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6914 handle the new API. Currently only ECB, CBC modes supported. Add new
6917 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6918 Encryption Standard (AES) Ciphersuites for Transport Layer
6919 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6920 not enabled by default and were not part of the "ALL" ciphersuite
6921 alias because they were not yet official; they could be
6922 explicitly requested by specifying the "AESdraft" ciphersuite
6923 group alias. In the final release of OpenSSL 0.9.7, the group
6924 alias is called "AES" and is part of "ALL".)
6925 [Ben Laurie, Steve Henson, Bodo Moeller]
6927 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6928 request to response.
6931 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6932 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6933 extract information from a certificate request. OCSP_response_create()
6934 creates a response and optionally adds a basic response structure.
6935 OCSP_basic_add1_status() adds a complete single response to a basic
6936 response and returns the OCSP_SINGLERESP structure just added (to allow
6937 extensions to be included for example). OCSP_basic_add1_cert() adds a
6938 certificate to a basic response and OCSP_basic_sign() signs a basic
6939 response with various flags. New helper functions ASN1_TIME_check()
6940 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6941 (converts ASN1_TIME to GeneralizedTime).
6944 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6945 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6946 structure from a certificate. X509_pubkey_digest() digests the public_key
6947 contents: this is used in various key identifiers.
6950 *) Make sk_sort() tolerate a NULL argument.
6951 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6953 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6954 passed by the function are trusted implicitly. If any of them signed the
6955 response then it is assumed to be valid and is not verified.
6958 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6959 to data. This was previously part of the PKCS7 ASN1 code. This
6960 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6961 [Steve Henson, reported by Kenneth R. Robinette
6962 <support@securenetterm.com>]
6964 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6965 routines: without these tracing memory leaks is very painful.
6966 Fix leaks in PKCS12 and PKCS7 routines.
6969 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6970 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6971 effectively meant GeneralizedTime would never be used. Now it
6972 is initialised to -1 but X509_time_adj() now has to check the value
6973 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6974 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6975 [Steve Henson, reported by Kenneth R. Robinette
6976 <support@securenetterm.com>]
6978 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6979 result in a zero length in the ASN1_INTEGER structure which was
6980 not consistent with the structure when d2i_ASN1_INTEGER() was used
6981 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6982 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6983 where it did not print out a minus for negative ASN1_INTEGER.
6986 *) Add summary printout to ocsp utility. The various functions which
6987 convert status values to strings have been renamed to:
6988 OCSP_response_status_str(), OCSP_cert_status_str() and
6989 OCSP_crl_reason_str() and are no longer static. New options
6990 to verify nonce values and to disable verification. OCSP response
6991 printout format cleaned up.
6994 *) Add additional OCSP certificate checks. These are those specified
6995 in RFC2560. This consists of two separate checks: the CA of the
6996 certificate being checked must either be the OCSP signer certificate
6997 or the issuer of the OCSP signer certificate. In the latter case the
6998 OCSP signer certificate must contain the OCSP signing extended key
6999 usage. This check is performed by attempting to match the OCSP
7000 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7001 in the OCSP_CERTID structures of the response.
7004 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7005 and related routines. This uses the standard OpenSSL certificate
7006 verify routines to perform initial checks (just CA validity) and
7007 to obtain the certificate chain. Then additional checks will be
7008 performed on the chain. Currently the root CA is checked to see
7009 if it is explicitly trusted for OCSP signing. This is used to set
7010 a root CA as a global signing root: that is any certificate that
7011 chains to that CA is an acceptable OCSP signing certificate.
7014 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7015 extensions from a separate configuration file.
7016 As when reading extensions from the main configuration file,
7017 the '-extensions ...' option may be used for specifying the
7019 [Massimiliano Pala <madwolf@comune.modena.it>]
7021 *) New OCSP utility. Allows OCSP requests to be generated or
7022 read. The request can be sent to a responder and the output
7023 parsed, outputed or printed in text form. Not complete yet:
7024 still needs to check the OCSP response validity.
7027 *) New subcommands for 'openssl ca':
7028 'openssl ca -status <serial>' prints the status of the cert with
7029 the given serial number (according to the index file).
7030 'openssl ca -updatedb' updates the expiry status of certificates
7032 [Massimiliano Pala <madwolf@comune.modena.it>]
7034 *) New '-newreq-nodes' command option to CA.pl. This is like
7035 '-newreq', but calls 'openssl req' with the '-nodes' option
7036 so that the resulting key is not encrypted.
7037 [Damien Miller <djm@mindrot.org>]
7039 *) New configuration for the GNU Hurd.
7040 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7042 *) Initial code to implement OCSP basic response verify. This
7043 is currently incomplete. Currently just finds the signer's
7044 certificate and verifies the signature on the response.
7047 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7048 value of OPENSSLDIR. This is available via the new '-d' option
7049 to 'openssl version', and is also included in 'openssl version -a'.
7052 *) Allowing defining memory allocation callbacks that will be given
7053 file name and line number information in additional arguments
7054 (a const char* and an int). The basic functionality remains, as
7055 well as the original possibility to just replace malloc(),
7056 realloc() and free() by functions that do not know about these
7057 additional arguments. To register and find out the current
7058 settings for extended allocation functions, the following
7059 functions are provided:
7061 CRYPTO_set_mem_ex_functions
7062 CRYPTO_set_locked_mem_ex_functions
7063 CRYPTO_get_mem_ex_functions
7064 CRYPTO_get_locked_mem_ex_functions
7066 These work the same way as CRYPTO_set_mem_functions and friends.
7067 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7068 extended allocation function is enabled.
7069 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7070 a conventional allocation function is enabled.
7071 [Richard Levitte, Bodo Moeller]
7073 *) Finish off removing the remaining LHASH function pointer casts.
7074 There should no longer be any prototype-casting required when using
7075 the LHASH abstraction, and any casts that remain are "bugs". See
7076 the callback types and macros at the head of lhash.h for details
7077 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7080 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7081 If /dev/[u]random devices are not available or do not return enough
7082 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7084 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7085 /etc/entropy will be queried once each in this sequence, quering stops
7086 when enough entropy was collected without querying more sockets.
7089 *) Change the Unix RAND_poll() variant to be able to poll several
7090 random devices, as specified by DEVRANDOM, until a sufficient amount
7091 of data has been collected. We spend at most 10 ms on each file
7092 (select timeout) and read in non-blocking mode. DEVRANDOM now
7093 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7094 (previously it was just the string "/dev/urandom"), so on typical
7095 platforms the 10 ms delay will never occur.
7096 Also separate out the Unix variant to its own file, rand_unix.c.
7097 For VMS, there's a currently-empty rand_vms.c.
7100 *) Move OCSP client related routines to ocsp_cl.c. These
7101 provide utility functions which an application needing
7102 to issue a request to an OCSP responder and analyse the
7103 response will typically need: as opposed to those which an
7104 OCSP responder itself would need which will be added later.
7106 OCSP_request_sign() signs an OCSP request with an API similar
7107 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7108 response. OCSP_response_get1_basic() extracts basic response
7109 from response. OCSP_resp_find_status(): finds and extracts status
7110 information from an OCSP_CERTID structure (which will be created
7111 when the request structure is built). These are built from lower
7112 level functions which work on OCSP_SINGLERESP structures but
7113 wont normally be used unless the application wishes to examine
7114 extensions in the OCSP response for example.
7116 Replace nonce routines with a pair of functions.
7117 OCSP_request_add1_nonce() adds a nonce value and optionally
7118 generates a random value. OCSP_check_nonce() checks the
7119 validity of the nonce in an OCSP response.
7122 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7123 This doesn't copy the supplied OCSP_CERTID and avoids the
7124 need to free up the newly created id. Change return type
7125 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7126 This can then be used to add extensions to the request.
7127 Deleted OCSP_request_new(), since most of its functionality
7128 is now in OCSP_REQUEST_new() (and the case insensitive name
7129 clash) apart from the ability to set the request name which
7130 will be added elsewhere.
7133 *) Update OCSP API. Remove obsolete extensions argument from
7134 various functions. Extensions are now handled using the new
7135 OCSP extension code. New simple OCSP HTTP function which
7136 can be used to send requests and parse the response.
7139 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7140 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7141 uses the special reorder version of SET OF to sort the attributes
7142 and reorder them to match the encoded order. This resolves a long
7143 standing problem: a verify on a PKCS7 structure just after signing
7144 it used to fail because the attribute order did not match the
7145 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7146 it uses the received order. This is necessary to tolerate some broken
7147 software that does not order SET OF. This is handled by encoding
7148 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7149 to produce the required SET OF.
7152 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7153 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7154 files to get correct declarations of the ASN.1 item variables.
7157 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7158 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7159 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7160 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7161 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7162 ASN1_ITEM and no wrapper functions.
7165 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7166 replace the old function pointer based I/O routines. Change most of
7167 the *_d2i_bio() and *_d2i_fp() functions to use these.
7170 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7171 lines, recognice more "algorithms" that can be deselected, and make
7172 it complain about algorithm deselection that isn't recognised.
7175 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7176 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7177 to use new functions. Add NO_ASN1_OLD which can be set to remove
7178 some old style ASN1 functions: this can be used to determine if old
7179 code will still work when these eventually go away.
7182 *) New extension functions for OCSP structures, these follow the
7183 same conventions as certificates and CRLs.
7186 *) New function X509V3_add1_i2d(). This automatically encodes and
7187 adds an extension. Its behaviour can be customised with various
7188 flags to append, replace or delete. Various wrappers added for
7189 certifcates and CRLs.
7192 *) Fix to avoid calling the underlying ASN1 print routine when
7193 an extension cannot be parsed. Correct a typo in the
7194 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7197 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7198 entries for variables.
7201 *) Add functionality to apps/openssl.c for detecting locking
7202 problems: As the program is single-threaded, all we have
7203 to do is register a locking callback using an array for
7204 storing which locks are currently held by the program.
7207 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7208 SSL_get_ex_data_X509_STORE_idx(), which is used in
7209 ssl_verify_cert_chain() and thus can be called at any time
7210 during TLS/SSL handshakes so that thread-safety is essential.
7211 Unfortunately, the ex_data design is not at all suited
7212 for multi-threaded use, so it probably should be abolished.
7215 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7216 [Broadcom, tweaked and integrated by Geoff Thorpe]
7218 *) Move common extension printing code to new function
7219 X509V3_print_extensions(). Reorganise OCSP print routines and
7220 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7223 *) New function X509_signature_print() to remove duplication in some
7227 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7228 set (this was treated exactly the same as SET OF previously). This
7229 is used to reorder the STACK representing the structure to match the
7230 encoding. This will be used to get round a problem where a PKCS7
7231 structure which was signed could not be verified because the STACK
7232 order did not reflect the encoded order.
7235 *) Reimplement the OCSP ASN1 module using the new code.
7238 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7239 for its ASN1 operations. The old style function pointers still exist
7240 for now but they will eventually go away.
7243 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7244 completely replaces the old ASN1 functionality with a table driven
7245 encoder and decoder which interprets an ASN1_ITEM structure describing
7246 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7247 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7248 has also been converted to the new form.
7251 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7252 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7253 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7254 for negative moduli.
7257 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7258 of not touching the result's sign bit.
7261 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7265 *) Changed the LHASH code to use prototypes for callbacks, and created
7266 macros to declare and implement thin (optionally static) functions
7267 that provide type-safety and avoid function pointer casting for the
7268 type-specific callbacks.
7271 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7273 [Veers Staats <staatsvr@asc.hpc.mil>,
7274 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7276 *) Reformat the FAQ so the different questions and answers can be divided
7277 in sections depending on the subject.
7280 *) Have the zlib compression code load ZLIB.DLL dynamically under
7284 *) New function BN_mod_sqrt for computing square roots modulo a prime
7285 (using the probabilistic Tonelli-Shanks algorithm unless
7286 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7287 be handled deterministically).
7288 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7290 *) Make BN_mod_inverse faster by explicitly handling small quotients
7291 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7292 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7295 *) New function BN_kronecker.
7298 *) Fix BN_gcd so that it works on negative inputs; the result is
7299 positive unless both parameters are zero.
7300 Previously something reasonably close to an infinite loop was
7301 possible because numbers could be growing instead of shrinking
7302 in the implementation of Euclid's algorithm.
7305 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7306 sign of the number in question.
7308 Fix BN_is_word(a,w) to work correctly for w == 0.
7310 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7311 because its test if the absolute value of 'a' equals 'w'.
7312 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7313 it exists mostly for use in the implementations of BN_is_zero(),
7314 BN_is_one(), and BN_is_word().
7317 *) New function BN_swap.
7320 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7321 the exponentiation functions are more likely to produce reasonable
7322 results on negative inputs.
7325 *) Change BN_mod_mul so that the result is always non-negative.
7326 Previously, it could be negative if one of the factors was negative;
7327 I don't think anyone really wanted that behaviour.
7330 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7331 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7332 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7333 and add new functions:
7342 BN_mod_lshift1_quick
7346 These functions always generate non-negative results.
7348 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7349 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7351 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7352 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7353 be reduced modulo m.
7354 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7357 The following entry accidentally appeared in the CHANGES file
7358 distributed with OpenSSL 0.9.7. The modifications described in
7359 it do *not* apply to OpenSSL 0.9.7.
7361 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7362 was actually never needed) and in BN_mul(). The removal in BN_mul()
7363 required a small change in bn_mul_part_recursive() and the addition
7364 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7365 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7366 bn_sub_words() and bn_add_words() except they take arrays with
7371 *) In 'openssl passwd', verify passwords read from the terminal
7372 unless the '-salt' option is used (which usually means that
7373 verification would just waste user's time since the resulting
7374 hash is going to be compared with some given password hash)
7375 or the new '-noverify' option is used.
7377 This is an incompatible change, but it does not affect
7378 non-interactive use of 'openssl passwd' (passwords on the command
7379 line, '-stdin' option, '-in ...' option) and thus should not
7383 *) Remove all references to RSAref, since there's no more need for it.
7386 *) Make DSO load along a path given through an environment variable
7387 (SHLIB_PATH) with shl_load().
7390 *) Constify the ENGINE code as a result of BIGNUM constification.
7391 Also constify the RSA code and most things related to it. In a
7392 few places, most notable in the depth of the ASN.1 code, ugly
7393 casts back to non-const were required (to be solved at a later
7397 *) Make it so the openssl application has all engines loaded by default.
7400 *) Constify the BIGNUM routines a little more.
7403 *) Add the following functions:
7405 ENGINE_load_cswift()
7407 ENGINE_load_atalla()
7409 ENGINE_load_builtin_engines()
7411 That way, an application can itself choose if external engines that
7412 are built-in in OpenSSL shall ever be used or not. The benefit is
7413 that applications won't have to be linked with libdl or other dso
7414 libraries unless it's really needed.
7416 Changed 'openssl engine' to load all engines on demand.
7417 Changed the engine header files to avoid the duplication of some
7418 declarations (they differed!).
7421 *) 'openssl engine' can now list capabilities.
7424 *) Better error reporting in 'openssl engine'.
7427 *) Never call load_dh_param(NULL) in s_server.
7430 *) Add engine application. It can currently list engines by name and
7431 identity, and test if they are actually available.
7434 *) Improve RPM specification file by forcing symbolic linking and making
7435 sure the installed documentation is also owned by root.root.
7436 [Damien Miller <djm@mindrot.org>]
7438 *) Give the OpenSSL applications more possibilities to make use of
7439 keys (public as well as private) handled by engines.
7442 *) Add OCSP code that comes from CertCo.
7445 *) Add VMS support for the Rijndael code.
7448 *) Added untested support for Nuron crypto accelerator.
7451 *) Add support for external cryptographic devices. This code was
7452 previously distributed separately as the "engine" branch.
7453 [Geoff Thorpe, Richard Levitte]
7455 *) Rework the filename-translation in the DSO code. It is now possible to
7456 have far greater control over how a "name" is turned into a filename
7457 depending on the operating environment and any oddities about the
7458 different shared library filenames on each system.
7461 *) Support threads on FreeBSD-elf in Configure.
7464 *) Fix for SHA1 assembly problem with MASM: it produces
7465 warnings about corrupt line number information when assembling
7466 with debugging information. This is caused by the overlapping
7468 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7471 NCONF_get_number() has no error checking at all. As a replacement,
7472 NCONF_get_number_e() is defined (_e for "error checking") and is
7473 promoted strongly. The old NCONF_get_number is kept around for
7474 binary backward compatibility.
7475 Make it possible for methods to load from something other than a BIO,
7476 by providing a function pointer that is given a name instead of a BIO.
7477 For example, this could be used to load configuration data from an
7481 *) Fix for non blocking accept BIOs. Added new I/O special reason
7482 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7483 with non blocking I/O was not possible because no retry code was
7484 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7488 *) Added the beginnings of Rijndael support.
7491 *) Fix for bug in DirectoryString mask setting. Add support for
7492 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7493 to allow certificate printing to more controllable, additional
7494 'certopt' option to 'x509' to allow new printing options to be
7498 *) Clean old EAY MD5 hack from e_os.h.
7501 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7503 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7504 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7505 [Joe Orton, Steve Henson]
7507 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7509 *) Fix additional bug revealed by the NISCC test suite:
7511 Stop bug triggering large recursion when presented with
7512 certain ASN.1 tags (CVE-2003-0851)
7515 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7517 *) Fix various bugs revealed by running the NISCC test suite:
7519 Stop out of bounds reads in the ASN1 code when presented with
7520 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7522 If verify callback ignores invalid public key errors don't try to check
7523 certificate signature with the NULL public key.
7527 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7528 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7532 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7533 extra data after the compression methods not only for TLS 1.0
7534 but also for SSL 3.0 (as required by the specification).
7535 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7537 *) Change X509_certificate_type() to mark the key as exported/exportable
7538 when it's 512 *bits* long, not 512 bytes.
7541 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7543 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7544 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7545 a protocol version number mismatch like a decryption error
7546 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7549 *) Turn on RSA blinding by default in the default implementation
7550 to avoid a timing attack. Applications that don't want it can call
7551 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7552 They would be ill-advised to do so in most cases.
7553 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7555 *) Change RSA blinding code so that it works when the PRNG is not
7556 seeded (in this case, the secret RSA exponent is abused as
7557 an unpredictable seed -- if it is not unpredictable, there
7558 is no point in blinding anyway). Make RSA blinding thread-safe
7559 by remembering the creator's thread ID in rsa->blinding and
7560 having all other threads use local one-time blinding factors
7561 (this requires more computation than sharing rsa->blinding, but
7562 avoids excessive locking; and if an RSA object is not shared
7563 between threads, blinding will still be very fast).
7566 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7568 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7569 via timing by performing a MAC computation even if incorrrect
7570 block cipher padding has been found. This is a countermeasure
7571 against active attacks where the attacker has to distinguish
7572 between bad padding and a MAC verification error. (CVE-2003-0078)
7574 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7575 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7576 Martin Vuagnoux (EPFL, Ilion)]
7578 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7580 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7581 memory from it's contents. This is done with a counter that will
7582 place alternating values in each byte. This can be used to solve
7583 two issues: 1) the removal of calls to memset() by highly optimizing
7584 compilers, and 2) cleansing with other values than 0, since those can
7585 be read through on certain media, for example a swap space on disk.
7588 *) Bugfix: client side session caching did not work with external caching,
7589 because the session->cipher setting was not restored when reloading
7590 from the external cache. This problem was masked, when
7591 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7592 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7595 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7596 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7597 [Zeev Lieber <zeev-l@yahoo.com>]
7599 *) Undo an undocumented change introduced in 0.9.6e which caused
7600 repeated calls to OpenSSL_add_all_ciphers() and
7601 OpenSSL_add_all_digests() to be ignored, even after calling
7605 *) Change the default configuration reader to deal with last line not
7606 being properly terminated.
7609 *) Change X509_NAME_cmp() so it applies the special rules on handling
7610 DN values that are of type PrintableString, as well as RDNs of type
7611 emailAddress where the value has the type ia5String.
7612 [stefank@valicert.com via Richard Levitte]
7614 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7615 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7616 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7617 the bitwise-OR of the two for use by the majority of applications
7618 wanting this behaviour, and update the docs. The documented
7619 behaviour and actual behaviour were inconsistent and had been
7620 changing anyway, so this is more a bug-fix than a behavioural
7622 [Geoff Thorpe, diagnosed by Nadav Har'El]
7624 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7625 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7628 *) Fix initialization code race conditions in
7629 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7630 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7631 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7632 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7633 ssl2_get_cipher_by_char(),
7634 ssl3_get_cipher_by_char().
7635 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7637 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7638 the cached sessions are flushed, as the remove_cb() might use ex_data
7639 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7640 (see [openssl.org #212]).
7641 [Geoff Thorpe, Lutz Jaenicke]
7643 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7644 length, instead of the encoding length to d2i_ASN1_OBJECT.
7647 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7649 *) [In 0.9.6g-engine release:]
7650 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7651 [Lynn Gazis <lgazis@rainbow.com>]
7653 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7655 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7656 and get fix the header length calculation.
7657 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7658 Alon Kantor <alonk@checkpoint.com> (and others),
7661 *) Use proper error handling instead of 'assertions' in buffer
7662 overflow checks added in 0.9.6e. This prevents DoS (the
7663 assertions could call abort()).
7664 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7666 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7668 *) Add various sanity checks to asn1_get_length() to reject
7669 the ASN1 length bytes if they exceed sizeof(long), will appear
7670 negative or the content length exceeds the length of the
7672 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7674 *) Fix cipher selection routines: ciphers without encryption had no flags
7675 for the cipher strength set and where therefore not handled correctly
7676 by the selection routines (PR #130).
7679 *) Fix EVP_dsa_sha macro.
7683 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7684 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7685 that was added in OpenSSL 0.9.6d.
7687 As the countermeasure turned out to be incompatible with some
7688 broken SSL implementations, the new option is part of SSL_OP_ALL.
7689 SSL_OP_ALL is usually employed when compatibility with weird SSL
7690 implementations is desired (e.g. '-bugs' option to 's_client' and
7691 's_server'), so the new option is automatically set in many
7695 *) Changes in security patch:
7697 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7698 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7699 Air Force Materiel Command, USAF, under agreement number
7702 *) Add various sanity checks to asn1_get_length() to reject
7703 the ASN1 length bytes if they exceed sizeof(long), will appear
7704 negative or the content length exceeds the length of the
7705 supplied buffer. (CVE-2002-0659)
7706 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7708 *) Assertions for various potential buffer overflows, not known to
7710 [Ben Laurie (CHATS)]
7712 *) Various temporary buffers to hold ASCII versions of integers were
7713 too small for 64 bit platforms. (CVE-2002-0655)
7714 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7716 *) Remote buffer overflow in SSL3 protocol - an attacker could
7717 supply an oversized session ID to a client. (CVE-2002-0656)
7718 [Ben Laurie (CHATS)]
7720 *) Remote buffer overflow in SSL2 protocol - an attacker could
7721 supply an oversized client master key. (CVE-2002-0656)
7722 [Ben Laurie (CHATS)]
7724 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7726 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7727 encoded as NULL) with id-dsa-with-sha1.
7728 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7730 *) Check various X509_...() return values in apps/req.c.
7731 [Nils Larsch <nla@trustcenter.de>]
7733 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7734 an end-of-file condition would erronously be flagged, when the CRLF
7735 was just at the end of a processed block. The bug was discovered when
7736 processing data through a buffering memory BIO handing the data to a
7737 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7738 <ptsekov@syntrex.com> and Nedelcho Stanev.
7741 *) Implement a countermeasure against a vulnerability recently found
7742 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7743 before application data chunks to avoid the use of known IVs
7744 with data potentially chosen by the attacker.
7747 *) Fix length checks in ssl3_get_client_hello().
7750 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7751 to prevent ssl3_read_internal() from incorrectly assuming that
7752 ssl3_read_bytes() found application data while handshake
7753 processing was enabled when in fact s->s3->in_read_app_data was
7754 merely automatically cleared during the initial handshake.
7755 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7757 *) Fix object definitions for Private and Enterprise: they were not
7758 recognized in their shortname (=lowercase) representation. Extend
7759 obj_dat.pl to issue an error when using undefined keywords instead
7760 of silently ignoring the problem (Svenning Sorensen
7761 <sss@sss.dnsalias.net>).
7764 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7765 generators, i.e. generators other than 2 and 5. (Previously, the
7766 code did not properly initialise the 'add' and 'rem' values to
7767 BN_generate_prime().)
7769 In the new general case, we do not insist that 'generator' is
7770 actually a primitive root: This requirement is rather pointless;
7771 a generator of the order-q subgroup is just as good, if not
7775 *) Map new X509 verification errors to alerts. Discovered and submitted by
7776 Tom Wu <tom@arcot.com>.
7779 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7780 returning non-zero before the data has been completely received
7781 when using non-blocking I/O.
7782 [Bodo Moeller; problem pointed out by John Hughes]
7784 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7785 [Ben Laurie, Lutz Jaenicke]
7787 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7788 Yoram Zahavi <YoramZ@gilian.com>).
7791 *) Add information about CygWin 1.3 and on, and preserve proper
7792 configuration for the versions before that.
7793 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7795 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7796 check whether we deal with a copy of a session and do not delete from
7797 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7798 <izhar@checkpoint.com>.
7801 *) Do not store session data into the internal session cache, if it
7802 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7803 flag is set). Proposed by Aslam <aslam@funk.com>.
7806 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7810 *) [In 0.9.6d-engine release:]
7811 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7812 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7814 *) Add the configuration target linux-s390x.
7815 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7817 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7818 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7819 variable as an indication that a ClientHello message has been
7820 received. As the flag value will be lost between multiple
7821 invocations of ssl3_accept when using non-blocking I/O, the
7822 function may not be aware that a handshake has actually taken
7823 place, thus preventing a new session from being added to the
7826 To avoid this problem, we now set s->new_session to 2 instead of
7827 using a local variable.
7828 [Lutz Jaenicke, Bodo Moeller]
7830 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7831 if the SSL_R_LENGTH_MISMATCH error is detected.
7832 [Geoff Thorpe, Bodo Moeller]
7834 *) New 'shared_ldflag' column in Configure platform table.
7837 *) Fix EVP_CIPHER_mode macro.
7838 ["Dan S. Camper" <dan@bti.net>]
7840 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7841 type, we must throw them away by setting rr->length to 0.
7842 [D P Chang <dpc@qualys.com>]
7844 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7846 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7847 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7848 worked incorrectly for those cases where range = 10..._2 and
7849 3*range is two bits longer than range.)
7852 *) Only add signing time to PKCS7 structures if it is not already
7856 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7857 OBJ_ld_ce should be OBJ_id_ce.
7858 Also some ip-pda OIDs in crypto/objects/objects.txt were
7859 incorrect (cf. RFC 3039).
7860 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7862 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7863 returns early because it has nothing to do.
7864 [Andy Schneider <andy.schneider@bjss.co.uk>]
7866 *) [In 0.9.6c-engine release:]
7867 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7868 [Andy Schneider <andy.schneider@bjss.co.uk>]
7870 *) [In 0.9.6c-engine release:]
7871 Add support for Cryptographic Appliance's keyserver technology.
7872 (Use engine 'keyclient')
7873 [Cryptographic Appliances and Geoff Thorpe]
7875 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7876 is called via tools/c89.sh because arguments have to be
7877 rearranged (all '-L' options must appear before the first object
7879 [Richard Shapiro <rshapiro@abinitio.com>]
7881 *) [In 0.9.6c-engine release:]
7882 Add support for Broadcom crypto accelerator cards, backported
7884 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7886 *) [In 0.9.6c-engine release:]
7887 Add support for SureWare crypto accelerator cards from
7888 Baltimore Technologies. (Use engine 'sureware')
7889 [Baltimore Technologies and Mark Cox]
7891 *) [In 0.9.6c-engine release:]
7892 Add support for crypto accelerator cards from Accelerated
7893 Encryption Processing, www.aep.ie. (Use engine 'aep')
7894 [AEP Inc. and Mark Cox]
7896 *) Add a configuration entry for gcc on UnixWare.
7897 [Gary Benson <gbenson@redhat.com>]
7899 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7900 messages are stored in a single piece (fixed-length part and
7901 variable-length part combined) and fix various bugs found on the way.
7904 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7905 instead. BIO_gethostbyname() does not know what timeouts are
7906 appropriate, so entries would stay in cache even when they have
7908 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7910 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7911 faced with a pathologically small ClientHello fragment that does
7912 not contain client_version: Instead of aborting with an error,
7913 simply choose the highest available protocol version (i.e.,
7914 TLS 1.0 unless it is disabled). In practice, ClientHello
7915 messages are never sent like this, but this change gives us
7916 strictly correct behaviour at least for TLS.
7919 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7920 never resets s->method to s->ctx->method when called from within
7921 one of the SSL handshake functions.
7922 [Bodo Moeller; problem pointed out by Niko Baric]
7924 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7925 (sent using the client's version number) if client_version is
7926 smaller than the protocol version in use. Also change
7927 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7928 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7929 the client will at least see that alert.
7932 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7936 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7937 client receives HelloRequest while in a handshake.
7938 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7940 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7941 should end in 'break', not 'goto end' which circumvents various
7942 cleanups done in state SSL_ST_OK. But session related stuff
7943 must be disabled for SSL_ST_OK in the case that we just sent a
7946 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7947 before just sending a HelloRequest.
7948 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7950 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7951 reveal whether illegal block cipher padding was found or a MAC
7952 verification error occurred. (Neither SSLerr() codes nor alerts
7953 are directly visible to potential attackers, but the information
7954 may leak via logfiles.)
7956 Similar changes are not required for the SSL 2.0 implementation
7957 because the number of padding bytes is sent in clear for SSL 2.0,
7958 and the extra bytes are just ignored. However ssl/s2_pkt.c
7959 failed to verify that the purported number of padding bytes is in
7963 *) Add OpenUNIX-8 support including shared libraries
7964 (Boyd Lynn Gerber <gerberb@zenez.com>).
7967 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7968 'wristwatch attack' using huge encoding parameters (cf.
7969 James H. Manger's CRYPTO 2001 paper). Note that the
7970 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7971 encoding parameters and hence was not vulnerable.
7974 *) BN_sqr() bug fix.
7975 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7977 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7978 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7979 followed by modular reduction.
7980 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7982 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7983 equivalent based on BN_pseudo_rand() instead of BN_rand().
7986 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7987 This function was broken, as the check for a new client hello message
7988 to handle SGC did not allow these large messages.
7989 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7992 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7995 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7996 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7999 *) Rework the configuration and shared library support for Tru64 Unix.
8000 The configuration part makes use of modern compiler features and
8001 still retains old compiler behavior for those that run older versions
8002 of the OS. The shared library support part includes a variant that
8003 uses the RPATH feature, and is available through the special
8004 configuration target "alpha-cc-rpath", which will never be selected
8006 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8008 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8009 with the same message size as in ssl3_get_certificate_request().
8010 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8011 messages might inadvertently be reject as too long.
8012 [Petr Lampa <lampa@fee.vutbr.cz>]
8014 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8017 *) Modified SSL library such that the verify_callback that has been set
8018 specificly for an SSL object with SSL_set_verify() is actually being
8019 used. Before the change, a verify_callback set with this function was
8020 ignored and the verify_callback() set in the SSL_CTX at the time of
8021 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8022 to allow the necessary settings.
8025 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8026 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8027 done automatically (in contradiction to the requirements of the C
8028 standard). This made problems when used from OpenSSH.
8031 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8032 dh->length and always used
8034 BN_rand_range(priv_key, dh->p).
8036 BN_rand_range() is not necessary for Diffie-Hellman, and this
8037 specific range makes Diffie-Hellman unnecessarily inefficient if
8038 dh->length (recommended exponent length) is much smaller than the
8039 length of dh->p. We could use BN_rand_range() if the order of
8040 the subgroup was stored in the DH structure, but we only have
8045 BN_rand(priv_key, l, ...)
8047 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8053 RSA_eay_public_encrypt
8054 RSA_eay_private_decrypt
8055 RSA_eay_private_encrypt (signing)
8056 RSA_eay_public_decrypt (signature verification)
8058 (default implementations for RSA_public_encrypt,
8059 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8060 always reject numbers >= n.
8063 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8064 to synchronize access to 'locking_thread'. This is necessary on
8065 systems where access to 'locking_thread' (an 'unsigned long'
8066 variable) is not atomic.
8069 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8070 *before* setting the 'crypto_lock_rand' flag. The previous code had
8071 a race condition if 0 is a valid thread ID.
8072 [Travis Vitek <vitek@roguewave.com>]
8074 *) Add support for shared libraries under Irix.
8075 [Albert Chin-A-Young <china@thewrittenword.com>]
8077 *) Add configuration option to build on Linux on both big-endian and
8079 [Ralf Baechle <ralf@uni-koblenz.de>]
8081 *) Add the possibility to create shared libraries on HP-UX.
8084 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8086 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8087 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8088 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8089 PRNG state recovery was possible based on the output of
8090 one PRNG request appropriately sized to gain knowledge on
8091 'md' followed by enough consecutive 1-byte PRNG requests
8092 to traverse all of 'state'.
8094 1. When updating 'md_local' (the current thread's copy of 'md')
8095 during PRNG output generation, hash all of the previous
8096 'md_local' value, not just the half used for PRNG output.
8098 2. Make the number of bytes from 'state' included into the hash
8099 independent from the number of PRNG bytes requested.
8101 The first measure alone would be sufficient to avoid
8102 Markku-Juhani's attack. (Actually it had never occurred
8103 to me that the half of 'md_local' used for chaining was the
8104 half from which PRNG output bytes were taken -- I had always
8105 assumed that the secret half would be used.) The second
8106 measure makes sure that additional data from 'state' is never
8107 mixed into 'md_local' in small portions; this heuristically
8108 further strengthens the PRNG.
8111 *) Fix crypto/bn/asm/mips3.s.
8114 *) When only the key is given to "enc", the IV is undefined. Print out
8115 an error message in this case.
8118 *) Handle special case when X509_NAME is empty in X509 printing routines.
8121 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8122 positive and less than q.
8125 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8126 used: it isn't thread safe and the add_lock_callback should handle
8128 [Paul Rose <Paul.Rose@bridge.com>]
8130 *) Verify that incoming data obeys the block size in
8131 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8135 [Ulf Möller, Bodo Möller]
8137 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8138 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8139 when fixing the server behaviour for backwards-compatible 'client
8140 hello' messages. (Note that the attack is impractical against
8141 SSL 3.0 and TLS 1.0 anyway because length and version checking
8142 means that the probability of guessing a valid ciphertext is
8143 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8146 Before 0.9.5, the countermeasure (hide the error by generating a
8147 random 'decryption result') did not work properly because
8148 ERR_clear_error() was missing, meaning that SSL_get_error() would
8149 detect the supposedly ignored error.
8151 Both problems are now fixed.
8154 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8155 (previously it was 1024).
8158 *) Fix for compatibility mode trust settings: ignore trust settings
8159 unless some valid trust or reject settings are present.
8162 *) Fix for blowfish EVP: its a variable length cipher.
8165 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8166 parameters in DSA public key structures and return an error in the
8167 DSA routines if parameters are absent.
8170 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8171 in the current directory if neither $RANDFILE nor $HOME was set.
8172 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8173 caused some confusion to Windows users who haven't defined $HOME.
8174 Thus RAND_file_name() is changed again: e_os.h can define a
8175 DEFAULT_HOME, which will be used if $HOME is not set.
8176 For Windows, we use "C:"; on other platforms, we still require
8177 environment variables.
8179 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8180 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8181 having multiple threads call RAND_poll() concurrently.
8184 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8185 combination of a flag and a thread ID variable.
8186 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8187 flag), *other* threads can enter ssleay_add_bytes without obeying
8188 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8189 that they do not hold after the first thread unsets add_do_not_lock).
8192 *) Change bctest again: '-x' expressions are not available in all
8196 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8198 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8199 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8201 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8202 the default extension for executables, if any. Also, make the perl
8203 scripts that use symlink() to test if it really exists and use "cp"
8204 if it doesn't. All this made OpenSSL compilable and installable in
8208 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8209 If SEQUENCE is length is indefinite just set c->slen to the total
8210 amount of data available.
8211 [Steve Henson, reported by shige@FreeBSD.org]
8212 [This change does not apply to 0.9.7.]
8214 *) Change bctest to avoid here-documents inside command substitution
8215 (workaround for FreeBSD /bin/sh bug).
8216 For compatibility with Ultrix, avoid shell functions (introduced
8217 in the bctest version that searches along $PATH).
8220 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8221 with des_encrypt() defined on some operating systems, like Solaris
8225 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8226 On the Importance of Eliminating Errors in Cryptographic
8227 Computations, J. Cryptology 14 (2001) 2, 101-119,
8228 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8231 *) MIPS assembler BIGNUM division bug fix.
8234 *) Disabled incorrect Alpha assembler code.
8237 *) Fix PKCS#7 decode routines so they correctly update the length
8238 after reading an EOC for the EXPLICIT tag.
8240 [This change does not apply to 0.9.7.]
8242 *) Fix bug in PKCS#12 key generation routines. This was triggered
8243 if a 3DES key was generated with a 0 initial byte. Include
8244 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8245 (but broken) behaviour.
8248 *) Enhance bctest to search for a working bc along $PATH and print
8250 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8252 *) Fix memory leaks in err.c: free err_data string if necessary;
8253 don't write to the wrong index in ERR_set_error_data.
8256 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8260 *) Replace rdtsc with _emit statements for VC++ version 5.
8261 [Jeremy Cooper <jeremy@baymoo.org>]
8263 *) Make it possible to reuse SSLv2 sessions.
8266 *) In copy_email() check for >= 0 as a return value for
8267 X509_NAME_get_index_by_NID() since 0 is a valid index.
8268 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8270 *) Avoid coredump with unsupported or invalid public keys by checking if
8271 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8272 PKCS7_verify() fails with non detached data.
8275 *) Don't use getenv in library functions when run as setuid/setgid.
8276 New function OPENSSL_issetugid().
8279 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8280 due to incorrect handling of multi-threading:
8282 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8284 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8286 3. Count how many times MemCheck_off() has been called so that
8287 nested use can be treated correctly. This also avoids
8288 inband-signalling in the previous code (which relied on the
8289 assumption that thread ID 0 is impossible).
8292 *) Add "-rand" option also to s_client and s_server.
8295 *) Fix CPU detection on Irix 6.x.
8296 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8297 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8299 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8302 [This change does not apply to 0.9.7.]
8304 *) Use the cached encoding of an X509_NAME structure rather than
8305 copying it. This is apparently the reason for the libsafe "errors"
8306 but the code is actually correct.
8309 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8310 Bleichenbacher's DSA attack.
8311 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8312 to be set and top=0 forces the highest bit to be set; top=-1 is new
8313 and leaves the highest bit random.
8314 [Ulf Moeller, Bodo Moeller]
8316 *) In the NCONF_...-based implementations for CONF_... queries
8317 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8318 a temporary CONF structure with the data component set to NULL
8319 (which gives segmentation faults in lh_retrieve).
8320 Instead, use NULL for the CONF pointer in CONF_get_string and
8321 CONF_get_number (which may use environment variables) and directly
8322 return NULL from CONF_get_section.
8325 *) Fix potential buffer overrun for EBCDIC.
8328 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8329 keyUsage if basicConstraints absent for a CA.
8332 *) Make SMIME_write_PKCS7() write mail header values with a format that
8333 is more generally accepted (no spaces before the semicolon), since
8334 some programs can't parse those values properly otherwise. Also make
8335 sure BIO's that break lines after each write do not create invalid
8339 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8340 macros previously used would not encode an empty SEQUENCE OF
8341 and break the signature.
8343 [This change does not apply to 0.9.7.]
8345 *) Zero the premaster secret after deriving the master secret in
8349 *) Add some EVP_add_digest_alias registrations (as found in
8350 OpenSSL_add_all_digests()) to SSL_library_init()
8351 aka OpenSSL_add_ssl_algorithms(). This provides improved
8352 compatibility with peers using X.509 certificates
8353 with unconventional AlgorithmIdentifier OIDs.
8356 *) Fix for Irix with NO_ASM.
8357 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8359 *) ./config script fixes.
8360 [Ulf Moeller, Richard Levitte]
8362 *) Fix 'openssl passwd -1'.
8365 *) Change PKCS12_key_gen_asc() so it can cope with non null
8366 terminated strings whose length is passed in the passlen
8367 parameter, for example from PEM callbacks. This was done
8368 by adding an extra length parameter to asc2uni().
8369 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8371 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8372 call failed, free the DSA structure.
8375 *) Fix to uni2asc() to cope with zero length Unicode strings.
8376 These are present in some PKCS#12 files.
8379 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8380 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8381 when writing a 32767 byte record.
8382 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8384 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8385 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8387 (RSA objects have a reference count access to which is protected
8388 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8389 so they are meant to be shared between threads.)
8390 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8391 "Reddie, Steven" <Steven.Reddie@ca.com>]
8393 *) Fix a deadlock in CRYPTO_mem_leaks().
8396 *) Use better test patterns in bntest.
8399 *) rand_win.c fix for Borland C.
8402 *) BN_rshift bugfix for n == 0.
8405 *) Add a 'bctest' script that checks for some known 'bc' bugs
8406 so that 'make test' does not abort just because 'bc' is broken.
8409 *) Store verify_result within SSL_SESSION also for client side to
8410 avoid potential security hole. (Re-used sessions on the client side
8411 always resulted in verify_result==X509_V_OK, not using the original
8412 result of the server certificate verification.)
8415 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8416 SSL3_RT_APPLICATION_DATA, return 0.
8417 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8421 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8422 releases, have been re-implemented by renaming the previous
8423 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8424 and ssl3_read_internal, respectively, and adding 'peek' parameters
8425 to them. The new ssl[23]_{read,peek} functions are calls to
8426 ssl[23]_read_internal with the 'peek' flag set appropriately.
8427 A 'peek' parameter has also been added to ssl3_read_bytes, which
8428 does the actual work for ssl3_read_internal.
8431 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8432 the method-specific "init()" handler. Also clean up ex_data after
8433 calling the method-specific "finish()" handler. Previously, this was
8434 happening the other way round.
8437 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8438 The previous value, 12, was not always sufficient for BN_mod_exp().
8441 *) Make sure that shared libraries get the internal name engine with
8442 the full version number and not just 0. This should mark the
8443 shared libraries as not backward compatible. Of course, this should
8444 be changed again when we can guarantee backward binary compatibility.
8447 *) Fix typo in get_cert_by_subject() in by_dir.c
8448 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8450 *) Rework the system to generate shared libraries:
8452 - Make note of the expected extension for the shared libraries and
8453 if there is a need for symbolic links from for example libcrypto.so.0
8454 to libcrypto.so.0.9.7. There is extended info in Configure for
8457 - Make as few rebuilds of the shared libraries as possible.
8459 - Still avoid linking the OpenSSL programs with the shared libraries.
8461 - When installing, install the shared libraries separately from the
8465 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8467 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8468 and not in SSL_clear because the latter is also used by the
8469 accept/connect functions; previously, the settings made by
8470 SSL_set_read_ahead would be lost during the handshake.
8471 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8473 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8474 Previously, it would create entries for disabled algorithms no
8478 *) Added several new manual pages for SSL_* function.
8481 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8483 *) In ssl23_get_client_hello, generate an error message when faced
8484 with an initial SSL 3.0/TLS record that is too small to contain the
8485 first two bytes of the ClientHello message, i.e. client_version.
8486 (Note that this is a pathologic case that probably has never happened
8487 in real life.) The previous approach was to use the version number
8488 from the record header as a substitute; but our protocol choice
8489 should not depend on that one because it is not authenticated
8490 by the Finished messages.
8493 *) More robust randomness gathering functions for Windows.
8494 [Jeffrey Altman <jaltman@columbia.edu>]
8496 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8497 not set then we don't setup the error code for issuer check errors
8498 to avoid possibly overwriting other errors which the callback does
8499 handle. If an application does set the flag then we assume it knows
8500 what it is doing and can handle the new informational codes
8504 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8505 a general "ANY" type, as such it should be able to decode anything
8506 including tagged types. However it didn't check the class so it would
8507 wrongly interpret tagged types in the same way as their universal
8508 counterpart and unknown types were just rejected. Changed so that the
8509 tagged and unknown types are handled in the same way as a SEQUENCE:
8510 that is the encoding is stored intact. There is also a new type
8511 "V_ASN1_OTHER" which is used when the class is not universal, in this
8512 case we have no idea what the actual type is so we just lump them all
8516 *) On VMS, stdout may very well lead to a file that is written to
8517 in a record-oriented fashion. That means that every write() will
8518 write a separate record, which will be read separately by the
8519 programs trying to read from it. This can be very confusing.
8521 The solution is to put a BIO filter in the way that will buffer
8522 text until a linefeed is reached, and then write everything a
8523 line at a time, so every record written will be an actual line,
8524 not chunks of lines and not (usually doesn't happen, but I've
8525 seen it once) several lines in one record. BIO_f_linebuffer() is
8528 Currently, it's a VMS-only method, because that's where it has
8529 been tested well enough.
8532 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8533 it can return incorrect results.
8534 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8535 but it was in 0.9.6-beta[12].)
8538 *) Disable the check for content being present when verifying detached
8539 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8540 include zero length content when signing messages.
8543 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8544 BIO_ctrl (for BIO pairs).
8547 *) Add DSO method for VMS.
8550 *) Bug fix: Montgomery multiplication could produce results with the
8554 *) Add RPM specification openssl.spec and modify it to build three
8555 packages. The default package contains applications, application
8556 documentation and run-time libraries. The devel package contains
8557 include files, static libraries and function documentation. The
8558 doc package contains the contents of the doc directory. The original
8559 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8562 *) Add a large number of documentation files for many SSL routines.
8563 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8565 *) Add a configuration entry for Sony News 4.
8566 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8568 *) Don't set the two most significant bits to one when generating a
8569 random number < q in the DSA library.
8572 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8573 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8574 the underlying transport is blocking) if a handshake took place.
8575 (The default behaviour is needed by applications such as s_client
8576 and s_server that use select() to determine when to use SSL_read;
8577 but for applications that know in advance when to expect data, it
8578 just makes things more complicated.)
8581 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8585 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8586 work better on such systems.
8587 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8589 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8590 Update PKCS12_parse() so it copies the friendlyName and the
8591 keyid to the certificates aux info.
8594 *) Fix bug in PKCS7_verify() which caused an infinite loop
8595 if there was more than one signature.
8596 [Sven Uszpelkat <su@celocom.de>]
8598 *) Major change in util/mkdef.pl to include extra information
8599 about each symbol, as well as presenting variables as well
8600 as functions. This change means that there's n more need
8601 to rebuild the .num files when some algorithms are excluded.
8604 *) Allow the verify time to be set by an application,
8605 rather than always using the current time.
8608 *) Phase 2 verify code reorganisation. The certificate
8609 verify code now looks up an issuer certificate by a
8610 number of criteria: subject name, authority key id
8611 and key usage. It also verifies self signed certificates
8612 by the same criteria. The main comparison function is
8613 X509_check_issued() which performs these checks.
8615 Lot of changes were necessary in order to support this
8616 without completely rewriting the lookup code.
8618 Authority and subject key identifier are now cached.
8620 The LHASH 'certs' is X509_STORE has now been replaced
8621 by a STACK_OF(X509_OBJECT). This is mainly because an
8622 LHASH can't store or retrieve multiple objects with
8623 the same hash value.
8625 As a result various functions (which were all internal
8626 use only) have changed to handle the new X509_STORE
8627 structure. This will break anything that messed round
8628 with X509_STORE internally.
8630 The functions X509_STORE_add_cert() now checks for an
8631 exact match, rather than just subject name.
8633 The X509_STORE API doesn't directly support the retrieval
8634 of multiple certificates matching a given criteria, however
8635 this can be worked round by performing a lookup first
8636 (which will fill the cache with candidate certificates)
8637 and then examining the cache for matches. This is probably
8638 the best we can do without throwing out X509_LOOKUP
8639 entirely (maybe later...).
8641 The X509_VERIFY_CTX structure has been enhanced considerably.
8643 All certificate lookup operations now go via a get_issuer()
8644 callback. Although this currently uses an X509_STORE it
8645 can be replaced by custom lookups. This is a simple way
8646 to bypass the X509_STORE hackery necessary to make this
8647 work and makes it possible to use more efficient techniques
8648 in future. A very simple version which uses a simple
8649 STACK for its trusted certificate store is also provided
8650 using X509_STORE_CTX_trusted_stack().
8652 The verify_cb() and verify() callbacks now have equivalents
8653 in the X509_STORE_CTX structure.
8655 X509_STORE_CTX also has a 'flags' field which can be used
8656 to customise the verify behaviour.
8659 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8660 excludes S/MIME capabilities.
8663 *) When a certificate request is read in keep a copy of the
8664 original encoding of the signed data and use it when outputing
8665 again. Signatures then use the original encoding rather than
8666 a decoded, encoded version which may cause problems if the
8667 request is improperly encoded.
8670 *) For consistency with other BIO_puts implementations, call
8671 buffer_write(b, ...) directly in buffer_puts instead of calling
8674 In BIO_puts, increment b->num_write as in BIO_write.
8675 [Peter.Sylvester@EdelWeb.fr]
8677 *) Fix BN_mul_word for the case where the word is 0. (We have to use
8678 BN_zero, we may not return a BIGNUM with an array consisting of
8682 *) Avoid calling abort() from within the library when problems are
8683 detected, except if preprocessor symbols have been defined
8684 (such as REF_CHECK, BN_DEBUG etc.).
8687 *) New openssl application 'rsautl'. This utility can be
8688 used for low level RSA operations. DER public key
8689 BIO/fp routines also added.
8692 *) New Configure entry and patches for compiling on QNX 4.
8693 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8695 *) A demo state-machine implementation was sponsored by
8696 Nuron (http://www.nuron.com/) and is now available in
8697 demos/state_machine.
8700 *) New options added to the 'dgst' utility for signature
8701 generation and verification.
8704 *) Unrecognized PKCS#7 content types are now handled via a
8705 catch all ASN1_TYPE structure. This allows unsupported
8706 types to be stored as a "blob" and an application can
8707 encode and decode it manually.
8710 *) Fix various signed/unsigned issues to make a_strex.c
8712 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8714 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8715 length if passed a buffer. ASN1_INTEGER_to_BN failed
8716 if passed a NULL BN and its argument was negative.
8717 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8719 *) Modification to PKCS#7 encoding routines to output definite
8720 length encoding. Since currently the whole structures are in
8721 memory there's not real point in using indefinite length
8722 constructed encoding. However if OpenSSL is compiled with
8723 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8726 *) Added BIO_vprintf() and BIO_vsnprintf().
8729 *) Added more prefixes to parse for in the the strings written
8730 through a logging bio, to cover all the levels that are available
8731 through syslog. The prefixes are now:
8733 PANIC, EMERG, EMR => LOG_EMERG
8734 ALERT, ALR => LOG_ALERT
8735 CRIT, CRI => LOG_CRIT
8736 ERROR, ERR => LOG_ERR
8737 WARNING, WARN, WAR => LOG_WARNING
8738 NOTICE, NOTE, NOT => LOG_NOTICE
8739 INFO, INF => LOG_INFO
8740 DEBUG, DBG => LOG_DEBUG
8742 and as before, if none of those prefixes are present at the
8743 beginning of the string, LOG_ERR is chosen.
8745 On Win32, the LOG_* levels are mapped according to this:
8747 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8748 LOG_WARNING => EVENTLOG_WARNING_TYPE
8749 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8753 *) Made it possible to reconfigure with just the configuration
8754 argument "reconf" or "reconfigure". The command line arguments
8755 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8756 and are retrieved from there when reconfiguring.
8760 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8762 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8765 *) The obj_dat.pl script was messing up the sorting of object
8766 names. The reason was that it compared the quoted version
8767 of strings as a result "OCSP" > "OCSP Signing" because
8768 " > SPACE. Changed script to store unquoted versions of
8769 names and add quotes on output. It was also omitting some
8770 names from the lookup table if they were given a default
8771 value (that is if SN is missing it is given the same
8772 value as LN and vice versa), these are now added on the
8773 grounds that if an object has a name we should be able to
8774 look it up. Finally added warning output when duplicate
8775 short or long names are found.
8778 *) Changes needed for Tandem NSK.
8779 [Scott Uroff <scott@xypro.com>]
8781 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8782 RSA_padding_check_SSLv23(), special padding was never detected
8783 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8784 version rollback attacks was not effective.
8786 In s23_clnt.c, don't use special rollback-attack detection padding
8787 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8788 client; similarly, in s23_srvr.c, don't do the rollback check if
8789 SSL 2.0 is the only protocol enabled in the server.
8792 *) Make it possible to get hexdumps of unprintable data with 'openssl
8793 asn1parse'. By implication, the functions ASN1_parse_dump() and
8794 BIO_dump_indent() are added.
8797 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8798 these print out strings and name structures based on various
8799 flags including RFC2253 support and proper handling of
8800 multibyte characters. Added options to the 'x509' utility
8801 to allow the various flags to be set.
8804 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8805 Also change the functions X509_cmp_current_time() and
8806 X509_gmtime_adj() work with an ASN1_TIME structure,
8807 this will enable certificates using GeneralizedTime in validity
8808 dates to be checked.
8811 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8812 negative public key encodings) on by default,
8813 NO_NEG_PUBKEY_BUG can be set to disable it.
8816 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8817 content octets. An i2c_ASN1_OBJECT is unnecessary because
8818 the encoding can be trivially obtained from the structure.
8821 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8822 not read locks (CRYPTO_r_[un]lock).
8825 *) A first attempt at creating official support for shared
8826 libraries through configuration. I've kept it so the
8827 default is static libraries only, and the OpenSSL programs
8828 are always statically linked for now, but there are
8829 preparations for dynamic linking in place.
8830 This has been tested on Linux and Tru64.
8833 *) Randomness polling function for Win9x, as described in:
8834 Peter Gutmann, Software Generation of Practically Strong
8838 *) Fix so PRNG is seeded in req if using an already existing
8842 *) New options to smime application. -inform and -outform
8843 allow alternative formats for the S/MIME message including
8844 PEM and DER. The -content option allows the content to be
8845 specified separately. This should allow things like Netscape
8846 form signing output easier to verify.
8849 *) Fix the ASN1 encoding of tags using the 'long form'.
8852 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8853 STRING types. These convert content octets to and from the
8854 underlying type. The actual tag and length octets are
8855 already assumed to have been read in and checked. These
8856 are needed because all other string types have virtually
8857 identical handling apart from the tag. By having versions
8858 of the ASN1 functions that just operate on content octets
8859 IMPLICIT tagging can be handled properly. It also allows
8860 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8861 and ASN1_INTEGER are identical apart from the tag.
8864 *) Change the handling of OID objects as follows:
8866 - New object identifiers are inserted in objects.txt, following
8867 the syntax given in objects.README.
8868 - objects.pl is used to process obj_mac.num and create a new
8870 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8873 This is currently kind of a hack, and the perl code in objects.pl
8874 isn't very elegant, but it works as I intended. The simplest way
8875 to check that it worked correctly is to look in obj_dat.h and
8876 check the array nid_objs and make sure the objects haven't moved
8877 around (this is important!). Additions are OK, as well as
8878 consistent name changes.
8881 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8884 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8885 The given file adds to whatever has already been seeded into the
8886 random pool through the RANDFILE configuration file option or
8887 environment variable, or the default random state file.
8890 *) mkstack.pl now sorts each macro group into lexical order.
8891 Previously the output order depended on the order the files
8892 appeared in the directory, resulting in needless rewriting
8896 *) Patches to make OpenSSL compile under Win32 again. Mostly
8897 work arounds for the VC++ problem that it treats func() as
8898 func(void). Also stripped out the parts of mkdef.pl that
8899 added extra typesafe functions: these no longer exist.
8902 *) Reorganisation of the stack code. The macros are now all
8903 collected in safestack.h . Each macro is defined in terms of
8904 a "stack macro" of the form SKM_<name>(type, a, b). The
8905 DEBUG_SAFESTACK is now handled in terms of function casts,
8906 this has the advantage of retaining type safety without the
8907 use of additional functions. If DEBUG_SAFESTACK is not defined
8908 then the non typesafe macros are used instead. Also modified the
8909 mkstack.pl script to handle the new form. Needs testing to see
8910 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8911 the default if no major problems. Similar behaviour for ASN1_SET_OF
8912 and PKCS12_STACK_OF.
8915 *) When some versions of IIS use the 'NET' form of private key the
8916 key derivation algorithm is different. Normally MD5(password) is
8917 used as a 128 bit RC4 key. In the modified case
8918 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8919 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8920 as the old Netscape_RSA functions except they have an additional
8921 'sgckey' parameter which uses the modified algorithm. Also added
8922 an -sgckey command line option to the rsa utility. Thanks to
8923 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8924 algorithm to openssl-dev.
8927 *) The evp_local.h macros were using 'c.##kname' which resulted in
8928 invalid expansion on some systems (SCO 5.0.5 for example).
8929 Corrected to 'c.kname'.
8930 [Phillip Porch <root@theporch.com>]
8932 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8933 a STACK of email addresses from a certificate or request, these look
8934 in the subject name and the subject alternative name extensions and
8935 omit any duplicate addresses.
8938 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8939 This makes DSA verification about 2 % faster.
8942 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8943 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8944 plus overhead for 1024 bit moduli).
8945 This makes exponentiations about 0.5 % faster for 1024 bit
8946 exponents (as measured by "openssl speed rsa2048").
8949 *) Rename memory handling macros to avoid conflicts with other
8951 Malloc => OPENSSL_malloc
8952 Malloc_locked => OPENSSL_malloc_locked
8953 Realloc => OPENSSL_realloc
8954 Free => OPENSSL_free
8957 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8958 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8961 *) CygWin32 support.
8962 [John Jarvie <jjarvie@newsguy.com>]
8964 *) The type-safe stack code has been rejigged. It is now only compiled
8965 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8966 by default all type-specific stack functions are "#define"d back to
8967 standard stack functions. This results in more streamlined output
8968 but retains the type-safety checking possibilities of the original
8972 *) The STACK code has been cleaned up, and certain type declarations
8973 that didn't make a lot of sense have been brought in line. This has
8974 also involved a cleanup of sorts in safestack.h to more correctly
8975 map type-safe stack functions onto their plain stack counterparts.
8976 This work has also resulted in a variety of "const"ifications of
8977 lots of the code, especially "_cmp" operations which should normally
8978 be prototyped with "const" parameters anyway.
8981 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8982 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8983 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8984 where all of 'md' is used each time the PRNG is used, but 'state'
8985 is used only indexed by a cyclic counter. As entropy may not be
8986 well distributed from the beginning, 'md' is important as a
8987 chaining variable. However, the output function chains only half
8988 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8989 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8990 in all of 'state' being rewritten, with the new values depending
8991 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8994 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8995 the handshake is continued after ssl_verify_cert_chain();
8996 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8997 can lead to 'unexplainable' connection aborts later.
8998 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9000 *) Major EVP API cipher revision.
9001 Add hooks for extra EVP features. This allows various cipher
9002 parameters to be set in the EVP interface. Support added for variable
9003 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9004 setting of RC2 and RC5 parameters.
9006 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9009 Remove lots of duplicated code from the EVP library. For example *every*
9010 cipher init() function handles the 'iv' in the same way according to the
9011 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9012 for CFB and OFB modes they zero ctx->num.
9014 New functionality allows removal of S/MIME code RC2 hack.
9016 Most of the routines have the same form and so can be declared in terms
9019 By shifting this to the top level EVP_CipherInit() it can be removed from
9020 all individual ciphers. If the cipher wants to handle IVs or keys
9021 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9024 Change lots of functions like EVP_EncryptUpdate() to now return a
9025 value: although software versions of the algorithms cannot fail
9026 any installed hardware versions can.
9029 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9030 this option is set, tolerate broken clients that send the negotiated
9031 protocol version number instead of the requested protocol version
9035 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9036 i.e. non-zero for export ciphersuites, zero otherwise.
9037 Previous versions had this flag inverted, inconsistent with
9038 rsa_tmp_cb (..._TMP_RSA_CB).
9039 [Bodo Moeller; problem reported by Amit Chopra]
9041 *) Add missing DSA library text string. Work around for some IIS
9042 key files with invalid SEQUENCE encoding.
9045 *) Add a document (doc/standards.txt) that list all kinds of standards
9046 and so on that are implemented in OpenSSL.
9049 *) Enhance c_rehash script. Old version would mishandle certificates
9050 with the same subject name hash and wouldn't handle CRLs at all.
9051 Added -fingerprint option to crl utility, to support new c_rehash
9055 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9058 *) Fix for SSL server purpose checking. Server checking was
9059 rejecting certificates which had extended key usage present
9060 but no ssl client purpose.
9061 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9063 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9064 is a little unclear about how a blank password is handled.
9065 Since the password in encoded as a BMPString with terminating
9066 double NULL a zero length password would end up as just the
9067 double NULL. However no password at all is different and is
9068 handled differently in the PKCS#12 key generation code. NS
9069 treats a blank password as zero length. MSIE treats it as no
9070 password on export: but it will try both on import. We now do
9071 the same: PKCS12_parse() tries zero length and no password if
9072 the password is set to "" or NULL (NULL is now a valid password:
9073 it wasn't before) as does the pkcs12 application.
9076 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9077 perror when PEM_read_bio_X509_REQ fails, the error message must
9078 be obtained from the error queue.
9081 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9082 it in ERR_remove_state if appropriate, and change ERR_get_state
9083 accordingly to avoid race conditions (this is necessary because
9084 thread_hash is no longer constant once set).
9087 *) Bugfix for linux-elf makefile.one.
9090 *) RSA_get_default_method() will now cause a default
9091 RSA_METHOD to be chosen if one doesn't exist already.
9092 Previously this was only set during a call to RSA_new()
9093 or RSA_new_method(NULL) meaning it was possible for
9094 RSA_get_default_method() to return NULL.
9097 *) Added native name translation to the existing DSO code
9098 that will convert (if the flag to do so is set) filenames
9099 that are sufficiently small and have no path information
9100 into a canonical native form. Eg. "blah" converted to
9101 "libblah.so" or "blah.dll" etc.
9104 *) New function ERR_error_string_n(e, buf, len) which is like
9105 ERR_error_string(e, buf), but writes at most 'len' bytes
9106 including the 0 terminator. For ERR_error_string_n, 'buf'
9108 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9110 *) CONF library reworked to become more general. A new CONF
9111 configuration file reader "class" is implemented as well as a
9112 new functions (NCONF_*, for "New CONF") to handle it. The now
9113 old CONF_* functions are still there, but are reimplemented to
9114 work in terms of the new functions. Also, a set of functions
9115 to handle the internal storage of the configuration data is
9116 provided to make it easier to write new configuration file
9117 reader "classes" (I can definitely see something reading a
9118 configuration file in XML format, for example), called _CONF_*,
9119 or "the configuration storage API"...
9121 The new configuration file reading functions are:
9123 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9124 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9126 NCONF_default, NCONF_WIN32
9128 NCONF_dump_fp, NCONF_dump_bio
9130 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9131 NCONF_new creates a new CONF object. This works in the same way
9132 as other interfaces in OpenSSL, like the BIO interface.
9133 NCONF_dump_* dump the internal storage of the configuration file,
9134 which is useful for debugging. All other functions take the same
9135 arguments as the old CONF_* functions wth the exception of the
9136 first that must be a `CONF *' instead of a `LHASH *'.
9138 To make it easer to use the new classes with the old CONF_* functions,
9139 the function CONF_set_default_method is provided.
9142 *) Add '-tls1' option to 'openssl ciphers', which was already
9143 mentioned in the documentation but had not been implemented.
9144 (This option is not yet really useful because even the additional
9145 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9148 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9149 OpenSSL-based applications) load shared libraries and bind to
9150 them in a portable way.
9151 [Geoff Thorpe, with contributions from Richard Levitte]
9153 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9155 *) Make sure _lrotl and _lrotr are only used with MSVC.
9157 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9158 (the default implementation of RAND_status).
9160 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9161 to '-clrext' (= clear extensions), as intended and documented.
9162 [Bodo Moeller; inconsistency pointed out by Michael Attili
9165 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9166 was larger than the MD block size.
9167 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9169 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9170 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9171 using the passed key: if the passed key was a private key the result
9172 of X509_print(), for example, would be to print out all the private key
9176 *) des_quad_cksum() byte order bug fix.
9177 [Ulf Möller, using the problem description in krb4-0.9.7, where
9178 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9180 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9182 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9184 *) For easily testing in shell scripts whether some command
9185 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9186 returns with exit code 0 iff no command of the given name is available.
9187 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9188 the output goes to stdout and nothing is printed to stderr.
9189 Additional arguments are always ignored.
9191 Since for each cipher there is a command of the same name,
9192 the 'no-cipher' compilation switches can be tested this way.
9194 ('openssl no-XXX' is not able to detect pseudo-commands such
9195 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9198 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9201 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9202 is set; it will be thrown away anyway because each handshake creates
9204 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9205 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9206 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9207 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9210 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9211 'Q' and 'R' lose their special meanings (quit/renegotiate).
9212 This is part of what -quiet does; unlike -quiet, -ign_eof
9213 does not suppress any output.
9216 *) Add compatibility options to the purpose and trust code. The
9217 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9218 accepts a certificate or CA, this was the previous behaviour,
9219 with all the associated security issues.
9221 X509_TRUST_COMPAT is the old trust behaviour: only and
9222 automatically trust self signed roots in certificate store. A
9223 new trust setting X509_TRUST_DEFAULT is used to specify that
9224 a purpose has no associated trust setting and it should instead
9225 use the value in the default purpose.
9228 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9229 and fix a memory leak.
9232 *) In util/mkerr.pl (which implements 'make errors'), preserve
9233 reason strings from the previous version of the .c file, as
9234 the default to have only downcase letters (and digits) in
9235 automatically generated reasons codes is not always appropriate.
9238 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9239 using strerror. Previously, ERR_reason_error_string() returned
9240 library names as reason strings for SYSerr; but SYSerr is a special
9241 case where small numbers are errno values, not library numbers.
9244 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9245 converts DSA parameters into DH parameters. (When creating parameters,
9246 DSA_generate_parameters is used.)
9249 *) Include 'length' (recommended exponent length) in C code generated
9250 by 'openssl dhparam -C'.
9253 *) The second argument to set_label in perlasm was already being used
9254 so couldn't be used as a "file scope" flag. Moved to third argument
9258 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9259 instead of RAND_bytes for encryption IVs and salts.
9262 *) Include RAND_status() into RAND_METHOD instead of implementing
9263 it only for md_rand.c Otherwise replacing the PRNG by calling
9264 RAND_set_rand_method would be impossible.
9267 *) Don't let DSA_generate_key() enter an infinite loop if the random
9268 number generation fails.
9271 *) New 'rand' application for creating pseudo-random output.
9274 *) Added configuration support for Linux/IA64
9275 [Rolf Haberrecker <rolf@suse.de>]
9277 *) Assembler module support for Mingw32.
9280 *) Shared library support for HPUX (in shlib/).
9281 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9283 *) Shared library support for Solaris gcc.
9284 [Lutz Behnke <behnke@trustcenter.de>]
9286 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9288 *) PKCS7_encrypt() was adding text MIME headers twice because they
9289 were added manually and by SMIME_crlf_copy().
9292 *) In bntest.c don't call BN_rand with zero bits argument.
9293 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9295 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9296 case was implemented. This caused BN_div_recp() to fail occasionally.
9299 *) Add an optional second argument to the set_label() in the perl
9300 assembly language builder. If this argument exists and is set
9301 to 1 it signals that the assembler should use a symbol whose
9302 scope is the entire file, not just the current function. This
9303 is needed with MASM which uses the format label:: for this scope.
9304 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9306 *) Change the ASN1 types so they are typedefs by default. Before
9307 almost all types were #define'd to ASN1_STRING which was causing
9308 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9312 *) Change names of new functions to the new get1/get0 naming
9313 convention: After 'get1', the caller owns a reference count
9314 and has to call ..._free; 'get0' returns a pointer to some
9315 data structure without incrementing reference counters.
9316 (Some of the existing 'get' functions increment a reference
9317 counter, some don't.)
9318 Similarly, 'set1' and 'add1' functions increase reference
9319 counters or duplicate objects.
9322 *) Allow for the possibility of temp RSA key generation failure:
9323 the code used to assume it always worked and crashed on failure.
9326 *) Fix potential buffer overrun problem in BIO_printf().
9327 [Ulf Möller, using public domain code by Patrick Powell; problem
9328 pointed out by David Sacerdote <das33@cornell.edu>]
9330 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9331 RAND_egd() and RAND_status(). In the command line application,
9332 the EGD socket can be specified like a seed file using RANDFILE
9336 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9337 Some CAs (e.g. Verisign) distribute certificates in this form.
9340 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9341 list to exclude them. This means that no special compilation option
9342 is needed to use anonymous DH: it just needs to be included in the
9346 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9347 EVP_MD_type. The old functionality is available in a new macro called
9348 EVP_MD_md(). Change code that uses it and update docs.
9351 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9352 where the 'void *' argument is replaced by a function pointer argument.
9353 Previously 'void *' was abused to point to functions, which works on
9354 many platforms, but is not correct. As these functions are usually
9355 called by macros defined in OpenSSL header files, most source code
9356 should work without changes.
9359 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9360 sections with information on -D... compiler switches used for
9361 compiling the library so that applications can see them. To enable
9362 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9363 must be defined. E.g.,
9364 #define OPENSSL_ALGORITHM_DEFINES
9365 #include <openssl/opensslconf.h>
9366 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9367 [Richard Levitte, Ulf and Bodo Möller]
9369 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9373 *) Change the 'other' type in certificate aux info to a STACK_OF
9374 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9375 the required ASN1 format: arbitrary types determined by an OID.
9378 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9379 argument to 'req'. This is not because the function is newer or
9380 better than others it just uses the work 'NEW' in the certificate
9381 request header lines. Some software needs this.
9384 *) Reorganise password command line arguments: now passwords can be
9385 obtained from various sources. Delete the PEM_cb function and make
9386 it the default behaviour: i.e. if the callback is NULL and the
9387 usrdata argument is not NULL interpret it as a null terminated pass
9388 phrase. If usrdata and the callback are NULL then the pass phrase
9389 is prompted for as usual.
9392 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9393 the support is automatically enabled. The resulting binaries will
9394 autodetect the card and use it if present.
9395 [Ben Laurie and Compaq Inc.]
9397 *) Work around for Netscape hang bug. This sends certificate request
9398 and server done in one record. Since this is perfectly legal in the
9399 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9400 the bugs/SSLv3 entry for more info.
9403 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9406 *) Add -rand argument to smime and pkcs12 applications and read/write
9410 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9413 *) Add command line password options to the remaining applications.
9416 *) Bug fix for BN_div_recp() for numerators with an even number of
9420 *) More tests in bntest.c, and changed test_bn output.
9423 *) ./config recognizes MacOS X now.
9426 *) Bug fix for BN_div() when the first words of num and divsor are
9427 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9430 *) Add support for various broken PKCS#8 formats, and command line
9431 options to produce them.
9434 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9435 get temporary BIGNUMs from a BN_CTX.
9438 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9442 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9443 include a #define from the old name to the new. The original intent
9444 was that statically linked binaries could for example just call
9445 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9446 link with digests. This never worked because SSLeay_add_all_digests()
9447 and SSLeay_add_all_ciphers() were in the same source file so calling
9448 one would link with the other. They are now in separate source files.
9451 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9454 *) Use a less unusual form of the Miller-Rabin primality test (it used
9455 a binary algorithm for exponentiation integrated into the Miller-Rabin
9456 loop, our standard modexp algorithms are faster).
9459 *) Support for the EBCDIC character set completed.
9460 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9462 *) Source code cleanups: use const where appropriate, eliminate casts,
9463 use void * instead of char * in lhash.
9466 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9467 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9468 this the server could overwrite ephemeral keys that the client
9472 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9473 using 50 iterations of the Rabin-Miller test.
9475 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9476 iterations of the Rabin-Miller test as required by the appendix
9477 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9478 As BN_is_prime_fasttest includes trial division, DSA parameter
9479 generation becomes much faster.
9481 This implies a change for the callback functions in DSA_is_prime
9482 and DSA_generate_parameters: The callback function is called once
9483 for each positive witness in the Rabin-Miller test, not just
9484 occasionally in the inner loop; and the parameters to the
9485 callback function now provide an iteration count for the outer
9486 loop rather than for the current invocation of the inner loop.
9487 DSA_generate_parameters additionally can call the callback
9488 function with an 'iteration count' of -1, meaning that a
9489 candidate has passed the trial division test (when q is generated
9490 from an application-provided seed, trial division is skipped).
9493 *) New function BN_is_prime_fasttest that optionally does trial
9494 division before starting the Rabin-Miller test and has
9495 an additional BN_CTX * argument (whereas BN_is_prime always
9496 has to allocate at least one BN_CTX).
9497 'callback(1, -1, cb_arg)' is called when a number has passed the
9498 trial division stage.
9501 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9505 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9508 *) New function BN_pseudo_rand().
9511 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9512 bignum version of BN_from_montgomery() with the working code from
9513 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9517 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9518 made it impossible to use the same SSL_SESSION data structure in
9519 SSL2 clients in multiple threads.
9522 *) The return value of RAND_load_file() no longer counts bytes obtained
9523 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9524 to seed the PRNG (previously an explicit byte count was required).
9525 [Ulf Möller, Bodo Möller]
9527 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9528 used (char *) instead of (void *) and had casts all over the place.
9531 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9534 *) Retain source code compatibility for BN_prime_checks macro:
9535 BN_is_prime(..., BN_prime_checks, ...) now uses
9536 BN_prime_checks_for_size to determine the appropriate number of
9537 Rabin-Miller iterations.
9540 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9541 DH_CHECK_P_NOT_SAFE_PRIME.
9542 (Check if this is true? OpenPGP calls them "strong".)
9545 *) Merge the functionality of "dh" and "gendh" programs into a new program
9546 "dhparam". The old programs are retained for now but will handle DH keys
9547 (instead of parameters) in future.
9550 *) Make the ciphers, s_server and s_client programs check the return values
9551 when a new cipher list is set.
9554 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9555 ciphers. Before when the 56bit ciphers were enabled the sorting was
9558 The syntax for the cipher sorting has been extended to support sorting by
9559 cipher-strength (using the strength_bits hard coded in the tables).
9560 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9562 Fix a bug in the cipher-command parser: when supplying a cipher command
9563 string with an "undefined" symbol (neither command nor alphanumeric
9564 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9565 an error is flagged.
9567 Due to the strength-sorting extension, the code of the
9568 ssl_create_cipher_list() function was completely rearranged. I hope that
9569 the readability was also increased :-)
9570 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9572 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9573 for the first serial number and places 2 in the serial number file. This
9574 avoids problems when the root CA is created with serial number zero and
9575 the first user certificate has the same issuer name and serial number
9579 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9580 the new code. Add documentation for this stuff.
9583 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9584 X509_*() to X509at_*() on the grounds that they don't handle X509
9585 structures and behave in an analogous way to the X509v3 functions:
9586 they shouldn't be called directly but wrapper functions should be used
9589 So we also now have some wrapper functions that call the X509at functions
9590 when passed certificate requests. (TO DO: similar things can be done with
9591 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9592 things. Some of these need some d2i or i2d and print functionality
9593 because they handle more complex structures.)
9596 *) Add missing #ifndefs that caused missing symbols when building libssl
9597 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9598 NO_RSA in ssl/s2*.c.
9599 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9601 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9602 has a return value which indicates the quality of the random data
9603 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9604 error queue. New function RAND_pseudo_bytes() generates output that is
9605 guaranteed to be unique but not unpredictable. RAND_add is like
9606 RAND_seed, but takes an extra argument for an entropy estimate
9607 (RAND_seed always assumes full entropy).
9610 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9611 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9612 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9613 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9614 false-positive rate of at most 2^-80 for random input.
9617 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9620 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9621 in the 0.9.5 release), this returns the chain
9622 from an X509_CTX structure with a dup of the stack and all
9623 the X509 reference counts upped: so the stack will exist
9624 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9627 Also make SSL_SESSION_print() print out the verify return
9631 *) Add manpage for the pkcs12 command. Also change the default
9632 behaviour so MAC iteration counts are used unless the new
9633 -nomaciter option is used. This improves file security and
9634 only older versions of MSIE (4.0 for example) need it.
9637 *) Honor the no-xxx Configure options when creating .DEF files.
9640 *) Add PKCS#10 attributes to field table: challengePassword,
9641 unstructuredName and unstructuredAddress. These are taken from
9642 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9643 international characters are used.
9645 More changes to X509_ATTRIBUTE code: allow the setting of types
9646 based on strings. Remove the 'loc' parameter when adding
9647 attributes because these will be a SET OF encoding which is sorted
9651 *) Initial changes to the 'req' utility to allow request generation
9652 automation. This will allow an application to just generate a template
9653 file containing all the field values and have req construct the
9656 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9657 used all over the place including certificate requests and PKCS#7
9658 structures. They are currently handled manually where necessary with
9659 some primitive wrappers for PKCS#7. The new functions behave in a
9660 manner analogous to the X509 extension functions: they allow
9661 attributes to be looked up by NID and added.
9663 Later something similar to the X509V3 code would be desirable to
9664 automatically handle the encoding, decoding and printing of the
9665 more complex types. The string types like challengePassword can
9666 be handled by the string table functions.
9668 Also modified the multi byte string table handling. Now there is
9669 a 'global mask' which masks out certain types. The table itself
9670 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9671 is useful when for example there is only one permissible type
9672 (as in countryName) and using the mask might result in no valid
9676 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
9677 SSL_get_peer_finished to allow applications to obtain the latest
9678 Finished messages sent to the peer or expected from the peer,
9679 respectively. (SSL_get_peer_finished is usually the Finished message
9680 actually received from the peer, otherwise the protocol will be aborted.)
9682 As the Finished message are message digests of the complete handshake
9683 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9684 be used for external authentication procedures when the authentication
9685 provided by SSL/TLS is not desired or is not enough.
9688 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9689 the host supports BWX extension and if Compaq C is present on the
9690 $PATH. Just exploiting of the BWX extension results in 20-30%
9691 performance kick for some algorithms, e.g. DES and RC4 to mention
9692 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9696 *) Add support for MS "fast SGC". This is arguably a violation of the
9697 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9698 weak crypto and after checking the certificate is SGC a second one
9699 with strong crypto. MS SGC stops the first handshake after receiving
9700 the server certificate message and sends a second client hello. Since
9701 a server will typically do all the time consuming operations before
9702 expecting any further messages from the client (server key exchange
9703 is the most expensive) there is little difference between the two.
9705 To get OpenSSL to support MS SGC we have to permit a second client
9706 hello message after we have sent server done. In addition we have to
9707 reset the MAC if we do get this second client hello.
9710 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9711 if a DER encoded private key is RSA or DSA traditional format. Changed
9712 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9713 format DER encoded private key. Newer code should use PKCS#8 format which
9714 has the key type encoded in the ASN1 structure. Added DER private key
9715 support to pkcs8 application.
9718 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9719 ciphersuites has been selected (as required by the SSL 3/TLS 1
9720 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9721 is set, we interpret this as a request to violate the specification
9722 (the worst that can happen is a handshake failure, and 'correct'
9723 behaviour would result in a handshake failure anyway).
9726 *) In SSL_CTX_add_session, take into account that there might be multiple
9727 SSL_SESSION structures with the same session ID (e.g. when two threads
9728 concurrently obtain them from an external cache).
9729 The internal cache can handle only one SSL_SESSION with a given ID,
9730 so if there's a conflict, we now throw out the old one to achieve
9734 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9735 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9736 some routines that use cipher OIDs: some ciphers do not have OIDs
9737 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9741 *) Simplify the trust setting structure and code. Now we just have
9742 two sequences of OIDs for trusted and rejected settings. These will
9743 typically have values the same as the extended key usage extension
9744 and any application specific purposes.
9746 The trust checking code now has a default behaviour: it will just
9747 check for an object with the same NID as the passed id. Functions can
9748 be provided to override either the default behaviour or the behaviour
9749 for a given id. SSL client, server and email already have functions
9750 in place for compatibility: they check the NID and also return "trusted"
9751 if the certificate is self signed.
9754 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9755 traditional format into an EVP_PKEY structure.
9758 *) Add a password callback function PEM_cb() which either prompts for
9759 a password if usr_data is NULL or otherwise assumes it is a null
9760 terminated password. Allow passwords to be passed on command line
9761 environment or config files in a few more utilities.
9764 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9765 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9766 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9767 Update documentation.
9770 *) Support for ASN1 "NULL" type. This could be handled before by using
9771 ASN1_TYPE but there wasn't any function that would try to read a NULL
9772 and produce an error if it couldn't. For compatibility we also have
9773 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9774 don't allocate anything because they don't need to.
9777 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9779 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9781 *) Rebuild of the memory allocation routines used by OpenSSL code and
9782 possibly others as well. The purpose is to make an interface that
9783 provide hooks so anyone can build a separate set of allocation and
9784 deallocation routines to be used by OpenSSL, for example memory
9785 pool implementations, or something else, which was previously hard
9786 since Malloc(), Realloc() and Free() were defined as macros having
9787 the values malloc, realloc and free, respectively (except for Win32
9788 compilations). The same is provided for memory debugging code.
9789 OpenSSL already comes with functionality to find memory leaks, but
9790 this gives people a chance to debug other memory problems.
9792 With these changes, a new set of functions and macros have appeared:
9794 CRYPTO_set_mem_debug_functions() [F]
9795 CRYPTO_get_mem_debug_functions() [F]
9796 CRYPTO_dbg_set_options() [F]
9797 CRYPTO_dbg_get_options() [F]
9798 CRYPTO_malloc_debug_init() [M]
9800 The memory debug functions are NULL by default, unless the library
9801 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9802 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9803 gives the standard debugging functions that come with OpenSSL) or
9804 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9805 provided by the library user) must be used. When the standard
9806 debugging functions are used, CRYPTO_dbg_set_options can be used to
9807 request additional information:
9808 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9809 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9811 Also, things like CRYPTO_set_mem_functions will always give the
9812 expected result (the new set of functions is used for allocation
9813 and deallocation) at all times, regardless of platform and compiler
9816 To finish it up, some functions that were never use in any other
9817 way than through macros have a new API and new semantic:
9820 CRYPTO_dbg_realloc()
9823 All macros of value have retained their old syntax.
9824 [Richard Levitte and Bodo Moeller]
9826 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9827 ordering of SMIMECapabilities wasn't in "strength order" and there
9828 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9832 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9833 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9834 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9836 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9837 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9838 functionality to handle multipart/signed properly) and a utility
9839 called 'smime' to call all this stuff. This is based on code I
9840 originally wrote for Celo who have kindly allowed it to be
9841 included in OpenSSL.
9844 *) Add variants des_set_key_checked and des_set_key_unchecked of
9845 des_set_key (aka des_key_sched). Global variable des_check_key
9846 decides which of these is called by des_set_key; this way
9847 des_check_key behaves as it always did, but applications and
9848 the library itself, which was buggy for des_check_key == 1,
9849 have a cleaner way to pick the version they need.
9852 *) New function PKCS12_newpass() which changes the password of a
9856 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9857 dynamic mix. In both cases the ids can be used as an index into the
9858 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9859 functions so they accept a list of the field values and the
9860 application doesn't need to directly manipulate the X509_TRUST
9864 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9868 *) Modify the way the V3 extension code looks up extensions. This now
9869 works in a similar way to the object code: we have some "standard"
9870 extensions in a static table which is searched with OBJ_bsearch()
9871 and the application can add dynamic ones if needed. The file
9872 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9873 updated whenever a new extension is added to the core code and kept
9874 in ext_nid order. There is a simple program 'tabtest.c' which checks
9875 this. New extensions are not added too often so this file can readily
9876 be maintained manually.
9878 There are two big advantages in doing things this way. The extensions
9879 can be looked up immediately and no longer need to be "added" using
9880 X509V3_add_standard_extensions(): this function now does nothing.
9881 [Side note: I get *lots* of email saying the extension code doesn't
9882 work because people forget to call this function]
9883 Also no dynamic allocation is done unless new extensions are added:
9884 so if we don't add custom extensions there is no need to call
9885 X509V3_EXT_cleanup().
9888 *) Modify enc utility's salting as follows: make salting the default. Add a
9889 magic header, so unsalted files fail gracefully instead of just decrypting
9890 to garbage. This is because not salting is a big security hole, so people
9891 should be discouraged from doing it.
9894 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9895 digest to be passed on the command line but it only used this
9896 parameter when signing a certificate. Modified so all relevant
9897 operations are affected by the digest parameter including the
9898 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9899 DSA key was used because it didn't fix the digest.
9902 *) Initial certificate chain verify code. Currently tests the untrusted
9903 certificates for consistency with the verify purpose (which is set
9904 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9906 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9907 this is because it will reject chains with invalid extensions whereas
9908 every previous version of OpenSSL and SSLeay made no checks at all.
9910 Trust code: checks the root CA for the relevant trust settings. Trust
9911 settings have an initial value consistent with the verify purpose: e.g.
9912 if the verify purpose is for SSL client use it expects the CA to be
9913 trusted for SSL client use. However the default value can be changed to
9914 permit custom trust settings: one example of this would be to only trust
9915 certificates from a specific "secure" set of CAs.
9917 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9918 which should be used for version portability: especially since the
9919 verify structure is likely to change more often now.
9921 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9922 to set them. If not set then assume SSL clients will verify SSL servers
9925 Two new options to the verify program: -untrusted allows a set of
9926 untrusted certificates to be passed in and -purpose which sets the
9927 intended purpose of the certificate. If a purpose is set then the
9928 new chain verify code is used to check extension consistency.
9931 *) Support for the authority information access extension.
9934 *) Modify RSA and DSA PEM read routines to transparently handle
9935 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9936 public keys in a format compatible with certificate
9937 SubjectPublicKeyInfo structures. Unfortunately there were already
9938 functions called *_PublicKey_* which used various odd formats so
9939 these are retained for compatibility: however the DSA variants were
9940 never in a public release so they have been deleted. Changed dsa/rsa
9941 utilities to handle the new format: note no releases ever handled public
9942 keys so we should be OK.
9944 The primary motivation for this change is to avoid the same fiasco
9945 that dogs private keys: there are several incompatible private key
9946 formats some of which are standard and some OpenSSL specific and
9947 require various evil hacks to allow partial transparent handling and
9948 even then it doesn't work with DER formats. Given the option anything
9949 other than PKCS#8 should be dumped: but the other formats have to
9950 stay in the name of compatibility.
9952 With public keys and the benefit of hindsight one standard format
9953 is used which works with EVP_PKEY, RSA or DSA structures: though
9954 it clearly returns an error if you try to read the wrong kind of key.
9956 Added a -pubkey option to the 'x509' utility to output the public key.
9957 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9958 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9959 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9960 that do the same as the EVP_PKEY_assign_*() except they up the
9961 reference count of the added key (they don't "swallow" the
9965 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9966 CRLs would fail if the file contained no certificates or no CRLs:
9967 added a new function to read in both types and return the number
9968 read: this means that if none are read it will be an error. The
9969 DER versions of the certificate and CRL reader would always fail
9970 because it isn't possible to mix certificates and CRLs in DER format
9971 without choking one or the other routine. Changed this to just read
9972 a certificate: this is the best we can do. Also modified the code
9973 in apps/verify.c to take notice of return codes: it was previously
9974 attempting to read in certificates from NULL pointers and ignoring
9975 any errors: this is one reason why the cert and CRL reader seemed
9976 to work. It doesn't check return codes from the default certificate
9977 routines: these may well fail if the certificates aren't installed.
9980 *) Code to support otherName option in GeneralName.
9983 *) First update to verify code. Change the verify utility
9984 so it warns if it is passed a self signed certificate:
9985 for consistency with the normal behaviour. X509_verify
9986 has been modified to it will now verify a self signed
9987 certificate if *exactly* the same certificate appears
9988 in the store: it was previously impossible to trust a
9989 single self signed certificate. This means that:
9990 openssl verify ss.pem
9991 now gives a warning about a self signed certificate but
9992 openssl verify -CAfile ss.pem ss.pem
9996 *) For servers, store verify_result in SSL_SESSION data structure
9997 (and add it to external session representation).
9998 This is needed when client certificate verifications fails,
9999 but an application-provided verification callback (set by
10000 SSL_CTX_set_cert_verify_callback) allows accepting the session
10001 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10002 but returns 1): When the session is reused, we have to set
10003 ssl->verify_result to the appropriate error code to avoid
10005 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10007 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10008 case in PKCS7_dataInit() where the signed PKCS7 structure
10009 didn't contain any existing data because it was being created.
10010 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10012 *) Add a salt to the key derivation routines in enc.c. This
10013 forms the first 8 bytes of the encrypted file. Also add a
10014 -S option to allow a salt to be input on the command line.
10017 *) New function X509_cmp(). Oddly enough there wasn't a function
10018 to compare two certificates. We do this by working out the SHA1
10019 hash and comparing that. X509_cmp() will be needed by the trust
10023 *) SSL_get1_session() is like SSL_get_session(), but increments
10024 the reference count in the SSL_SESSION returned.
10025 [Geoff Thorpe <geoff@eu.c2.net>]
10027 *) Fix for 'req': it was adding a null to request attributes.
10028 Also change the X509_LOOKUP and X509_INFO code to handle
10029 certificate auxiliary information.
10032 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10036 *) Add the possibility to add extra information to the memory leak
10037 detecting output, to form tracebacks, showing from where each
10038 allocation was originated: CRYPTO_push_info("constant string") adds
10039 the string plus current file name and line number to a per-thread
10040 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10041 is like calling CYRPTO_pop_info() until the stack is empty.
10042 Also updated memory leak detection code to be multi-thread-safe.
10045 *) Add options -text and -noout to pkcs7 utility and delete the
10046 encryption options which never did anything. Update docs.
10049 *) Add options to some of the utilities to allow the pass phrase
10050 to be included on either the command line (not recommended on
10051 OSes like Unix) or read from the environment. Update the
10052 manpages and fix a few bugs.
10055 *) Add a few manpages for some of the openssl commands.
10058 *) Fix the -revoke option in ca. It was freeing up memory twice,
10059 leaking and not finding already revoked certificates.
10062 *) Extensive changes to support certificate auxiliary information.
10063 This involves the use of X509_CERT_AUX structure and X509_AUX
10064 functions. An X509_AUX function such as PEM_read_X509_AUX()
10065 can still read in a certificate file in the usual way but it
10066 will also read in any additional "auxiliary information". By
10067 doing things this way a fair degree of compatibility can be
10068 retained: existing certificates can have this information added
10069 using the new 'x509' options.
10071 Current auxiliary information includes an "alias" and some trust
10072 settings. The trust settings will ultimately be used in enhanced
10073 certificate chain verification routines: currently a certificate
10074 can only be trusted if it is self signed and then it is trusted
10078 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10079 The problem was that one of the replacement routines had not been working
10080 since SSLeay releases. For now the offending routine has been replaced
10081 with non-optimised assembler. Even so, this now gives around 95%
10082 performance improvement for 1024 bit RSA signs.
10085 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10086 handling. Most clients have the effective key size in bits equal to
10087 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10088 A few however don't do this and instead use the size of the decrypted key
10089 to determine the RC2 key length and the AlgorithmIdentifier to determine
10090 the effective key length. In this case the effective key length can still
10091 be 40 bits but the key length can be 168 bits for example. This is fixed
10092 by manually forcing an RC2 key into the EVP_PKEY structure because the
10093 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10094 the key length and effective key length are equal.
10097 *) Add a bunch of functions that should simplify the creation of
10098 X509_NAME structures. Now you should be able to do:
10099 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10100 and have it automatically work out the correct field type and fill in
10101 the structures. The more adventurous can try:
10102 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10103 and it will (hopefully) work out the correct multibyte encoding.
10106 *) Change the 'req' utility to use the new field handling and multibyte
10107 copy routines. Before the DN field creation was handled in an ad hoc
10108 way in req, ca, and x509 which was rather broken and didn't support
10109 BMPStrings or UTF8Strings. Since some software doesn't implement
10110 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10111 using the dirstring_type option. See the new comment in the default
10112 openssl.cnf for more info.
10115 *) Make crypto/rand/md_rand.c more robust:
10116 - Assure unique random numbers after fork().
10117 - Make sure that concurrent threads access the global counter and
10118 md serializably so that we never lose entropy in them
10119 or use exactly the same state in multiple threads.
10120 Access to the large state is not always serializable because
10121 the additional locking could be a performance killer, and
10122 md should be large enough anyway.
10125 *) New file apps/app_rand.c with commonly needed functionality
10126 for handling the random seed file.
10128 Use the random seed file in some applications that previously did not:
10130 dsaparam -genkey (which also ignored its '-rand' option),
10133 x509 (when signing).
10134 Except on systems with /dev/urandom, it is crucial to have a random
10135 seed file at least for key creation, DSA signing, and for DH exchanges;
10136 for RSA signatures we could do without one.
10138 gendh and gendsa (unlike genrsa) used to read only the first byte
10139 of each file listed in the '-rand' option. The function as previously
10140 found in genrsa is now in app_rand.c and is used by all programs
10141 that support '-rand'.
10144 *) In RAND_write_file, use mode 0600 for creating files;
10145 don't just chmod when it may be too late.
10148 *) Report an error from X509_STORE_load_locations
10149 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10152 *) New function ASN1_mbstring_copy() this copies a string in either
10153 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10154 into an ASN1_STRING type. A mask of permissible types is passed
10155 and it chooses the "minimal" type to use or an error if not type
10159 *) Add function equivalents to the various macros in asn1.h. The old
10160 macros are retained with an M_ prefix. Code inside the library can
10161 use the M_ macros. External code (including the openssl utility)
10162 should *NOT* in order to be "shared library friendly".
10165 *) Add various functions that can check a certificate's extensions
10166 to see if it usable for various purposes such as SSL client,
10167 server or S/MIME and CAs of these types. This is currently
10168 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10169 verification. Also added a -purpose flag to x509 utility to
10170 print out all the purposes.
10173 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10177 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10178 for, obtain and decode and extension and obtain its critical flag.
10179 This allows all the necessary extension code to be handled in a
10180 single function call.
10183 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10184 platforms. See crypto/rc4/rc4_enc.c for further details.
10187 *) New -noout option to asn1parse. This causes no output to be produced
10188 its main use is when combined with -strparse and -out to extract data
10189 from a file (which may not be in ASN.1 format).
10192 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10193 when producing the local key id.
10194 [Richard Levitte <levitte@stacken.kth.se>]
10196 *) New option -dhparam in s_server. This allows a DH parameter file to be
10197 stated explicitly. If it is not stated then it tries the first server
10198 certificate file. The previous behaviour hard coded the filename
10202 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10203 a public key to be input or output. For example:
10204 openssl rsa -in key.pem -pubout -out pubkey.pem
10205 Also added necessary DSA public key functions to handle this.
10208 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10209 in the message. This was handled by allowing
10210 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10211 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10213 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10214 to the end of the strings whereas this didn't. This would cause problems
10215 if strings read with d2i_ASN1_bytes() were later modified.
10216 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10218 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10219 data and it contains EOF it will end up returning an error. This is
10220 caused by input 46 bytes long. The cause is due to the way base64
10221 BIOs find the start of base64 encoded data. They do this by trying a
10222 trial decode on each line until they find one that works. When they
10223 do a flag is set and it starts again knowing it can pass all the
10224 data directly through the decoder. Unfortunately it doesn't reset
10225 the context it uses. This means that if EOF is reached an attempt
10226 is made to pass two EOFs through the context and this causes the
10227 resulting error. This can also cause other problems as well. As is
10228 usual with these problems it takes *ages* to find and the fix is
10229 trivial: move one line.
10230 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10232 *) Ugly workaround to get s_client and s_server working under Windows. The
10233 old code wouldn't work because it needed to select() on sockets and the
10234 tty (for keypresses and to see if data could be written). Win32 only
10235 supports select() on sockets so we select() with a 1s timeout on the
10236 sockets and then see if any characters are waiting to be read, if none
10237 are present then we retry, we also assume we can always write data to
10238 the tty. This isn't nice because the code then blocks until we've
10239 received a complete line of data and it is effectively polling the
10240 keyboard at 1s intervals: however it's quite a bit better than not
10241 working at all :-) A dedicated Windows application might handle this
10242 with an event loop for example.
10245 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10246 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10247 will be called when RSA_sign() and RSA_verify() are used. This is useful
10248 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10249 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10250 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10251 This necessitated the support of an extra signature type NID_md5_sha1
10252 for SSL signatures and modifications to the SSL library to use it instead
10253 of calling RSA_public_decrypt() and RSA_private_encrypt().
10256 *) Add new -verify -CAfile and -CApath options to the crl program, these
10257 will lookup a CRL issuers certificate and verify the signature in a
10258 similar way to the verify program. Tidy up the crl program so it
10259 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10260 less strict. It will now permit CRL extensions even if it is not
10261 a V2 CRL: this will allow it to tolerate some broken CRLs.
10264 *) Initialize all non-automatic variables each time one of the openssl
10265 sub-programs is started (this is necessary as they may be started
10266 multiple times from the "OpenSSL>" prompt).
10267 [Lennart Bang, Bodo Moeller]
10269 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10270 removing all other RSA functionality (this is what NO_RSA does). This
10271 is so (for example) those in the US can disable those operations covered
10272 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10276 *) Non-copying interface to BIO pairs.
10277 (still largely untested)
10280 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10281 ASCII string. This was handled independently in various places before.
10284 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10285 UTF8 strings a character at a time.
10288 *) Use client_version from client hello to select the protocol
10289 (s23_srvr.c) and for RSA client key exchange verification
10290 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10293 *) Add various utility functions to handle SPKACs, these were previously
10294 handled by poking round in the structure internals. Added new function
10295 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10296 print, verify and generate SPKACs. Based on an original idea from
10297 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10300 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10303 *) Allow the config file extension section to be overwritten on the
10304 command line. Based on an original idea from Massimiliano Pala
10305 <madwolf@comune.modena.it>. The new option is called -extensions
10306 and can be applied to ca, req and x509. Also -reqexts to override
10307 the request extensions in req and -crlexts to override the crl extensions
10311 *) Add new feature to the SPKAC handling in ca. Now you can include
10312 the same field multiple times by preceding it by "XXXX." for example:
10315 this is the same syntax as used in the req config file.
10318 *) Allow certificate extensions to be added to certificate requests. These
10319 are specified in a 'req_extensions' option of the req section of the
10320 config file. They can be printed out with the -text option to req but
10321 are otherwise ignored at present.
10324 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10325 data read consists of only the final block it would not decrypted because
10326 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10327 A misplaced 'break' also meant the decrypted final block might not be
10328 copied until the next read.
10331 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10332 a few extra parameters to the DH structure: these will be useful if
10333 for example we want the value of 'q' or implement X9.42 DH.
10336 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10337 provides hooks that allow the default DSA functions or functions on a
10338 "per key" basis to be replaced. This allows hardware acceleration and
10339 hardware key storage to be handled without major modification to the
10340 library. Also added low level modexp hooks and CRYPTO_EX structure and
10341 associated functions.
10344 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10345 as "read only": it can't be written to and the buffer it points to will
10346 not be freed. Reading from a read only BIO is much more efficient than
10347 a normal memory BIO. This was added because there are several times when
10348 an area of memory needs to be read from a BIO. The previous method was
10349 to create a memory BIO and write the data to it, this results in two
10350 copies of the data and an O(n^2) reading algorithm. There is a new
10351 function BIO_new_mem_buf() which creates a read only memory BIO from
10352 an area of memory. Also modified the PKCS#7 routines to use read only
10356 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10357 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10358 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10359 but a retry condition occurred while trying to read the rest.
10362 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10363 NID_pkcs7_encrypted by default: this was wrong since this should almost
10364 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10365 the encrypted data type: this is a more sensible place to put it and it
10366 allows the PKCS#12 code to be tidied up that duplicated this
10370 *) Changed obj_dat.pl script so it takes its input and output files on
10371 the command line. This should avoid shell escape redirection problems
10375 *) Initial support for certificate extension requests, these are included
10376 in things like Xenroll certificate requests. Included functions to allow
10377 extensions to be obtained and added.
10380 *) -crlf option to s_client and s_server for sending newlines as
10381 CRLF (as required by many protocols).
10384 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10386 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10387 [Ralf S. Engelschall]
10389 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10390 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10392 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10396 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10397 DH parameters/keys (q is lost during that conversion, but the resulting
10398 DH parameters contain its length).
10400 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10401 much faster than DH_generate_parameters (which creates parameters
10402 where p = 2*q + 1), and also the smaller q makes DH computations
10403 much more efficient (160-bit exponentiation instead of 1024-bit
10404 exponentiation); so this provides a convenient way to support DHE
10405 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10406 utter importance to use
10407 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10409 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10410 when such DH parameters are used, because otherwise small subgroup
10411 attacks may become possible!
10414 *) Avoid memory leak in i2d_DHparams.
10417 *) Allow the -k option to be used more than once in the enc program:
10418 this allows the same encrypted message to be read by multiple recipients.
10421 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10422 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10423 it will always use the numerical form of the OID, even if it has a short
10427 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10428 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10429 otherwise bn_mod_exp was called. In the case of hardware keys for example
10430 no private key components need be present and it might store extra data
10431 in the RSA structure, which cannot be accessed from bn_mod_exp.
10432 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10433 private key operations.
10436 *) Added support for SPARC Linux.
10439 *) pem_password_cb function type incompatibly changed from
10440 typedef int pem_password_cb(char *buf, int size, int rwflag);
10442 ....(char *buf, int size, int rwflag, void *userdata);
10443 so that applications can pass data to their callbacks:
10444 The PEM[_ASN1]_{read,write}... functions and macros now take an
10445 additional void * argument, which is just handed through whenever
10446 the password callback is called.
10447 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10449 New function SSL_CTX_set_default_passwd_cb_userdata.
10451 Compatibility note: As many C implementations push function arguments
10452 onto the stack in reverse order, the new library version is likely to
10453 interoperate with programs that have been compiled with the old
10454 pem_password_cb definition (PEM_whatever takes some data that
10455 happens to be on the stack as its last argument, and the callback
10456 just ignores this garbage); but there is no guarantee whatsoever that
10459 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10460 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10461 problems not only on Windows, but also on some Unix platforms.
10462 To avoid problematic command lines, these definitions are now in an
10463 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10464 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10467 *) MIPS III/IV assembler module is reimplemented.
10470 *) More DES library cleanups: remove references to srand/rand and
10471 delete an unused file.
10474 *) Add support for the the free Netwide assembler (NASM) under Win32,
10475 since not many people have MASM (ml) and it can be hard to obtain.
10476 This is currently experimental but it seems to work OK and pass all
10477 the tests. Check out INSTALL.W32 for info.
10480 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10481 without temporary keys kept an extra copy of the server key,
10482 and connections with temporary keys did not free everything in case
10486 *) New function RSA_check_key and new openssl rsa option -check
10487 for verifying the consistency of RSA keys.
10488 [Ulf Moeller, Bodo Moeller]
10490 *) Various changes to make Win32 compile work:
10491 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10492 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10493 comparison" warnings.
10494 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10497 *) Add a debugging option to PKCS#5 v2 key generation function: when
10498 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10499 derived keys are printed to stderr.
10502 *) Copy the flags in ASN1_STRING_dup().
10503 [Roman E. Pavlov <pre@mo.msk.ru>]
10505 *) The x509 application mishandled signing requests containing DSA
10506 keys when the signing key was also DSA and the parameters didn't match.
10508 It was supposed to omit the parameters when they matched the signing key:
10509 the verifying software was then supposed to automatically use the CA's
10510 parameters if they were absent from the end user certificate.
10512 Omitting parameters is no longer recommended. The test was also
10513 the wrong way round! This was probably due to unusual behaviour in
10514 EVP_cmp_parameters() which returns 1 if the parameters match.
10515 This meant that parameters were omitted when they *didn't* match and
10516 the certificate was useless. Certificates signed with 'ca' didn't have
10518 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10520 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10521 The interface is as follows:
10522 Applications can use
10523 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10524 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10525 "off" is now the default.
10526 The library internally uses
10527 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10528 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10529 to disable memory-checking temporarily.
10531 Some inconsistent states that previously were possible (and were
10532 even the default) are now avoided.
10534 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10535 with each memory chunk allocated; this is occasionally more helpful
10536 than just having a counter.
10538 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10540 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10544 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10545 which largely parallels "options", but is for changing API behaviour,
10546 whereas "options" are about protocol behaviour.
10547 Initial "mode" flags are:
10549 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10550 a single record has been written.
10551 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10552 retries use the same buffer location.
10553 (But all of the contents must be
10557 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10560 *) Fix problems with no-hmac etc.
10561 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10563 *) New functions RSA_get_default_method(), RSA_set_method() and
10564 RSA_get_method(). These allows replacement of RSA_METHODs without having
10565 to mess around with the internals of an RSA structure.
10568 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10569 Also really enable memory leak checks in openssl.c and in some
10571 [Chad C. Mulligan, Bodo Moeller]
10573 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10574 up the length of negative integers. This has now been simplified to just
10575 store the length when it is first determined and use it later, rather
10576 than trying to keep track of where data is copied and updating it to
10578 [Steve Henson, reported by Brien Wheeler
10579 <bwheeler@authentica-security.com>]
10581 *) Add a new function PKCS7_signatureVerify. This allows the verification
10582 of a PKCS#7 signature but with the signing certificate passed to the
10583 function itself. This contrasts with PKCS7_dataVerify which assumes the
10584 certificate is present in the PKCS#7 structure. This isn't always the
10585 case: certificates can be omitted from a PKCS#7 structure and be
10586 distributed by "out of band" means (such as a certificate database).
10589 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10590 function prototypes in pem.h, also change util/mkdef.pl to add the
10591 necessary function names.
10594 *) mk1mf.pl (used by Windows builds) did not properly read the
10595 options set by Configure in the top level Makefile, and Configure
10596 was not even able to write more than one option correctly.
10597 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10600 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10601 file to be loaded from a BIO or FILE pointer. The BIO version will
10602 for example allow memory BIOs to contain config info.
10605 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10606 Whoever hopes to achieve shared-library compatibility across versions
10607 must use this, not the compile-time macro.
10608 (Exercise 0.9.4: Which is the minimum library version required by
10610 Note: All this applies only to multi-threaded programs, others don't
10614 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10615 through a BIO pair triggered the default case, i.e.
10616 SSLerr(...,SSL_R_UNKNOWN_STATE).
10619 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10620 can use the SSL library even if none of the specific BIOs is
10624 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10625 for the encoded length.
10626 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10628 *) Add initial documentation of the X509V3 functions.
10631 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10632 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10633 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10634 secure PKCS#8 private key format with a high iteration count.
10637 *) Fix determination of Perl interpreter: A perl or perl5
10638 _directory_ in $PATH was also accepted as the interpreter.
10639 [Ralf S. Engelschall]
10641 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10642 wrong with it but it was very old and did things like calling
10643 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10644 unusual formatting.
10647 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10648 to use the new extension code.
10651 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10652 with macros. This should make it easier to change their form, add extra
10653 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10657 *) Add to configuration table a new entry that can specify an alternative
10658 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10659 according to Mark Crispin <MRC@Panda.COM>.
10663 *) DES CBC did not update the IV. Weird.
10666 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10667 Changing the behaviour of the former might break existing programs --
10668 where IV updating is needed, des_ncbc_encrypt can be used.
10671 *) When bntest is run from "make test" it drives bc to check its
10672 calculations, as well as internally checking them. If an internal check
10673 fails, it needs to cause bc to give a non-zero result or make test carries
10674 on without noticing the failure. Fixed.
10677 *) DES library cleanups.
10680 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10681 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10682 ciphers. NOTE: although the key derivation function has been verified
10683 against some published test vectors it has not been extensively tested
10684 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10688 *) Instead of "mkdir -p", which is not fully portable, use new
10689 Perl script "util/mkdir-p.pl".
10692 *) Rewrite the way password based encryption (PBE) is handled. It used to
10693 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10694 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10695 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10696 the 'parameter' field of the AlgorithmIdentifier is passed to the
10697 underlying key generation function so it must do its own ASN1 parsing.
10698 This has also changed the EVP_PBE_CipherInit() function which now has a
10699 'parameter' argument instead of literal salt and iteration count values
10700 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10703 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10704 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10705 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10706 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10707 value was just used as a "magic string" and not used directly its
10708 value doesn't matter.
10711 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10715 *) "linux-sparc64" configuration (ultrapenguin).
10716 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10717 "linux-sparc" configuration.
10718 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10720 *) config now generates no-xxx options for missing ciphers.
10723 *) Support the EBCDIC character set (work in progress).
10724 File ebcdic.c not yet included because it has a different license.
10725 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10727 *) Support BS2000/OSD-POSIX.
10728 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10730 *) Make callbacks for key generation use void * instead of char *.
10733 *) Make S/MIME samples compile (not yet tested).
10736 *) Additional typesafe stacks.
10739 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10743 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10745 *) New configuration variant "sco5-gcc".
10747 *) Updated some demos.
10748 [Sean O Riordain, Wade Scholine]
10750 *) Add missing BIO_free at exit of pkcs12 application.
10753 *) Fix memory leak in conf.c.
10756 *) Updates for Win32 to assembler version of MD5.
10759 *) Set #! path to perl in apps/der_chop to where we found it
10760 instead of using a fixed path.
10763 *) SHA library changes for irix64-mips4-cc.
10766 *) Improvements for VMS support.
10770 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10772 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10773 This also avoids the problems with SC4.2 and unpatched SC5.
10774 [Andy Polyakov <appro@fy.chalmers.se>]
10776 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10777 These are required because of the typesafe stack would otherwise break
10778 existing code. If old code used a structure member which used to be STACK
10779 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10780 sk_num or sk_value it would produce an error because the num, data members
10781 are not present in STACK_OF. Now it just produces a warning. sk_set
10782 replaces the old method of assigning a value to sk_value
10783 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10784 that does this will no longer work (and should use sk_set instead) but
10785 this could be regarded as a "questionable" behaviour anyway.
10788 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10789 correctly handle encrypted S/MIME data.
10792 *) Change type of various DES function arguments from des_cblock
10793 (which means, in function argument declarations, pointer to char)
10794 to des_cblock * (meaning pointer to array with 8 char elements),
10795 which allows the compiler to do more typechecking; it was like
10796 that back in SSLeay, but with lots of ugly casts.
10798 Introduce new type const_des_cblock.
10801 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10802 problems: find RecipientInfo structure that matches recipient certificate
10803 and initialise the ASN1 structures properly based on passed cipher.
10806 *) Belatedly make the BN tests actually check the results.
10809 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10810 to and from BNs: it was completely broken. New compilation option
10811 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10812 key elements as negative integers.
10815 *) Reorganize and speed up MD5.
10816 [Andy Polyakov <appro@fy.chalmers.se>]
10819 [Richard Levitte <richard@levitte.org>]
10821 *) New option -out to asn1parse to allow the parsed structure to be
10822 output to a file. This is most useful when combined with the -strparse
10823 option to examine the output of things like OCTET STRINGS.
10826 *) Make SSL library a little more fool-proof by not requiring any longer
10827 that SSL_set_{accept,connect}_state be called before
10828 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10829 in many applications because usually everything *appeared* to work as
10830 intended anyway -- now it really works as intended).
10833 *) Move openssl.cnf out of lib/.
10836 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10837 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10838 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10839 [Ralf S. Engelschall]
10841 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10842 handle PKCS#7 enveloped data properly.
10843 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10845 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10846 copying pointers. The cert_st handling is changed by this in
10847 various ways (and thus what used to be known as ctx->default_cert
10848 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10849 any longer when s->cert does not give us what we need).
10850 ssl_cert_instantiate becomes obsolete by this change.
10851 As soon as we've got the new code right (possibly it already is?),
10852 we have solved a couple of bugs of the earlier code where s->cert
10853 was used as if it could not have been shared with other SSL structures.
10855 Note that using the SSL API in certain dirty ways now will result
10856 in different behaviour than observed with earlier library versions:
10857 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10858 does not influence s as it used to.
10860 In order to clean up things more thoroughly, inside SSL_SESSION
10861 we don't use CERT any longer, but a new structure SESS_CERT
10862 that holds per-session data (if available); currently, this is
10863 the peer's certificate chain and, for clients, the server's certificate
10864 and temporary key. CERT holds only those values that can have
10865 meaningful defaults in an SSL_CTX.
10868 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10869 from the internal representation. Various PKCS#7 fixes: remove some
10870 evil casts and set the enc_dig_alg field properly based on the signing
10874 *) Allow PKCS#12 password to be set from the command line or the
10875 environment. Let 'ca' get its config file name from the environment
10876 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10880 *) Allow certificate policies extension to use an IA5STRING for the
10881 organization field. This is contrary to the PKIX definition but
10882 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10886 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10887 without disallowing inline assembler and the like for non-pedantic builds.
10890 *) Support Borland C++ builder.
10891 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10893 *) Support Mingw32.
10896 *) SHA-1 cleanups and performance enhancements.
10897 [Andy Polyakov <appro@fy.chalmers.se>]
10899 *) Sparc v8plus assembler for the bignum library.
10900 [Andy Polyakov <appro@fy.chalmers.se>]
10902 *) Accept any -xxx and +xxx compiler options in Configure.
10905 *) Update HPUX configuration.
10908 *) Add missing sk_<type>_unshift() function to safestack.h
10909 [Ralf S. Engelschall]
10911 *) New function SSL_CTX_use_certificate_chain_file that sets the
10912 "extra_cert"s in addition to the certificate. (This makes sense
10913 only for "PEM" format files, as chains as a whole are not
10917 *) Support verify_depth from the SSL API.
10918 x509_vfy.c had what can be considered an off-by-one-error:
10919 Its depth (which was not part of the external interface)
10920 was actually counting the number of certificates in a chain;
10921 now it really counts the depth.
10924 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10925 instead of X509err, which often resulted in confusing error
10926 messages since the error codes are not globally unique
10927 (e.g. an alleged error in ssl3_accept when a certificate
10928 didn't match the private key).
10930 *) New function SSL_CTX_set_session_id_context that allows to set a default
10931 value (so that you don't need SSL_set_session_id_context for each
10932 connection using the SSL_CTX).
10935 *) OAEP decoding bug fix.
10938 *) Support INSTALL_PREFIX for package builders, as proposed by
10942 *) New Configure options "threads" and "no-threads". For systems
10943 where the proper compiler options are known (currently Solaris
10944 and Linux), "threads" is the default.
10947 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10950 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10951 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10952 such as /usr/local/bin.
10955 *) "make linux-shared" to build shared libraries.
10956 [Niels Poppe <niels@netbox.org>]
10958 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10961 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10962 extension adding in x509 utility.
10965 *) Remove NOPROTO sections and error code comments.
10968 *) Partial rewrite of the DEF file generator to now parse the ANSI
10972 *) New Configure options --prefix=DIR and --openssldir=DIR.
10975 *) Complete rewrite of the error code script(s). It is all now handled
10976 by one script at the top level which handles error code gathering,
10977 header rewriting and C source file generation. It should be much better
10978 than the old method: it now uses a modified version of Ulf's parser to
10979 read the ANSI prototypes in all header files (thus the old K&R definitions
10980 aren't needed for error creation any more) and do a better job of
10981 translating function codes into names. The old 'ASN1 error code imbedded
10982 in a comment' is no longer necessary and it doesn't use .err files which
10983 have now been deleted. Also the error code call doesn't have to appear all
10984 on one line (which resulted in some large lines...).
10987 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10990 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10991 0 (which usually indicates a closed connection), but continue reading.
10994 *) Fix some race conditions.
10997 *) Add support for CRL distribution points extension. Add Certificate
10998 Policies and CRL distribution points documentation.
11001 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11004 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11005 8 of keying material. Merlin has also confirmed interop with this fix
11006 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11007 [Merlin Hughes <merlin@baltimore.ie>]
11009 *) Fix lots of warnings.
11010 [Richard Levitte <levitte@stacken.kth.se>]
11012 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11013 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11014 [Richard Levitte <levitte@stacken.kth.se>]
11016 *) Fix problems with sizeof(long) == 8.
11017 [Andy Polyakov <appro@fy.chalmers.se>]
11019 *) Change functions to ANSI C.
11022 *) Fix typos in error codes.
11023 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11025 *) Remove defunct assembler files from Configure.
11028 *) SPARC v8 assembler BIGNUM implementation.
11029 [Andy Polyakov <appro@fy.chalmers.se>]
11031 *) Support for Certificate Policies extension: both print and set.
11032 Various additions to support the r2i method this uses.
11035 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11036 return a const string when you are expecting an allocated buffer.
11039 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11040 types DirectoryString and DisplayText.
11043 *) Add code to allow r2i extensions to access the configuration database,
11044 add an LHASH database driver and add several ctx helper functions.
11047 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11048 fail when they extended the size of a BIGNUM.
11051 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11052 support typesafe stack.
11055 *) Fix typo in SSL_[gs]et_options().
11056 [Nils Frostberg <nils@medcom.se>]
11058 *) Delete various functions and files that belonged to the (now obsolete)
11059 old X509V3 handling code.
11062 *) New Configure option "rsaref".
11065 *) Don't auto-generate pem.h.
11068 *) Introduce type-safe ASN.1 SETs.
11071 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11072 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11074 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11075 that links with OpenSSL (well at least cause lots of warnings), but fear
11076 not: the conversion is trivial, and it eliminates loads of evil casts. A
11077 few STACKed things have been converted already. Feel free to convert more.
11078 In the fullness of time, I'll do away with the STACK type altogether.
11081 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11082 specified in <certfile> by updating the entry in the index.txt file.
11083 This way one no longer has to edit the index.txt file manually for
11084 revoking a certificate. The -revoke option does the gory details now.
11085 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11087 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11088 `-text' option at all and this way the `-noout -text' combination was
11089 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11090 [Ralf S. Engelschall]
11092 *) Make sure a corresponding plain text error message exists for the
11093 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11094 verify callback function determined that a certificate was revoked.
11095 [Ralf S. Engelschall]
11097 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11098 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11099 all available cipers including rc5, which was forgotten until now.
11100 In order to let the testing shell script know which algorithms
11101 are available, a new (up to now undocumented) command
11102 "openssl list-cipher-commands" is used.
11105 *) Bugfix: s_client occasionally would sleep in select() when
11106 it should have checked SSL_pending() first.
11109 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11110 the raw DSA values prior to ASN.1 encoding.
11113 *) Tweaks to Configure
11114 [Niels Poppe <niels@netbox.org>]
11116 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11120 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11123 *) New config option to avoid instructions that are illegal on the 80386.
11124 The default code is faster, but requires at least a 486.
11127 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11128 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11129 same as SSL2_VERSION anyway.
11132 *) New "-showcerts" option for s_client.
11135 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11136 application. Various cleanups and fixes.
11139 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11140 modify error routines to work internally. Add error codes and PBE init
11141 to library startup routines.
11144 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11145 packing functions to asn1 and evp. Changed function names and error
11146 codes along the way.
11149 *) PKCS12 integration: and so it begins... First of several patches to
11150 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11151 objects to objects.h
11154 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11155 and display support for Thawte strong extranet extension.
11158 *) Add LinuxPPC support.
11159 [Jeff Dubrule <igor@pobox.org>]
11161 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11162 bn_div_words in alpha.s.
11163 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11165 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11166 OAEP isn't supported when OpenSSL is built with RSAref.
11167 [Ulf Moeller <ulf@fitug.de>]
11169 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11170 so they no longer are missing under -DNOPROTO.
11171 [Soren S. Jorvang <soren@t.dk>]
11174 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11176 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11177 doesn't work when the session is reused. Coming soon!
11180 *) Fix a security hole, that allows sessions to be reused in the wrong
11181 context thus bypassing client cert protection! All software that uses
11182 client certs and session caches in multiple contexts NEEDS PATCHING to
11183 allow session reuse! A fuller solution is in the works.
11184 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11186 *) Some more source tree cleanups (removed obsolete files
11187 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11188 permission on "config" script to be executable) and a fix for the INSTALL
11190 [Ulf Moeller <ulf@fitug.de>]
11192 *) Remove some legacy and erroneous uses of malloc, free instead of
11194 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11196 *) Make rsa_oaep_test return non-zero on error.
11197 [Ulf Moeller <ulf@fitug.de>]
11199 *) Add support for native Solaris shared libraries. Configure
11200 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11201 if someone would make that last step automatic.
11202 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11204 *) ctx_size was not built with the right compiler during "make links". Fixed.
11207 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11208 except NULL ciphers". This means the default cipher list will no longer
11209 enable NULL ciphers. They need to be specifically enabled e.g. with
11210 the string "DEFAULT:eNULL".
11213 *) Fix to RSA private encryption routines: if p < q then it would
11214 occasionally produce an invalid result. This will only happen with
11215 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11218 *) Be less restrictive and allow also `perl util/perlpath.pl
11219 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11220 because this way one can also use an interpreter named `perl5' (which is
11221 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11222 installed as `perl').
11223 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11225 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11226 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11228 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11229 advapi32.lib to Win32 build and change the pem test comparison
11230 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11231 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11232 and crypto/des/ede_cbcm_enc.c.
11235 *) DES quad checksum was broken on big-endian architectures. Fixed.
11238 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11239 Win32 test batch file so it (might) work again. The Win32 test batch file
11240 is horrible: I feel ill....
11243 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11244 in e_os.h. Audit of header files to check ANSI and non ANSI
11245 sections: 10 functions were absent from non ANSI section and not exported
11246 from Windows DLLs. Fixed up libeay.num for new functions.
11249 *) Make `openssl version' output lines consistent.
11250 [Ralf S. Engelschall]
11252 *) Fix Win32 symbol export lists for BIO functions: Added
11253 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11254 to ms/libeay{16,32}.def.
11255 [Ralf S. Engelschall]
11257 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11258 fine under Unix and passes some trivial tests I've now added. But the
11259 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11260 added to make sure no one expects that this stuff really works in the
11261 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11262 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11264 [Ralf S. Engelschall]
11266 *) Fix the generation of two part addresses in perl.
11267 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11269 *) Add config entry for Linux on MIPS.
11270 [John Tobey <jtobey@channel1.com>]
11272 *) Make links whenever Configure is run, unless we are on Windoze.
11275 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11276 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11280 *) Add a useful kludge to allow package maintainers to specify compiler and
11281 other platforms details on the command line without having to patch the
11282 Configure script everytime: One now can use ``perl Configure
11283 <id>:<details>'', i.e. platform ids are allowed to have details appended
11284 to them (separated by colons). This is treated as there would be a static
11285 pre-configured entry in Configure's %table under key <id> with value
11286 <details> and ``perl Configure <id>'' is called. So, when you want to
11287 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11288 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11289 now, which overrides the FreeBSD-elf entry on-the-fly.
11290 [Ralf S. Engelschall]
11292 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11295 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11296 on the `perl Configure ...' command line. This way one can compile
11297 OpenSSL libraries with Position Independent Code (PIC) which is needed
11298 for linking it into DSOs.
11299 [Ralf S. Engelschall]
11301 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11305 *) Cleaned up the LICENSE document: The official contact for any license
11306 questions now is the OpenSSL core team under openssl-core@openssl.org.
11307 And add a paragraph about the dual-license situation to make sure people
11308 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11309 to the OpenSSL toolkit.
11310 [Ralf S. Engelschall]
11312 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11313 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11314 Additionally cleaned up the `make links' target: Remove unnecessary
11315 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11316 to speed processing and no longer clutter the display with confusing
11317 stuff. Instead only the actually done links are displayed.
11318 [Ralf S. Engelschall]
11320 *) Permit null encryption ciphersuites, used for authentication only. It used
11321 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11322 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11326 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11327 signed attributes when verifying signatures (this would break them),
11328 the detached data encoding was wrong and public keys obtained using
11329 X509_get_pubkey() weren't freed.
11332 *) Add text documentation for the BUFFER functions. Also added a work around
11333 to a Win95 console bug. This was triggered by the password read stuff: the
11334 last character typed gets carried over to the next fread(). If you were
11335 generating a new cert request using 'req' for example then the last
11336 character of the passphrase would be CR which would then enter the first
11340 *) Added the new `Includes OpenSSL Cryptography Software' button as
11341 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11342 button and can be used by applications based on OpenSSL to show the
11343 relationship to the OpenSSL project.
11344 [Ralf S. Engelschall]
11346 *) Remove confusing variables in function signatures in files
11347 ssl/ssl_lib.c and ssl/ssl.h.
11348 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11350 *) Don't install bss_file.c under PREFIX/include/
11351 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11353 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11354 functions that return function pointers and has support for NT specific
11355 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11356 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11357 unsigned to signed types: this was killing the Win32 compile.
11360 *) Add new certificate file to stack functions,
11361 SSL_add_dir_cert_subjects_to_stack() and
11362 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11363 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11364 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11365 This means that Apache-SSL and similar packages don't have to mess around
11366 to add as many CAs as they want to the preferred list.
11369 *) Experiment with doxygen documentation. Currently only partially applied to
11371 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11372 openssl.doxy as the configuration file.
11375 *) Get rid of remaining C++-style comments which strict C compilers hate.
11376 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11378 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11379 compiled in by default: it has problems with large keys.
11382 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11383 DH private keys and/or callback functions which directly correspond to
11384 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11385 is needed for applications which have to configure certificates on a
11386 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11388 For the RSA certificate situation is makes no difference, but
11389 for the DSA certificate situation this fixes the "no shared cipher"
11390 problem where the OpenSSL cipher selection procedure failed because the
11391 temporary keys were not overtaken from the context and the API provided
11392 no way to reconfigure them.
11393 The new functions now let applications reconfigure the stuff and they
11394 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11395 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11396 non-public-API function ssl_cert_instantiate() is used as a helper
11397 function and also to reduce code redundancy inside ssl_rsa.c.
11398 [Ralf S. Engelschall]
11400 *) Move s_server -dcert and -dkey options out of the undocumented feature
11401 area because they are useful for the DSA situation and should be
11402 recognized by the users.
11403 [Ralf S. Engelschall]
11405 *) Fix the cipher decision scheme for export ciphers: the export bits are
11406 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11407 SSL_EXP_MASK. So, the original variable has to be used instead of the
11408 already masked variable.
11409 [Richard Levitte <levitte@stacken.kth.se>]
11411 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11412 [Richard Levitte <levitte@stacken.kth.se>]
11414 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11415 from `int' to `unsigned int' because it's a length and initialized by
11416 EVP_DigestFinal() which expects an `unsigned int *'.
11417 [Richard Levitte <levitte@stacken.kth.se>]
11419 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11420 script. Instead use the usual Shell->Perl transition trick.
11421 [Ralf S. Engelschall]
11423 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11424 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11425 -noout -modulus' as it's already the case for `openssl rsa -noout
11426 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11427 currently the public key is printed (a decision which was already done by
11428 `openssl dsa -modulus' in the past) which serves a similar purpose.
11429 Additionally the NO_RSA no longer completely removes the whole -modulus
11430 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11432 [Ralf S. Engelschall]
11434 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11435 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11436 [Arne Ansper <arne@ats.cyber.ee>]
11438 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11439 to be added. Now both 'req' and 'ca' can use new objects defined in the
11443 *) Add cool BIO that does syslog (or event log on NT).
11444 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11446 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11447 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11448 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11449 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11452 *) Add preliminary config info for new extension code.
11455 *) Make RSA_NO_PADDING really use no padding.
11456 [Ulf Moeller <ulf@fitug.de>]
11458 *) Generate errors when private/public key check is done.
11461 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11462 for some CRL extensions and new objects added.
11465 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11466 key usage extension and fuller support for authority key id.
11469 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11470 padding method for RSA, which is recommended for new applications in PKCS
11471 #1 v2.0 (RFC 2437, October 1998).
11472 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11473 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11474 against Bleichbacher's attack on RSA.
11475 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11478 *) Updates to the new SSL compression code
11479 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11481 *) Fix so that the version number in the master secret, when passed
11482 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11483 (because the server will not accept higher), that the version number
11484 is 0x03,0x01, not 0x03,0x00
11485 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11487 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11488 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11489 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11492 *) Support for RAW extensions where an arbitrary extension can be
11493 created by including its DER encoding. See apps/openssl.cnf for
11497 *) Make sure latest Perl versions don't interpret some generated C array
11498 code as Perl array code in the crypto/err/err_genc.pl script.
11499 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11501 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11502 not many people have the assembler. Various Win32 compilation fixes and
11503 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11504 build instructions.
11507 *) Modify configure script 'Configure' to automatically create crypto/date.h
11508 file under Win32 and also build pem.h from pem.org. New script
11509 util/mkfiles.pl to create the MINFO file on environments that can't do a
11510 'make files': perl util/mkfiles.pl >MINFO should work.
11513 *) Major rework of DES function declarations, in the pursuit of correctness
11514 and purity. As a result, many evil casts evaporated, and some weirdness,
11515 too. You may find this causes warnings in your code. Zapping your evil
11516 casts will probably fix them. Mostly.
11519 *) Fix for a typo in asn1.h. Bug fix to object creation script
11520 obj_dat.pl. It considered a zero in an object definition to mean
11521 "end of object": none of the objects in objects.h have any zeros
11522 so it wasn't spotted.
11523 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11525 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11526 Masking (CBCM). In the absence of test vectors, the best I have been able
11527 to do is check that the decrypt undoes the encrypt, so far. Send me test
11528 vectors if you have them.
11531 *) Correct calculation of key length for export ciphers (too much space was
11532 allocated for null ciphers). This has not been tested!
11535 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11536 message is now correct (it understands "crypto" and "ssl" on its
11537 command line). There is also now an "update" option. This will update
11538 the util/ssleay.num and util/libeay.num files with any new functions.
11540 perl util/mkdef.pl crypto ssl update
11541 it will update them.
11544 *) Overhauled the Perl interface (perl/*):
11545 - ported BN stuff to OpenSSL's different BN library
11546 - made the perl/ source tree CVS-aware
11547 - renamed the package from SSLeay to OpenSSL (the files still contain
11548 their history because I've copied them in the repository)
11549 - removed obsolete files (the test scripts will be replaced
11550 by better Test::Harness variants in the future)
11551 [Ralf S. Engelschall]
11553 *) First cut for a very conservative source tree cleanup:
11554 1. merge various obsolete readme texts into doc/ssleay.txt
11555 where we collect the old documents and readme texts.
11556 2. remove the first part of files where I'm already sure that we no
11557 longer need them because of three reasons: either they are just temporary
11558 files which were left by Eric or they are preserved original files where
11559 I've verified that the diff is also available in the CVS via "cvs diff
11560 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11561 the crypto/md/ stuff).
11562 [Ralf S. Engelschall]
11564 *) More extension code. Incomplete support for subject and issuer alt
11565 name, issuer and authority key id. Change the i2v function parameters
11566 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11567 what that's for :-) Fix to ASN1 macro which messed up
11568 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11571 *) Preliminary support for ENUMERATED type. This is largely copied from the
11575 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11576 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11578 *) Make sure `make rehash' target really finds the `openssl' program.
11579 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11581 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11582 like to hear about it if this slows down other processors.
11585 *) Add CygWin32 platform information to Configure script.
11586 [Alan Batie <batie@aahz.jf.intel.com>]
11588 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11589 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11591 *) New program nseq to manipulate netscape certificate sequences
11594 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11598 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11599 but the BN code had some problems that would cause failures when
11600 doing certificate verification and some other functions.
11601 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11603 *) Add ASN1 and PEM code to support netscape certificate sequences.
11606 *) Add ASN1 and PEM code to support netscape certificate sequences.
11609 *) Add several PKIX and private extended key usage OIDs.
11612 *) Modify the 'ca' program to handle the new extension code. Modify
11613 openssl.cnf for new extension format, add comments.
11616 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11617 and add a sample to openssl.cnf so req -x509 now adds appropriate
11621 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11622 error code, add initial support to X509_print() and x509 application.
11625 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11626 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11627 stuff is currently isolated and isn't even compiled yet.
11630 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11631 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11632 Removed the versions check from X509 routines when loading extensions:
11633 this allows certain broken certificates that don't set the version
11634 properly to be processed.
11637 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11638 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11639 can still be regenerated with "make depend".
11642 *) Spelling mistake in C version of CAST-128.
11643 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11645 *) Changes to the error generation code. The perl script err-code.pl
11646 now reads in the old error codes and retains the old numbers, only
11647 adding new ones if necessary. It also only changes the .err files if new
11648 codes are added. The makefiles have been modified to only insert errors
11649 when needed (to avoid needlessly modifying header files). This is done
11650 by only inserting errors if the .err file is newer than the auto generated
11651 C file. To rebuild all the error codes from scratch (the old behaviour)
11652 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11653 or delete all the .err files.
11656 *) CAST-128 was incorrectly implemented for short keys. The C version has
11657 been fixed, but is untested. The assembler versions are also fixed, but
11658 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11659 to regenerate it if needed.
11660 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11661 Hagino <itojun@kame.net>]
11663 *) File was opened incorrectly in randfile.c.
11664 [Ulf Möller <ulf@fitug.de>]
11666 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11667 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11668 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11669 al: it's just almost always a UTCTime. Note this patch adds new error
11670 codes so do a "make errors" if there are problems.
11673 *) Correct Linux 1 recognition in config.
11674 [Ulf Möller <ulf@fitug.de>]
11676 *) Remove pointless MD5 hash when using DSA keys in ca.
11677 [Anonymous <nobody@replay.com>]
11679 *) Generate an error if given an empty string as a cert directory. Also
11680 generate an error if handed NULL (previously returned 0 to indicate an
11681 error, but didn't set one).
11682 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11684 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11687 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11688 parameters. This was causing a warning which killed off the Win32 compile.
11691 *) Remove C++ style comments from crypto/bn/bn_local.h.
11692 [Neil Costigan <neil.costigan@celocom.com>]
11694 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11695 based on a text string, looking up short and long names and finally
11696 "dot" format. The "dot" format stuff didn't work. Added new function
11697 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11698 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11699 OID is not part of the table.
11702 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11703 X509_LOOKUP_by_alias().
11706 *) Sort openssl functions by name.
11709 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11710 encryption from sample DSA keys (in case anyone is interested the password
11714 *) Make _all_ *_free functions accept a NULL pointer.
11715 [Frans Heymans <fheymans@isaserver.be>]
11717 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11719 [Anonymous <nobody@replay.com>]
11721 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11722 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11724 *) Don't blow it for numeric -newkey arguments to apps/req.
11725 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11727 *) Temp key "for export" tests were wrong in s3_srvr.c.
11728 [Anonymous <nobody@replay.com>]
11730 *) Add prototype for temp key callback functions
11731 SSL_CTX_set_tmp_{rsa,dh}_callback().
11734 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11735 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11738 *) X509_name_add_entry() freed the wrong thing after an error.
11739 [Arne Ansper <arne@ats.cyber.ee>]
11741 *) rsa_eay.c would attempt to free a NULL context.
11742 [Arne Ansper <arne@ats.cyber.ee>]
11744 *) BIO_s_socket() had a broken should_retry() on Windoze.
11745 [Arne Ansper <arne@ats.cyber.ee>]
11747 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11748 [Arne Ansper <arne@ats.cyber.ee>]
11750 *) Make sure the already existing X509_STORE->depth variable is initialized
11751 in X509_STORE_new(), but document the fact that this variable is still
11752 unused in the certificate verification process.
11753 [Ralf S. Engelschall]
11755 *) Fix the various library and apps files to free up pkeys obtained from
11756 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11759 *) Fix reference counting in X509_PUBKEY_get(). This makes
11760 demos/maurice/example2.c work, amongst others, probably.
11761 [Steve Henson and Ben Laurie]
11763 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11764 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11765 are no longer created. This way we have a single and consistent command
11766 line interface `openssl <command>', similar to `cvs <command>'.
11767 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11769 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11770 BIT STRING wrapper always have zero unused bits.
11773 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11776 *) Make the top-level INSTALL documentation easier to understand.
11779 *) Makefiles updated to exit if an error occurs in a sub-directory
11780 make (including if user presses ^C) [Paul Sutton]
11782 *) Make Montgomery context stuff explicit in RSA data structure.
11785 *) Fix build order of pem and err to allow for generated pem.h.
11788 *) Fix renumbering bug in X509_NAME_delete_entry().
11791 *) Enhanced the err-ins.pl script so it makes the error library number
11792 global and can add a library name. This is needed for external ASN1 and
11793 other error libraries.
11796 *) Fixed sk_insert which never worked properly.
11799 *) Fix ASN1 macros so they can handle indefinite length constructed
11800 EXPLICIT tags. Some non standard certificates use these: they can now
11804 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11805 into a single doc/ssleay.txt bundle. This way the information is still
11806 preserved but no longer messes up this directory. Now it's new room for
11807 the new set of documentation files.
11808 [Ralf S. Engelschall]
11810 *) SETs were incorrectly DER encoded. This was a major pain, because they
11811 shared code with SEQUENCEs, which aren't coded the same. This means that
11812 almost everything to do with SETs or SEQUENCEs has either changed name or
11813 number of arguments.
11814 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11816 *) Fix test data to work with the above.
11819 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11820 was already fixed by Eric for 0.9.1 it seems.
11821 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11823 *) Autodetect FreeBSD3.
11826 *) Fix various bugs in Configure. This affects the following platforms:
11830 unixware-2.0-pentium
11834 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11835 before they are needed.
11838 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11842 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11844 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11845 changed SSLeay to OpenSSL in version strings.
11846 [Ralf S. Engelschall]
11848 *) Some fixups to the top-level documents.
11851 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11852 because the symlink to include/ was missing.
11853 [Ralf S. Engelschall]
11855 *) Incorporated the popular no-RSA/DSA-only patches
11856 which allow to compile a RSA-free SSLeay.
11857 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11859 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11860 when "ssleay" is still not found.
11861 [Ralf S. Engelschall]
11863 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11864 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11866 *) Updated the README file.
11867 [Ralf S. Engelschall]
11869 *) Added various .cvsignore files in the CVS repository subdirs
11870 to make a "cvs update" really silent.
11871 [Ralf S. Engelschall]
11873 *) Recompiled the error-definition header files and added
11874 missing symbols to the Win32 linker tables.
11875 [Ralf S. Engelschall]
11877 *) Cleaned up the top-level documents;
11878 o new files: CHANGES and LICENSE
11879 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11880 o merged COPYRIGHT into LICENSE
11881 o removed obsolete TODO file
11882 o renamed MICROSOFT to INSTALL.W32
11883 [Ralf S. Engelschall]
11885 *) Removed dummy files from the 0.9.1b source tree:
11886 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11887 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11888 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11889 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11890 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11891 [Ralf S. Engelschall]
11893 *) Added various platform portability fixes.
11896 *) The Genesis of the OpenSSL rpject:
11897 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11898 Young and Tim J. Hudson created while they were working for C2Net until
11900 [The OpenSSL Project]
11903 Changes between 0.9.0b and 0.9.1b [not released]
11905 *) Updated a few CA certificates under certs/
11908 *) Changed some BIGNUM api stuff.
11911 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11912 DGUX x86, Linux Alpha, etc.
11915 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11916 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11920 *) Add -strparse option to asn1pars program which parses nested
11922 [Dr Stephen Henson <shenson@bigfoot.com>]
11924 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11927 *) DSA fix for "ca" program.
11930 *) Added "-genkey" option to "dsaparam" program.
11933 *) Added RIPE MD160 (rmd160) message digest.
11936 *) Added -a (all) option to "ssleay version" command.
11939 *) Added PLATFORM define which is the id given to Configure.
11942 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11945 *) Extended the ASN.1 parser routines.
11948 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11951 *) Added a BN_CTX to the BN library.
11954 *) Fixed the weak key values in DES library
11957 *) Changed API in EVP library for cipher aliases.
11960 *) Added support for RC2/64bit cipher.
11963 *) Converted the lhash library to the crypto/mem.c functions.
11966 *) Added more recognized ASN.1 object ids.
11969 *) Added more RSA padding checks for SSL/TLS.
11972 *) Added BIO proxy/filter functionality.
11975 *) Added extra_certs to SSL_CTX which can be used
11976 send extra CA certificates to the client in the CA cert chain sending
11977 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11980 *) Now Fortezza is denied in the authentication phase because
11981 this is key exchange mechanism is not supported by SSLeay at all.
11984 *) Additional PKCS1 checks.
11987 *) Support the string "TLSv1" for all TLS v1 ciphers.
11990 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11991 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11994 *) Fixed a few memory leaks.
11997 *) Fixed various code and comment typos.
12000 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12001 bytes sent in the client random.
12002 [Edward Bishop <ebishop@spyglass.com>]